literature review of health insurance

Health Insurance Literacy and Health Disparities in the United States: A Literature Review

• Emily Vardell Emporia State University

This paper presents a literature review of health insurance literacy with a focus on specialized populations in the U.S. and how limited health literacy skills exacerbate health disparities. This discussion places this issue within the context of contemporary U.S. health care reform and makes connections between health insurance coverage and health disparities. This overview of the research on health insurance literacy covers research across the health insurance spectrum, from awareness of health insurance options to assessments of health literacy skills in specific populations as well as from readability of health insurance informational materials to the availability of multilingual services. In exploring the demographic variables associated with lower health insurance literacy skills, this paper reviews the body of current research in this area to make connections between populations more likely to have unequal access to health care and how having limited skills in navigating the U.S. health care system may compound these disparities. In addition, this paper proposes an Integrated Framework for Health Insurance Literacy as a method for further studying the connections between demographic factors, health coverage, health status, and health insurance literacy skills.

Author Biography

Emily vardell, emporia state university.

Emily Vardell ([email protected]) is an Assistant Professor in the School of Library and Information Management at Emporia State University. She teaches graduate courses on the foundations of library and information science, research methods, reference, consumer health, and health sciences librarianship. Her research interests are in the area of health information behavior with a focus on health insurance literacy and how people make health insurance decisions. Dr. Vardell earned her PhD from the School of Information and Library Science (SILS) at the University of North Carolina at Chapel Hill in 2017 and her Master of Library Science degree from Texas Woman’s University in 2007. Dr. Vardell has received grants from the Medical Library Association and the National Network of Libraries of Medicine, been awarded the Beta Phi Mu Eugene Garfield Doctoral Dissertation Fellowship, and served as a Fulbright scholar in Vienna, Austria.

Copyright (c) 2019 Emily Vardell

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License .

Authors who publish with this journal agree to the following terms: see Author Agreement . Articles are licensed under a CC-BY-NC-ND license.

Information

• For Readers
• For Authors
• For Librarians

ISSN 2574-3430

IJIDI logo created by Craig Taylor.

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here .

Loading metrics

Open Access

Peer-reviewed

Research Article

The impact of public health insurance on health care utilisation, financial protection and health status in low- and middle-income countries: A systematic review

Roles Conceptualization, Data curation, Formal analysis, Investigation, Writing – original draft, Writing – review & editing

* E-mail: [email protected]

Affiliation Department of Health Sciences, University of York, York, England, United Kingdom

Roles Investigation, Methodology, Supervision, Writing – review & editing

Affiliations Centre of Health Economics, University of York, York, England, United Kingdom, Luxembourg Institute of Socio-economic Research (LISER), Luxembourg

Roles Conceptualization, Methodology, Supervision, Writing – review & editing

Affiliations Department of Health Sciences, University of York, York, England, United Kingdom, Department of Epidemiology and Biostatistics, Schulich School of Medicine and Dentistry, Western University, London, Ontario, Canada

Roles Conceptualization, Investigation, Supervision, Writing – review & editing

• Darius Erlangga, 
• Marc Suhrcke, 
• Shehzad Ali, 
• Karen Bloor

• Published: August 28, 2019
• https://doi.org/10.1371/journal.pone.0219731
• Reader Comments

7 Nov 2019: Erlangga D, Suhrcke M, Ali S, Bloor K (2019) Correction: The impact of public health insurance on health care utilisation, financial protection and health status in low- and middle-income countries: A systematic review. PLOS ONE 14(11): e0225237. https://doi.org/10.1371/journal.pone.0225237 View correction

Expanding public health insurance seeks to attain several desirable objectives, including increasing access to healthcare services, reducing the risk of catastrophic healthcare expenditures, and improving health outcomes. The extent to which these objectives are met in a real-world policy context remains an empirical question of increasing research and policy interest in recent years.

We reviewed systematically empirical studies published from July 2010 to September 2016 using Medline, Embase, Econlit, CINAHL Plus via EBSCO, and Web of Science and grey literature databases. No language restrictions were applied. Our focus was on both randomised and observational studies, particularly those including explicitly attempts to tackle selection bias in estimating the treatment effect of health insurance. The main outcomes are: (1) utilisation of health services, (2) financial protection for the target population, and (3) changes in health status.

8755 abstracts and 118 full-text articles were assessed. Sixty-eight studies met the inclusion criteria including six randomised studies, reflecting a substantial increase in the quantity and quality of research output compared to the time period before 2010. Overall, health insurance schemes in low- and middle-income countries (LMICs) have been found to improve access to health care as measured by increased utilisation of health care facilities (32 out of 40 studies). There also appeared to be a favourable effect on financial protection (26 out of 46 studies), although several studies indicated otherwise. There is moderate evidence that health insurance schemes improve the health of the insured (9 out of 12 studies).

Interpretation

Increased health insurance coverage generally appears to increase access to health care facilities, improve financial protection and improve health status, although findings are not totally consistent. Understanding the drivers of differences in the outcomes of insurance reforms is critical to inform future implementations of publicly funded health insurance to achieve the broader goal of universal health coverage.

Citation: Erlangga D, Suhrcke M, Ali S, Bloor K (2019) The impact of public health insurance on health care utilisation, financial protection and health status in low- and middle-income countries: A systematic review. PLoS ONE 14(8): e0219731. https://doi.org/10.1371/journal.pone.0219731

Editor: Sandra C. Buttigieg, University of Malta Faculty of Health Sciences, MALTA

Received: March 19, 2018; Accepted: July 2, 2019; Published: August 28, 2019

Copyright: © 2019 Erlangga et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: The search strategy for this review is available in Supporting Information files.

Funding: The authors received no specific funding for this work.

Competing interests: The authors have declared that no competing interests exist.

Introduction

In recent decades, achieving universal health coverage (UHC) has been a major health policy focus globally.[ 1 – 3 ] UHC entitles all people to access healthcare services through publicly organised risk pooling,[ 4 ] safeguarding against the risk of catastrophic healthcare expenditures.[ 5 ] Low- and middle-income countries (LMICs) face particular challenges in achieving UHC due to particularly limited public resources for health care, inefficient allocation, over-reliance on out-of-pocket payments, and often large population size.[ 5 ] As a result, access to health care and the burden of financial cost in LMICs tends to be worse for the poor, often resulting in forgone care.[ 6 – 8 ]

Introducing and increasing the coverage of publicly organised and financed health insurance is widely seen as the most promising way of achieving UHC,[ 9 , 10 ] since private insurance is mostly unaffordable for the poor.[ 11 ] Historically, social health insurance, tax-based insurance, or a mix of the two have been the dominant health insurance models amongst high income countries and some LMICs, including Brazil, Colombia, Costa Rica, Mexico, and Thailand.[ 12 ] This is partly influenced by the size of the formal sector economy from which taxes and payroll contributions can be collected. In recent decades, community-based health insurance (CBHI) or “mutual health organizations” have become increasingly popular among LMICs, particularly in Sub-Saharan Africa (e.g. Burkina Faso,[ 13 ] Senegal[ 14 ] and Rwanda[ 15 ]) as well as Asia (e.g. China[ 16 ] and India[ 17 ]). CBHI has emerged as an alternative health financing strategy, particularly in cases where the public sector has failed to provide adequate access to health care.[ 18 ]

We searched for existing systematic reviews on health insurance in the Cochrane Database for Systematic Reviews, Medline, Embase, and Econlit. Search terms “health insurance”, “low-middle income countries”, and “utilisation” were used alongside methodological search strategy to locate reviews. Seven systematic reviews were identified of varying levels of quality, [ 19 – 26 ] with Acharya et al.[ 27 ] being the most comprehensive. The majority of existing reviews has suggested that publicly-funded health insurance has typically shown a positive impact on access to care, while the picture for financial protection was mixed, and evidence of the impact on health status was very sparse.

This study reviews systematically the recent fast-growing evidence on the impact of health insurance on health care utilisation, financial protection and health status in LMICs. Since the publication of Acharya et al. (which conducted literature searches in July 2010), the empirical evidence on the impact of health insurance has expanded significantly in terms of quantity and quality, with growing use of sophisticated techniques to account for statistical challenges[ 28 ] (particularly insurance selection bias). This study makes an important contribution towards our understanding of the impact of health insurance in LMICs, taking particular care in appraising the quality of studies. We recognise the heterogeneity of insurance schemes implemented in LMICs and therefore do not attempt to generalise findings, but we aim to explore the pattern emerging from various studies and to extract common factors that may affect the effectiveness of health insurance, that should be the focus of future policy and research. Furthermore, we explore evidence of moral hazard in insurance membership, an aspect that was not addressed in the Acharya et al review.[ 27 ]

This review was planned, conducted, and reported in adherence with PRISMA standards of quality for reporting systematic reviews.[ 29 ]

Participants

Studies focusing on LMICs are included, as measured by per capita gross national income (GNI) estimated using the World Bank Atlas method per July 2016.[ 30 ]

Intervention

Classification of health insurance can be complicated due to the many characteristics defining its structure, including the mode of participation (compulsory or voluntary), benefit entitlement, level of membership (individual or household), methods for raising funds (taxes, flat premium, or income-based premium) and the mechanism and extent of risk pooling [ 31 ]. For the purpose of this review, we included all health insurance schemes organised by government, comprising social health insurance and tax-based health insurance. Private health insurance was excluded from our review, but we recognise the presence of community-based health insurance (CBHI) in many LMICs, especially in Africa and Asia [ 18 ]. We also therefore included CBHI if it was scaled up nationally or was actively promoted by national government. Primary studies that included both public and private health insurance were also considered for inclusion if a clear distinction between the two was made in the primary paper. Studies examining other types of financial incentives to increase the demand for healthcare services, such as voucher schemes or cash transfers, were excluded.

Control group

In order to provide robust evidence on the effect on insurance, it is necessary to compare an insured group with an appropriate control group. In this review, we selected studies that used an uninsured population as the control group. Multiple comparison groups were allowed, but an uninsured group had to be one of them.

Outcome measures

We focus on three main outcomes:

• Utilisation of health care facilities or services (e.g. immunisation coverage, number of visits, rates of hospitalisation).
• Financial protection, as measured by changes in out-of-pocket (OOP) health expenditure at household or individual level, and also catastrophic health expenditure or impoverishment from medical expenses.
• Health status, as measured by morbidity and mortality rates, indicators of risk factors (e.g. nutritional status), and self-reported health status.

The scope of this review is not restricted to any level of healthcare delivery (i.e. primary or secondary care). All types of health services were considered in this review.

Types of studies

The review includes randomized controlled trials, quasi-experimental studies (or “natural experiments”[ 32 ]), and observational studies that account for selection bias due to insurance endogeneity (i.e. bias caused by insurance decisions that are correlated with the expected level of utilisation and/or OOP expenditure). Observational studies that did not take account of selection bias were excluded.

Databases and search terms

A search for relevant articles was conducted on 6 September 2016 using peer-reviewed databases (Medline, Embase, Econlit, CINAHL Plus via EBSCO and Web of Science) and grey literatures (WHO, World Bank, and PAHO). Our search was restricted to studies published since July 2010, immediately after the period covered by the earlier Acharya et al. (2012) review. No language restrictions were applied. Full details of our search strategy are available in the supporting information ( S1 Table ).

Screening and data extraction

Two independent reviewers (DE and MS) screened all titles and abstracts of the initially identified studies to determine whether they satisfied the inclusion criteria. Any disagreement was resolved through mutual consensus. Full texts were retrieved for the studies that met the inclusion criteria. A data collection form was used to extract the relevant information from the included studies.

Assessment of study quality

We used the Grades of Assessment, Development and Evaluation (GRADE) system checklist[ 33 , 34 ] which is commonly used for quality assessment in systematic reviews. However, GRADE does not rate observational studies based on whether they controlled for selection bias. Therefore, we supplemented the GRADE score with the ‘Quality of Effectiveness Estimates from Non-randomised Studies’ (QuEENS) checklist.[ 35 ]

cRandomised studies were considered to have low risk of bias. Non-randomised studies that account for selection on observable variables, such as propensity score matching (PSM), were categorised as high risk of bias unless they provided adequate assumption checks or compared the results to those from other methods, in which case they may be classed as medium risk. Non-randomised studies that account for selection on both observables and unobservables, such as regression with difference-in-differences (DiD) or Heckman sample selection models, were considered to have medium risk of bias–some of these studies were graded as high or low risk depending on sufficiency of assumption checks and comparison with results from other methods.

Heterogeneity of health insurance programmes across countries and variability in empirical methods used across studies precluded a formal meta-analysis. We therefore conducted a narrative synthesis of the literature and did not report the effect size. Throughout this review, we only considered three possible effects: positive outcome, negative outcome, or no statistically significant effect (here defined as p-value > 0.1).

Results of the search

Our database search identified 8,755 studies. Five additional studies were retrieved from grey literature. After screening of titles and abstracts, 118 studies were identified as potentially relevant. After reviewing the full-texts, 68 studies were included in the systematic review (see Fig 1 for the PRISMA diagram). A full description of the included studies is presented in the supporting information ( S2 Table ). Of the 68 included studies, 40 studies examined the effect on utilisation, 46 studies on financial protection, and only 12 studies on health status (see Table 1 ).

• PPT PowerPoint slide
• PNG larger image
• TIFF original image

https://doi.org/10.1371/journal.pone.0219731.g001

https://doi.org/10.1371/journal.pone.0219731.t001

Utilisation of health care

Table 2 collates evidence on the effects of health insurance on utilisation of healthcare services. Three main findings were observed:

• Evidence on utilisation of curative care generally suggested a positive effect, with 30 out of 38 studies reporting a statistically significant positive effect.
• Evidence on preventive care is less clear with 4 out of 7 studies reporting a positive effect, two studies finding a negative effect and one study reporting no effect.
• Among the higher quality studies, i.e. those that suitably controlled for selection bias reflected by moderate or low GRADE score and low risk of bias (score = 3) on QuEENS, seven studies reported a positive relationship between insurance and utilisation. One study[ 36 ] reported no statistically significant effect, and another study found a statistically significant negative effect.[ 37 ]

https://doi.org/10.1371/journal.pone.0219731.t002

Financial protection

Overall, evidence on the impact of health insurance on financial protection is less clear than that for utilisation (see Table 3 ). 34 of the 46 studies reported the impact of health insurance on the level of out-of-pocket health expenditure. Among those 34 studies, 17 found a positive effect (i.e. a reduction in out-of-pocket expenditure), 15 studies found no statistically significant effect, and two studies–from Indonesia[ 59 ] and Peru[ 62 ]–reported a negative effect (i.e. an increase in out-of-pocket expenditure).

https://doi.org/10.1371/journal.pone.0219731.t003

Another financial protection measure is the probability of incurring catastrophic health expenditure defined as OOP exceeding a certain threshold percentage of total expenditure or income. Of the 14 studies reporting this measure, nine reported reduction in the risk of catastrophic expenditure, three found no statistically significant difference, and two found a negative effect of health insurance. Only four studies reported sensitivity analysis varying changes in the threshold level,[ 59 , 62 , 75 , 76 ] though this did not materially affect the findings.

• Two studies used a different measure of financial protection, the probability of impoverishment due to catastrophic health expenditure, reporting conflicting findings.[ 77 , 78 ] Finally, four studies evaluated the effect on financial protection by assessing the impact of insurance on non-healthcare consumption or saving behaviour, such as non-medical related consumption[ 79 ], probability of financing medical bills via asset sales or borrowing[ 40 ], and household saving[ 80 ]. No clear pattern can be observed from those four studies.

Health status

Improving health is one of the main objectives of health insurance, yet very few studies thus far have attempted to evaluate health outcomes. We identified 12 studies, with considerable variation in the precise health measure considered (see Table 4 ). There was some evidence of positive impact on health status: nine studies found a positive effect, one study reported a negative effect, and two studies reported no effect.

https://doi.org/10.1371/journal.pone.0219731.t004

Type of insurance and countries

Considering the heterogeneity of insurance schemes among different countries, we attempted to explore the aggregate results by the type of insurance scheme and by country. Table 5 provides a summary of results classified by three type of insurance scheme: community-based health insurance, voluntary health insurance (non-CBHI), and compulsory health insurance. This division is based on the mode of participation (compulsory vs voluntary), which may affect the presence of adverse selection and moral hazard. Premiums are typically community-rated in CBHI, risk-rated in voluntary schemes and income-rated in compulsory schemes.

https://doi.org/10.1371/journal.pone.0219731.t005

In principle, CBHI is also considered a voluntary scheme, but we separated it to explore whether the larger size of pooling from non-CBHI schemes may affect the outcomes. Social health insurance is theoretically a mandatory scheme that requires contribution from the enrolees. However, in the context of LMICs, the mandatory element is hard to enforce, and in practice the scheme adopts a voluntary enrolment. Additionally, the government may also want to subsidise the premium for poor people. Therefore, in this review SHI schemes can fall into either the voluntary health insurance (non-CBHI) or compulsory health insurance (non-CBHI), depending on the target population defined in the evaluation study. Lastly, we chose studies with high quality/low risk only to provide more robust results.

Based on the summary in Table 5 , the effect on utilisation overall does not differ based on type of insurance, with most evidence suggesting an overall increase in utilisation by the insured. The two studies showing no effect or reduced consumption of care were conducted in two different areas of India, which may–somewhat tentatively–suggest a common factor unique to India’s health system that may compromise the effectiveness of health insurance in increasing utilisation.

Regarding financial protection, the evidence for both CBHI and non-CBHI voluntary health insurance is inconclusive. Furthermore, there is an indication of heterogeneity by supply side factors captured by proximity to health facilities. Evidence from studies exploring subsidised schemes suggests no effect on financial protection, even a negative effect among the insured in Peru.

Lastly, evidence for health status may be influenced by how health outcomes are measured. Studies exploring specific health status, (examples included health indexes, wasting, C-reactive protein, and low birth weight), show a positive effect, whereas studies using mortality rates tends to show no effect or even negative effects. Studies exploring CBHI scheme did not find any evidence of positive effect on health status, as measured either by mortality rate or specific health status.

This review synthesises the recent, burgeoning empirical literature on the impact of health insurance in LMICs. We identified a total of 68 eligible studies over a period of six years–double the amount identified by the previous review by Acharya et al. over an approximately 60-year time horizon (1950—July 2010). We used two quality assessment checklists to scrutinise the study methodology, taking more explicit account of the methodological robustness of non-experimental designs.

Programme evaluation has been of interest to many researchers for reporting on the effectiveness of a public policy to policymakers. In theory, the gold standard for a programme evaluation is the randomised control trial, in which the treatment is randomly assigned to the participants. The treatment assignment process has to be exogenous to ensure that any observed effect between the treated and control groups can only be caused by the difference in the treatment assignment. Unfortunately, this ideal scenario is often not feasible in a public policy setting. Our findings showed that only three papers between 2010 and 2016 were able to conduct a randomised study to evaluate the impact of health insurance programmes in developing countries, particularly CBHI [ 38 , 75 , 103 ]. Policymakers may believe in the value of an intervention regardless of its actual evidence base, or they may believe that the intervention is beneficial and that no one in need should be denied it. In addition, policymakers are inclined to demonstrate the effectiveness of an intervention that they want implemented in the most promising contexts, as opposed to random allocation [ 104 ].

Consequently, programme evaluators often have to deal with a non-randomised treatment assignment which may result in selection bias problems. Selection bias is defined as a spurious relationship between the treatment and the outcome of interest due to the systematic differences between the treated and the control groups [ 105 ]. In the case of health insurance, an individual who chooses to enrol in the scheme may have different characteristics to an individual who chooses not to enrol. When those important characteristics are unobservable, the analyst needs to apply more advanced techniques and, sometimes, stronger assumptions. Based on our findings, we noted several popular methods, including propensity score matching (N = 8), difference-in-difference (N = 10), fixed or random effects of panel data (N = 6), instrumental variables (N = 12) and regression discontinuity (N = 6). Those methods have varying degree of success in controlling the unobserved selection bias and analysts should explore the robustness of their findings by comparing initial findings with other methods by testing important assumptions. We noted some papers combining two common methods, such as difference-in-difference with propensity score matching (N = 10) and fixed effects with instrumental variables (N = 8), in order to obtain more robust results.

Overall effect

Compared with the earlier review, our study has found stronger and more consistent evidence of positive effects of health insurance on health care utilisation, but less clear evidence on financial protection. Restricting the evidence base to the small subset of randomised studies, the effects on financial protection appear more consistently positive, i.e. three cluster randomised studies[ 39 , 75 , 76 ] showed a decline in OOP expenditure and one randomised study[ 36 ] found no significant effect.

Besides the impact on utilisation and financial protection, this review identified a number of good quality studies measuring the impact of health insurance on health outcomes. Twelve studies were identified (i.e. twice as many as those published before 2010), nine of which showed a beneficial health effect. This holds for the subset of papers with stronger methodology for tackling selection bias.[ 39 , 49 , 89 , 103 ] In cases where a health insurance programme does not have a positive effect on either utilisation, financial protection, and health status, it is particularly important to understand the underlying reasons.

Possible explanation of heterogeneity

Payment system..

Heterogeneity of the impact of health insurance may be explained by differences in health systems and/or health insurance programmes. Robyn et al. (2012) and Fink et al (2013) argued that the lack of significant effect of insurance in Burkina Faso may have been partially influenced by the capitation payment system. As the health workers relied heavily on user fees for their income, the change of payment system from fee-for-services to capitation may have discouraged provision of high quality services. If enrolees perceive the quality of contracted providers as bad, they might delay seeking treatment, which in turn could impact negatively on health.

Several studies from China found the utilisation of expensive treatment and higher-level health care facilities to have increased following the introduction of the insurance scheme.[ 41 , 44 , 45 , 88 ] A fee-for-service payment system may have incentivised providers to include more expensive treatments.[ 43 , 83 , 88 ] Recent systematic reviews suggested that payment systems might play a key role in determining the success of insurance schemes,[ 23 , 106 ] but this evidence is still weak, as most of the included studies were observational studies that did not control sufficiently for selection bias.

Uncovered essential items.

Sood et al. (2014) found no statistically significant effect of community-based health insurance on utilisation in India. They argued that this could be caused by their inability to specify the medical conditions covered by the insurance, causing dilution of a potential true effect. In other countries, transportation costs[ 69 ] and treatments that were not covered by the insurance[ 59 , 60 ] may explain the absence of a reduction in out-of-pocket health expenditures.

Methodological differences.

Two studies in Georgia evaluated the same programme but with different conclusions.[ 50 , 51 ] This discrepancy may be explained by the difference in the estimated treatment effect: one used average treatment effect (ATE), finding no effect, and another used average treatment effect on the treated (ATT), reporting a positive effect. ATE is of prime interest when policymakers are interested in scaling up the programme, whereas ATT is useful to measure the effect on people who were actually exposed to insurance.[ 107 ]

Duration of health insurance.

We also found that the longer an insurance programme has been in place prior to the timing of the evaluation, the higher the odds of improved health outcomes. It is plausible that health insurance would not change the health status of population instantly upon implementation.[ 21 ] While there may be an appetite among policymakers to obtain favourable short term assessments, it is important to compare the impact over time, where feasible.

Moral hazard.

Acharya et al (2012) raised an important question about the possibility of a moral hazard effect as an unintended consequence of introducing (or expanding) health insurance in LMICs. We found seven studies exploring ex-ante moral hazard by estimating the effect on preventive care. If uninsured individuals expect to be covered in the future, they may reduce the consumption of preventive care or invest less in healthy behaviours.[ 108 , 109 ] Current overall evidence cannot suggest a definite conclusion considering the heterogeneity in chosen outcomes. One study found that the use of a self-treated bed nets to prevent malaria declined among the insured group in Ghana[ 54 ] while two studies reported an increase in vaccination rates[ 62 ] and the number of prenatal care visits[ 55 , 62 ]among the insured group. Another study reported no evidence that health insurance encouraged unhealthy behaviour or reduction of preventive efforts in Thailand.[ 66 ]

Two studies from Colombia found that the insured group is more likely to increase their demand for preventive treatment.[ 47 , 49 ] As preventive treatment is free for all, both authors attributed this increased demand to the scheme’s capitation system, incentivising providers to promote preventive care to avoid future costly treatments.[ 110 ] Another study of a different health insurance programme in Colombia found an opposite effect.[ 48 ]

Study limitations.

This review includes a large variety of study designs and indicators for assessing the multiple potential impacts of health insurance, making it hard to directly compare and aggregate findings. For those studies that used a control group, the use of self-selected controls in many cases creates potential bias. Studies of the effect of CBHI are often better at establishing the counterfactual by allowing the use of randomisation in a small area, whereas government schemes or social health insurance covering larger populations have limited opportunity to use randomisation. Non-randomised studies are more susceptible to confounding factors unobserved by the analysts. For a better understanding of the links between health insurance and relevant outcomes, there is also a need to go beyond quantitative evidence alone and combine the quantitative findings with qualitative insights. This is particularly important when trying to interpret some of the counterintuitive results encountered in some studies.

The impact of different health insurance schemes in many countries on utilisation generally shows a positive effect. This is aligned with the supply-demand theory in whichhealth insurance decreases the price of health care services resulting in increased demand. It is difficult to draw an overall conclusion about the impact of health insurance on financial protection, most likely because of differences in health insurance programmes. The impact of health insurance on health status suggests a promising positive effect, but more studies from different countries is required.

The interest in achieving UHC via publicly funded health insurance is likely to increase even further in the coming years, and it is one of the United Nation’s Sustainable Development Goals (SDGs) for 2030[ 111 ]. As public health insurance is still being widely implemented in many LMICs, the findings from this review should be of interest to health experts and policy-makers at the national and the international level.

Supporting information

S1 table. search strategies..

https://doi.org/10.1371/journal.pone.0219731.s001

S2 Table. Study characteristic and reported effect from the included studies (N = 68).

https://doi.org/10.1371/journal.pone.0219731.s002

S3 Table. PRISMA 2009 checklist.

https://doi.org/10.1371/journal.pone.0219731.s003

• View Article
• Google Scholar
• PubMed/NCBI
• 4. World Health Organization. Research for universal health coverage: World health report 2013. WHO. Geneva: World Health Organization; 2014.
• 5. World Health Organization. The world health report 2010. Health systems financing: the path to universal coverage. World Health Organization; 2010.
• 9. Maeda A, Araujo E, Cashin C, Harris J, Ikegami N, Reich MR. Universal Health Coverage for Inclusive and Sustainable Development: A Synthesis of 11 Country Case Studies. The World Bank; 2014. https://doi.org/10.1596/978-1-4648-0297-3
• 10. Jowett M, Kutzin J. Raising revenues for health in support of UHC: strategic issues for policy makers. world Health Organization. Geneva; 2015. Report No.: 1. https://doi.org/10.1080/13545701.2015.1088658
• 12. Wang H, Switlick K, Ortiz C, Zurita B, Connor C. Health Insurance Handbook. The World Bank; 2011. https://doi.org/10.1596/978-0-8213-8982-9
• 21. Giedion U, Alfonso EA, Díaz Y, Andrés Alfonso E, Díaz Y. The Impact of Universal Coverage Schemes in the Developing World: A Review of the Existing Evidence. Univers Heal Cover Stud Ser (UNICO), No 25. Washington DC: World; 2013;
• 31. A System of Health Accounts 2011. OECD; 2017. https://doi.org/10.1787/9789264270985-en
• 34. Cochrane . Cochrane handbook for systematic reviews of interventions. Higgins JPT, Green SE, editors. Wiley-Blackwell; 2008.
• 37. Sheth K. Evaluating Health-Seeking Behavior, Utilization of Care, and Health Risk: Evidence from a Community Based Insurance Model in India. 2014. Report No.: 36.
• 105. Wooldridge JM. Introductory Econometrics: A Modern Approach. Fifth Inte. Mason, Ohio: South-Western Cengage Learning; 2013.
• 111. World Health Organization. World Health Statistics 2016: Monitoring health for the SDGs. WHO. Geneva: World Health Organization; 2017.

An official website of the United States government

The .gov means it's official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you're on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings
• Browse Titles

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Institute of Medicine (US) Committee on the Consequences of Uninsurance. Care Without Coverage: Too Little, Too Late. Washington (DC): National Academies Press (US); 2002.

Care Without Coverage: Too Little, Too Late.

• Hardcopy Version at National Academies Press

3 Effects of Health Insurance on Health

This chapter presents the Committee's review of studies that address the impact of health insurance on various health-related outcomes. It examines research on the relationship between health insurance (or lack of insurance), use of medical care and health outcomes for specific conditions and types of services, and with overall health status and mortality. There is a consistent, positive relationship between health insurance coverage and health-related outcomes across a body of studies that use a variety of data sources and different analytic approaches. The best evidence suggests that health insurance is associated with more appropriate use of health care services and better health outcomes for adults.

The discussion of the research in this chapter is organized within sections that encompass virtually all of the research literature on health outcomes and insurance status that the Committee identified. The chapter sections include the following:

• Primary prevention and screening services
• Cancer care and outcomes
• Chronic disease management, with specific discussions of diabetes, hypertension, end-stage renal disease (ESRD), HIV disease, and mental illness
• Hospital-based care (emergency services, traumatic injury, cardiovascular disease)
• Overall mortality and general measures of health status

The Committee consolidated study results within categories that reflect both diseases and services because these frameworks helped in summarizing the individual studies and subsumed similar research structures and outcome measures. Older studies and those of lesser relevance or quality are not discussed within this chapter devoted to presenting study results and reaching Committee findings. However, all of the studies reviewed are described briefly in Appendix B .

The studies presented in some detail in this chapter are those that the Committee judged to be both methodologically sound and the most informative regarding health insurance effects on health-related outcomes. 1 Most studies report a positive relationship between health insurance coverage and measured outcomes. However, all studies with negative results that are contrary to the Committee's findings are presented and discussed in this chapter. Appendix B includes summaries of the complete set of studies that the Committee reviewed.

In the pages that follow, the Committee's findings introduce each of the five major sections listed above and also some of the subsections under chronic disease and hospital-based care. All of the Committee's specific findings are also presented together in Box 3.12 in the concluding section of this chapter. These findings are the basis for the Committee's overall conclusions in Chapter 4 .

Specific Committee Findings. Uninsured adults are less likely than adults with any kind of health coverage to receive preventive and screening services and less likely to receive these services on a timely basis. Health insurance that provides more extensive (more...)

• PRIMARY PREVENTION AND SCREENING SERVICES

Finding: Uninsured adults are less likely than adults with any kind of health coverage to receive preventive and screening services and less likely to receive these services on a timely basis. Health insurance that provides more extensive coverage of preventive and screening services is likely to result in greater and more appropriate use of these services.

Finding: Health insurance may reduce racial and ethnic disparities in the receipt of preventive and screening services.

These findings have important implications for health outcomes, as can be seen in the later sections on cancer and chronic diseases. For prevention and screening services, health insurance facilitates both the receipt of services and a continuing care relationship or regular source of care, which also increases the likelihood of receiving appropriate care.

Insurance benefits are less likely to include preventive and screening services ( Box 3.2 ) than they are physician visits for acute care or diagnostic tests for symptomatic conditions. However, over time, coverage of preventive and screening services has been increasing. In 1998, about three-quarters of adults with employment-based health insurance had a benefit package that included adult physical examinations; two years later in 2000, the proportion had risen to 90 percent (KPMG, 1998; Kaiser Family Foundation/HRET, 2000). Yet even if health insurance benefit packages do not cover preventive or screening services, those with health insurance are more likely to receive these recommended services because they are more likely to have a regular source of care, and having a regular source of care is independently associated with receiving recommended services (Bush and Langer, 1998; Gordon et al., 1998; Mandelblatt et al., 1999; Zambrana et al., 1999; Cummings et al., 2000; Hsia et al., 2000; Breen et al., 2001). The effect of having health insurance is more evident for relatively costly services, such as mammograms, than for less costly services, such as a clinical breast exam (CBE) or Pap test (Zambrana et al., 1999; Cummings et al., 2000; O'Malley et al., 2001).

Screening Services. The U.S. Preventive Services Task Force (USPSTF) recommends screening for the following conditions in the general adult population under age 65: cervical cancer (above age 18), breast and colorectal cancer (above age 50), hypertension (more...)

According to several large population surveys conducted within the past decade, adults without health insurance are less likely to receive recommended preventive and screening services and are less likely to receive them at the frequencies recommended by the United States Preventive Services Task Force than are insured adults. 2 The 1992 National Health Interview Survey (NHIS) documented receipt of mammography, CBE, Pap test, fecal occult blood test (FOBT), sigmoidoscopy, and digital rectal exam by adults under 65 (Potosky et al., 1998). Those with no health insurance had significantly lower screening rates compared to those with private coverage and compared to those with Medicaid for every service except sigmoidoscopy. The odds ratios (ORs) for receiving a screening service if uninsured compared with having private health insurance ranged from 0.27 for mammography to 0.43 for Pap test. 3

The 1998 NHIS found that, although rates of screening at appropriate intervals had increased generally over the preceding decade, they remained substantially lower for uninsured adults than for those with any kind of health insurance (Breen et al., 2001). 4 In a multivariable analysis that adjusted for age, race, education, and a regular source of care, uninsured adults were significantly less likely than those with any kind of coverage to receive a Pap test, mammography, and colorectal screening (FOBT or sigmoidoscopy) (ORs ranged from 0.37 to 0.5) (Breen et al., 2001). The study reported a strong relationship between having a regular source of care and timely receipt of these screening services in addition to the relationship between health insurance and screening.

Studies using other national samples report results consistent with those of the NHIS. A study of more than 31,000 women between ages 50 and 64 who responded to telephone surveys conducted between 1994 and 1997 about their receipt of mammograms, Pap smears, and colorectal cancer screening (either FOBT or sigmoidoscopy) found that uninsured women were significantly less likely to have received these tests than were women with private prepaid plan insurance (ORs ranging from 0.30 to 0.50) (Hsia et al., 2000). This study also found a strong relationship between having a regular source of care and receipt of screening services. Health insurance was an independently significant predictor. Another study based on several years of the Behavioral Risk Factor Surveillance System (BRFSS) for older adults (55 through 64) found that uninsured men and women were much less likely than their insured counterparts to receive cancer or heart disease screening tests and also much less likely to have a regular source of care (Powell-Griner et al., 1999; see Table 4.1 ).

Disparities Among Population Groups

A review of the literature on the interaction of race, ethnicity, and socioeconomic status (SES) with health insurance, concluded that health insurance makes a positive contribution to the likelihood of receiving appropriate screening services, although racial and ethnic disparities persist independent of health insurance (Haas and Adler, 2001). Studies of the use of preventive services by particular ethnic groups, such as Hispanics and African Americans, find that health insurance is associated with increased receipt of preventive services and increased likelihood of having a regular source of care, which improves one's chances of receiving appropriate preventive services (Solis et al., 1990; Mandelblatt et al., 1999; Zambrana et al., 1999; Wagner and Guendelman, 2000; Breen et al., 2001; O'Malley et al., 2001).

Breen and colleagues (2001) modeled the expected increase in screening rates for different ethnic groups if they were to gain health insurance coverage and a regular source of care. This “what-if” model suggests that those groups for whom screening rates are particularly low (e.g., receipt of mammography by Hispanic women, colorectal screening of African-American men) would make the largest gains (an 11 percentage-point increase in mammography rates for Hispanic women [to 77 percent] and a 5 percentage-point increase in colorectal screening for African-American men [to 31 percent] (Breen et al., 2001).

Extensiveness of Insurance Benefits

The type of health insurance and the continuity of coverage have also been found to affect receipt of appropriate preventive and screening services. Faulkner and Schauffler (1997) examined receipt of physical examinations, blood pressure screening, lipid screening for detection of cardiovascular disease, Pap test, CBE, and mammography and identified a positive and statistically significant “dose– response” relationship between the extent of coverage for preventive services (e.g., whether all such services, most, some, or none were covered by health insurance). Insurance coverage for preventive care increased men's receipt of preventive services more than it did that of women. Men with no coverage for preventive services were much less likely than men with complete coverage for such services to receive them (ORs for receipt of specific services ranged from 0.36 to 0.56). Women with no preventive services coverage also received fewer of these services than did women with full coverage for them (ORs for specific services ranged from 0.5 to 0.83) (Faulkner and Schauffler, 1997).

Ayanian and colleagues (2000) used the 1998 BRFSS data set to analyze the effect of length of time without coverage on receipt of preventive and screening services for adults between ages 18 and 65. Those without coverage for a year or longer were more likely than those uninsured for less than one year to go without appropriate preventive and screening services. For every generally recommended service (mammography, CBE, Pap smear, FOBT, sigmoidoscopy, hypertension screening, and cholesterol screening), the longer-term uninsured were significantly less likely than persons with any form of health insurance to receive these services (Ayanian et al., 2000).

Negative Findings

In the Committee's review, the one study that did not find a positive effect of insurance coverage compared mammography use among clients of various sites of care in Detroit, Michigan: two health department clinics, a health maintenance organization (HMO), and a private hospital (Burack et al., 1993). This study found no significant differences among women according to their health insurance status but did find that patients with more visits annually for any service (seven or more) were more likely to receive mammography. All women in this study had access to a primary care provider and, in the case of uninsured women, to clinics with the mission of serving the uninsured. These factors may explain why uninsured women had mammography rates as high as those of women with insurance.

• CANCER CARE AND OUTCOMES

Finding: Uninsured cancer patients generally have poorer outcomes and are more likely to die prematurely than persons with insurance, largely because of delayed diagnosis. This finding is supported by population-based studies of breast, cervical, colorectal, and prostate cancer and melanoma.

The studies analyzing health-related outcomes for cancer patients provide some of the most compelling evidence for the effect of health insurance status on health outcomes ( Box 3.3 ). This evidence comes from research based on area or statewide cancer registries, which provide large numbers of observations and reflect almost all cases occurring in a geographic region. Multivariable data analysis is used to determine the independent effects of health insurance, by controlling for demographic, SES, and clinical differences among study subjects.

Cancer. Cancers of all kinds have an overall incidence nationally of 400 cases per 100,000 people each year. More than 8.9 million Americans alive today have a history of cancer. Cancers account for approximately 550,000 deaths each year in the United (more...)

In addition to receiving fewer cancer screening services, uninsured adults are at greater risk of late-stage, often fatal cancer. Early diagnosis frequently improves the chances of surviving cancer. Generally, in studies examining the stage at which cancer is diagnosed, those with private health insurance have the best outcomes and those with no insurance have the worst (i.e., the highest proportion of late-stage diagnoses), with intermediate outcomes for Medicaid enrollees. In some studies however, the outcomes for Medicaid enrollees are comparable to those for uninsured cancer patients (Roetzheim et al., 1999). Both because of an assumption of similarity in SES between uninsured and Medicaid patients and because of small numbers of observations in the separate categories, some studies report combined results for Medicaid and uninsured patients and compare these findings with those for privately insured patients (e.g., Lee-Feldstein et al., 2000).

In studies assessing the outcomes for adults with cancer—stage of disease at diagnosis and mortality—Medicaid enrollees often do no better, and sometimes do worse, than uninsured patients. This similarity in experience between patients enrolled in Medicaid and those without any coverage may reflect the fact that uninsured persons in poor health, once they seek care, may become enrolled in Medicaid as a result of their frequent interactions with the health care system (Davidoff et al., 2001; see Box 2.1 ). Also, Medicaid enrollees tend to have discontinuous coverage and thus may have had less regular access to screening services. Consequently, persons with Medicaid at the time of a cancer diagnosis may have been without coverage for some prior period (Carrasquillo et al., 1998; IOM, 2001a; Perkins et al., 2001). For example, one study of women under 65 with Medi-Cal coverage (California's Medicaid and indigent care program) who were diagnosed with breast cancer found that, among those who had been uninsured during the year prior to their diagnosis (18 percent of all Medi-Cal enrollees), late-stage diagnosis was much more likely than among those who had been continuously enrolled for the previous 12 months (ORs of 3.9 for those who had been uninsured and 1.4 for those continuously covered by Medi-Cal, compared with all other women ages 30–64 diagnosed with breast cancer) (Perkins et al., 2001).

With this general background on the nature of the research examining health insurance status effects, the remainder of this section discusses study results for five specific cancers.

Breast Cancer

Uninsured women and women with Medicaid are more likely to receive a breast cancer diagnosis at a late stage of disease (regional or distant) and have a 30– 50 percent greater risk of dying than women with private coverage, as shown in studies based on three different state or regional cancer registries (Ayanian et al., 1993; Roetzheim et al., 1999, 2000; Lee-Feldstein et al., 2000).

In a study using the New Jersey Cancer Registry, Ayanian and colleagues (1993) identified 4,675 women 35 to 65 years of age diagnosed with breast cancer and assessed their stage of disease at diagnosis and their survival rates 4.5 to 7 years after diagnosis. The authors found that uninsured women were significantly more likely than privately insured women to be diagnosed with regional or late-stage cancer, as were patients with Medicaid. After controlling for stage of disease at diagnosis and other factors, uninsured women had an adjusted risk of death 49 percent higher than that of privately insured women, and women with Medicaid had a 40 percent higher risk of death than those who were privately insured.

Using a regional cancer registry and Census data for 1987 through 1993, Lee-Feldstein and colleagues (2000) examined the stage of disease at diagnosis, treatment, and survival experience of about 1,800 northern California women under the age of 65 diagnosed with breast cancer. They found that women who were uninsured and publicly insured (primarily Medicaid), taken together, were twice as likely as privately insured women with indemnity coverage to be diagnosed at a late stage of disease. Over a four- to ten-year follow-up, uninsured and publicly insured women had higher risks of death from both breast cancer (42 percent higher) and all causes (46 percent higher) than did privately insured women with indemnity coverage. The likelihood of receiving breast-conserving surgery did not differ between these two groups.

In a review of approximately 9,800 Florida residents diagnosed with breast cancer in 1994, Roetzheim and colleagues calculated that, after controlling for age, education, income, marital status, race, and comorbidity, women without insurance were more likely to be diagnosed with late-stage disease than women with private indemnity coverage (OR = 1.43) (Roetzheim et al., 1999). Women with Medicaid had an even greater likelihood of late-stage diagnosis compared with privately insured women (OR = 1.87). In a subsequent analysis of mortality using the same registry data, the authors estimated that the relative risk (RR) of dying was 31 percent higher for uninsured women and 58 percent higher for women with Medicaid over a three to four-year follow-up period (Roetzheim et al, 2000a). Further analysis suggested that stage of disease at diagnosis and, to a lesser extent, treatment modality appeared to account for the differences in survival by insurance status. Finally, uninsured women were less likely than women with private coverage to receive breast-conserving surgery when stage at diagnosis, comorbidities, and other personal characteristics were taken into account (OR = 0.70) (Roetzheim et al., 2000a).

Cervical Cancer

Uninsured women are more likely to receive a late-stage diagnosis for invasive cervical cancer than are privately insured women. Ferrante and colleagues (2000) analyzed 852 cases of invasive cervical cancer reported in the Florida tumor registry for 1994 to determine factors associated with late-stage diagnosis. In bivariate analysis, being uninsured was associated with an increased likelihood of late-stage diagnosis (OR = 1.6). In a multivariable analysis that adjusted for age, education, income, marital status, race, comorbidities, and smoking, uninsured women were more likely to present with a late-stage cancer compared to women with private indemnity coverage, although this finding was not statistically significant (OR = 1.49, confidence interval [CI]: 0.88–2.50). The outcome for Medicaid enrollees was similar to that of privately insured women in both bivariate and multivariable analysis (Ferrante et al., 2000).

Colorectal Cancer

Uninsured patients with colorectal cancer have a greater risk of dying than do patients with private indemnity insurance, even after adjusting for differences in the stage at which the cancer is diagnosed and the treatment modality. Using the Florida cancer registry for 1994, Roetzheim and colleagues (1999) analyzed the relative likelihood of late-stage diagnosis by insurance status for more than 8,000 cases of colorectal cancer. In a multivariable analysis adjusting for sociodemographic characteristics, smoking status, and comorbidities, uninsured patients were more likely to be diagnosed with late-stage colorectal cancer than were patients with private indemnity coverage (OR = 1.67). Medicaid enrollees had a statistically insignificant greater likelihood of late-stage disease compared to patients with indemnity coverage (OR = 1.44, CI: 0.92–2.25).

A subsequent analysis of largely the same data set (9,500 cases) that adjusted for sociodemographic factors and comorbidities but not for smoking estimated the adjusted mortality risk for uninsured patients with colorectal cancer to be 64 percent greater over a three- to four-year follow-up period than that for patients covered by private indemnity plans (Roetzheim et al., 2000b). 5 Even after adjusting for stage of disease at diagnosis, the risk of death for uninsured patients was 50 percent higher than that for the privately insured, and after further adjustment for treatment modality, the risk for uninsured patients was 40 percent higher (Roetzheim et al., 2000b).

Prostate Cancer

In addition to delayed diagnosis and greater risk of death, uninsured prostate cancer patients have been found to experience a decrease in health-related quality of life after their diagnosis and during treatment, unlike publicly and privately insured patients. A study of about 8,700 cases of newly diagnosed prostate cancer reported to the Florida cancer registry in 1994 found that uninsured men were more likely to be diagnosed at a late stage of the disease than were men with private indemnity insurance (OR = 1.47) (Roetzheim et al., 1999). A study of 860 men in 26 medical practices with newly diagnosed prostate cancer evaluated their health-related quality of life (HRQOL) at three- to six-month intervals over a two-year period (Penson et al., 2001). Although uninsured men diagnosed with prostate cancer did not have a lower HRQOL at diagnosis, their HRQOL decreased over the course of their disease and treatment, in contrast to that of HMO and Medicare patients. The authors suggest that “patients undergoing aggressive treatment, which can itself have deleterious effects on quality of life, are exposed to further hardships when they do not have comprehensive health insurance upon which to support their care” (Penson et al., 2001, p. 357).

Uninsured patients, as well as Medicaid patients have been found to be more likely to be diagnosed with late-stage melanoma than are privately insured patients. Among 1,500 patients diagnosed with melanoma, uninsured patients were more likely to have late-stage (regional or distant) disease than those with private indemnity coverage (OR = 2.6) (Roetzheim et al., 1999). The small number of Medicaid patients with melanoma (13) included in this study also had a much greater chance of being diagnosed with late-stage cancer.

• CHRONIC DISEASE CARE AND OUTCOMES

Finding: Uninsured people with chronic diseases are less likely to receive appropriate care to manage their health conditions than are those who have health insurance. For the five disease conditions that the Committee examined (diabetes, cardiovascular disease, end-stage renal disease, HIV infection, and mental illness), uninsured patients have worse clinical outcomes than insured patients.

Effective management of chronic conditions such as diabetes, hypertension, HIV, and depression ( Box 3.4 ) includes not only periodic services and care from health care professionals but also the active involvement of patients in modifying their behavior, monitoring their condition, and participating in treatment regimens (Wagner et al., 1996; Davis et al., 2000). Identifying chronic conditions early and providing appropriate health care on an ongoing and coordinated basis are health care system goals that have been developed over several decades and have been continuously refined as evidence for cost-effective interventions and practices has accumulated. Maintaining an ongoing relationship with a specific provider who keeps records, manages care, and is available for consultation between visits is a key to high-quality health care, particularly for those with chronic illnesses (O'Connor et al., 1998; IOM, 2001b).

Chronic Conditions. Chronic conditions are the leading causes of death, disability, and illness in the United States, accounting for one-third of the potential life years lost before age 65 (CDC, 2000a). Almost 100 million Americans have chronic conditions. (more...)

For persons with a chronic illness, health insurance may be most important in that it enhances the opportunities to acquire a regular source of care. If someone has coverage through a private or public managed care plan, a relationship with a primary care provider may be built into the insurance. Indemnity or fee-for-service (FFS) insurance coverage also improves the chances of having a regular source of care because having the resources to pay for services is often a prerequi-site to being seen in a medical practice. Uninsured adults are much less likely to have a regular source of care and are more likely to identify an emergency department as their regular source of care than are adults with any form of coverage (Weinick et al., 1997; Cunningham and Whitmore, 1998; Zuvekas and Weinick, 1999; Haley and Zuckerman, 2000). Loss of coverage also interrupts patterns of use of health care and results in delays in seeking needed care (Burstin et al., 1998; Kasper et al., 2000; Hoffman et al., 2001). For uninsured adults under age 65, 19 percent with heart disease and 14 percent with hypertension lack a usual source of care, compared to 8 and 4 percent, respectively, of their insured counterparts (Fish-Parcham, 2001). For uninsured patients without a regular source of care or those who identify an emergency department as their usual source, obtaining care that is consistent with recognized standards for effective disease management is a daunting challenge.

Providers with a commitment to serving uninsured clients, such as local public health and hospital clinics and federally funded community health centers, have sometimes instituted special interventions and programs for the chronically ill to promote continuity of care and disease management. These innovations are critically important to the identified, chronically ill patients who routinely receive care at such clinics and centers. The efforts of these providers, however, are limited in scale by funding and service capacity relative to the high need for care within their service areas (Baker et al., 1998; Chin et al., 2000; Piette, 2000; Philis-Tsimikas and Walker, 2001). As demonstrated in the following review of studies examining the care and outcomes for patients with specific chronic conditions, those who do not have health insurance coverage of any kind fare measurably worse than their insured counterparts.

Cardiovascular Disease

Finding: Uninsured adults with hypertension or high cholesterol have diminished access to care, are less likely to be screened, are less likely to take prescription medication if diagnosed, and experience worse health outcomes.

Across the spectrum of services and the course of development of cardiovascular disease ( Box 3.5 ), uninsured adults receive fewer services and experience worse health. They are less likely to receive screening for hypertension and high cholesterol and to have frequent monitoring of blood pressure once they develop hypertension. Uninsured adults are less likely to stay on drug therapy for hypertension both because they lack a regular provider and because they do not have insurance coverage. Loss of insurance coverage has been demonstrated to disrupt therapeutic relationships and worsen control of blood pressure.

Cardiovascular Disease. “Cardiovascular disease” encompasses a variety of diseases and conditions that affect the heart and blood vessels, including hypertension (high blood pressure), heart disease, and stroke. One-quarter of all Americans (more...)

Uninsured adults are less likely to receive routine screening services for cardiovascular disease. A nationwide household survey in 1997 found that adults who had been without health insurance for one year or longer were less likely than insured adults to have received recommended hypertension screening within the previous two years (80 percent compared with 94 percent) or cholesterol screening (60 percent compared with 82 percent) (Ayanian et al., 2000). Adults who were uninsured for less than one year received these screening services at rates intermediate between those for long-term uninsured and insured adults.

Health insurance coverage is associated with better blood pressure control for lower-income persons with hypertension, according to two studies, one prospective and experimental and the other a longitudinal analysis of a cohort of patients that either lost or maintained Medicaid coverage. The prospective study, the RAND Health Insurance Experiment, found that for patients with diagnosed hypertension, patients in the plan without any cost sharing had significantly lower blood pressure than those in health plans with any form of cost sharing (an overall difference of 1.9 mm Hg) (Keeler et al., 1985). A much greater effect of cost sharing on average blood pressure was found for low-income patients than for high-income patients (3.5 mm Hg. versus 1.1 mm Hg.). Patients in the plan without cost sharing also had greater compliance with drug and behavioral therapies. These differences were attributed to more frequent contact with health providers in the free care plan (Keeler et al., 1985). 6

In the longitudinal analysis, Lurie and colleagues (1984, 1986) followed a cohort of patients at a university ambulatory care clinic for one year after some lost their Medi-Cal coverage consequent to a state policy change. At six months after loss of coverage and again at one year, hypertensive patients who lost coverage had significantly worse blood pressure than did those who remained covered by MediCal, with an average increase in diastolic blood pressure of 6 mm Hg compared with a decrease in the insured control group of 3 mm Hg after a full year (Lurie et al., 1984, 1986). The percentage of patients with diastolic blood pressure greater than 100 mm Hg increased in the group that lost coverage from 3 percent at baseline to 31 percent at six months, and then declined to 19 percent at one year, while the proportion with diastolic blood pressure > 100 mm Hg in the continuously covered control group did not change significantly over the year (Lurie et al., 1986).

Deficits in the care of uninsured persons with hypertension place them at risk of complications and deterioration in their condition. The 1987 National Medical Expenditures Survey afforded an in-depth examination of the use of antihypertensive medications by health insurance status. Uninsured persons younger than 65 who had hypertension were less likely than either those with private insurance or Medicaid to have any antihypertensive medication therapy (ORs = 0.62 and 0.44, respectively) (Huttin et al., 2000). 7 An analysis of the third round of the National Health and Nutrition Examination Survey (NHANES), with data on 40,000 respondents for the period 1988–1994, found that 22 percent of uninsured adults under age 65 with diagnosed hypertension had gone for more than one year without a blood pressure check, compared to 10 percent of insured adults with hypertension (Fish-Parcham, 2001). While 75 percent of insured adults under 65 who had ever been diagnosed with high blood pressure and been told to take medication for it were in fact taking blood pressure medication, only 58 percent of their uninsured counterparts who had been advised to take medication were doing so. Among those adults under 65 who had been advised to take cholesterol-lowering medication, 43 percent of those without insurance failed to take such medication, compared to 29 percent among those with health insurance who did not comply with this advice (Fish-Parcham, 2001).

A study by Shea and colleagues (1992a, 1992b) of patients presenting to two New York hospital emergency departments between 1989 and 1991 found that uninsured patients were more likely to have severe, uncontrolled hypertension than were sociodemographically similar patients with any health insurance (OR = 2.2), while patients without a regular source of care had an even greater risk of severe and uncontrolled disease (OR = 4.4). When insurance status, having a regular source of care, and complying with a therapeutic regimen were all included in the analysis, the odds ratio for being uninsured was no longer statistically significant (OR = 1.9, CI: 0.8–4.6). This result is not surprising, given the strong association between having health insurance and having a regular source of care.

Finding: Uninsured persons with diabetes are less likely to receive recommended services. Lacking health insurance for longer periods increases the risk of inadequate care for this condition and can lead to uncontrolled blood sugar levels, which, over time, put diabetics at risk for additional chronic disease and disability.

Despite the demanding and costly care regimen that persons with diabetes face, adults with diabetes are almost as likely to lack health insurance as those without this disease. Of diabetic adults under age 65, 12 percent are uninsured compared with 15 percent of the comparable general population (Harris, 1999). Persons with diabetes who are uninsured are less likely to receive the professionally recommended standard of care than are those who have health insurance ( Box 3.6 ). One result of not receiving appropriate care may be uncontrolled blood sugar levels, which puts diabetics at increased risk of hospitalization for either hyper- or hypoglycemia, in addition to increasing the likelihood of comorbidities and disabilities (Palta et al., 1997).

Diabetes. Diabetes mellitus is a prevalent chronic disease that has been increasing in the U.S. population by 5–6 percent each year during the past decade. Approximately 800,000 new cases are diagnosed each year. More than 16 million Americans (more...)

Based on a 1994 survey, among adults diagnosed with diabetes who did not use insulin, those without health insurance were less likely than those with any kind of coverage to self-monitor blood glucose (OR = 0.5) or, within the past year, to have had their feet examined (OR = 0.4), or a dilated eye exam (OR = 0.5) (Beckles et al., 1998). 8 Persons with diabetes who used insulin and were uninsured were also less likely than those with health insurance to have had a foot examination (OR = 0.25) or a dilated eye examination (OR = 0.34) (Beckles et al., 1998).

A later analysis, using 1998 data from the same annual survey, found that 25 percent of adults younger than 65 who had diabetes and were uninsured for a year or more had not had a routine checkup within the past two years, compared with 7 percent of diabetics who were uninsured for less than a year and 5 percent of diabetics with health insurance (Ayanian et al., 2000). Adjusting results for the demographic characteristics of the national population, persons with diabetes who were uninsured for a year or longer were significantly less likely to have had a foot examination, a dilated eye examination, a cholesterol measurement, or a flu shot than were insured diabetics ( Figure 3.1 ) (Ayanian et al., 2000).

Diabetes management among insured and uninsured adults, ages 18–64. NOTE: Proportions adjusted to demographic characteristics of study cohort.

End-Stage Renal Disease

Finding: Uninsured patients with end-stage renal disease begin dialysis at a later stage of disease than do insured patients and have poorer clinical measures of their condition at the time they begin dialysis.

Insurance status affects the timing and quality of care ( Box 3.7 ) and may contribute to the longevity of dialysis patients, which is substantially lower than that of others of the same age (Obrador et al., 1999). The clinical goals for patients with kidney disease are to slow the progression of renal failure, manage complications, and prevent or manage comorbidities effectively. Although professional consensus about when dialysis should begin is not complete, there is agreement that the point in the progression of the disease at which dialysis begins affects patient outcomes (Kausz et al., 2000).

End-Stage Renal Disease. In 2000, 90,000 people in the United States developed end-stage renal disease (kidney failure). Dialysis and transplantation are the two standard treatments. Approximately 300,000 patients are on dialysis and 80,000 have received (more...)

The Medicare ESRD program maintains extensive clinical and sociodemographic information on all dialysis patients, including information on patient health insurance status before beginning dialysis. This database provides opportunities to analyze the health care experience of all Americans who eventually develop ESRD, rather than just a sample of the population. One study that used this database analyzed the characteristics of 155,000 chronic dialysis patients who entered dialysis over a 27-month period between 1995 and 1997 (Obrador et al., 1999). This study found that uninsured patients were sicker at initiation of dialysis and less likely to have received erythropoietin (EPO) therapy than patients with any kind of insurance pre-ESRD. Uninsured patients also had an increased likelihood of hypoalbuminemia than those who had previously been privately insured (OR = 1.37) and a greater likelihood of low hematocrit (<28 percent) 9 than the privately insured (OR = 1.34), after controlling for patients' sociodemographic and clinical characteristics, including comorbidities. Uninsured patients were also less likely than privately insured patients to have received EPO prior to dialysis (OR = 0.49) (Obrador et al., 1999). A second study based on the same data set found that patients without insurance were more likely to begin dialysis late 10 than were patients with any form of insurance (OR = 1.55) (Kausz et al., 2000).

Human Immunodeficiency Virus (HIV) Infection

Finding: Uninsured adults with HIV infection are less likely to receive highly effective medications that have been shown to improve survival.

A strong body of research about HIV infection confirms the findings of the general literature on insurance status and access to and use of services: uninsured adults diagnosed with HIV face greater delays in care than those with health insurance. They are less likely to receive regular care and drug therapy and are more likely to go without needed care than patients with any kind of coverage (Cunningham et al., 1995, 1999; Katz et al., 1995; Shapiro et al., 1999).

BOX 3.8 HIV Infection

• As of the beginning of 2000, the Centers for Disease Control and Prevention estimated that about 800,000 to 900,000 people were living with HIV infection or AIDS in the United States (CDC, 2001a).
• In each of the years 1997, 1998, and 1999, between 40,000 and 50,000 new cases of AIDS were reported.
• By 1996, combination antiretroviral therapy including protease inhibitors and nonnucleoside reverse transcriptase inhibitors, referred to as highly active antiretroviral therapies were becoming established as the treatment of choice for HIV infection (Carpenter et al., 1996). Largely as a result of these therapies, deaths among persons with AIDS dropped for the first time between 1996 and 1997 (by 42 percent) and declined 8 percent between 1998 and 1999 (CDC, 2001a).
• About half of all adults with HIV infection see a provider at least once every six months (Bozzette et al., 1998).
• Studies of HIV infection and health insurance examine a variety of health-related outcomes: general measures of access and utilization such as routine care visits and emergency department visits without hospitalization, delays between diagnosis and initiation of therapy, use of recommended drug therapies, and clinical outcomes such as CD4 lymphocyte counts.

A number of analyses have been based on national, longitudinal surveys evaluating access to care for persons with HIV infection (Niemcryk et al., 1998; Joyce et al., 1999; Shapiro et al., 1999; Andersen et al., 2000; Cunningham et al., 1999, 2000; Turner et al., 2000, Goldman et al., 2001). 11 These surveys allow assessment of the relationship between health insurance and access to care, use of services, receipt and timeliness of recommended therapies, and mortality as related to health insurance status. The research based on one of these surveys, the HIV Cost and Services Utilization Study (HCSUS), represents some of the most carefully designed studies of access to care and receipt of recommended therapies for specific conditions. In addition, there are several smaller, local studies based on hospital records or patient surveys (Katz et al., 1992, 1995; Bennett et al., 1995; Cunningham et al., 1995, 1996; Palacio et al, 1999; Sorvillo et al., 1999).

Access to a Regular Source of Care

Several studies suggest that the positive effects of health insurance for HIV-infected adults are achieved through the mechanism of having a regular source of care. Sorvillo and colleagues (1999) surveyed 339 HIV-positive adults in Los Angeles county in 1996–1997, and found that two-thirds of insured patients used protease inhibitors (PIs), while just half of uninsured patients were using them. When the site of care (private clinic, HMO, or public clinic) was included in a multivariable analysis, insurance status was no longer significantly related to receipt of PIs because of the concentration of uninsured patients in public clinics, which were less likely to prescribe PIs, especially at the beginning of the study period (Sorvillo et al., 1999).

Uninsured patients appear to face greater delays in beginning care following a diagnosis of HIV infection. In bivariate analysis of HCSUS data, uninsured patients were significantly more likely to have their first office visit more than three months after diagnosis with HIV than were privately insured patients (37 percent of uninsured patients had delays compared to 25 percent of privately insured patients in 1993; by 1995, those patients with delays decreased to 22 percent of uninsured patients and 14 percent of privately insured) (Turner et al., 2000). However, in a multivariable analysis, being uninsured was no longer a significant predictor of late initiation, while not having a regular source of care remained an important predictor (Turner et al., 2000).

Findings regarding emergency department (ED) use and hospitalization have changed over time. The most recent analysis, based on HCSUS, finds greater use of EDs, without hospitalization, and hospitalization more frequently than every six months for uninsured HIV patients (Shapiro et al., 1999), suggesting poorer access to other kinds of outpatient care. Studies based on earlier data report that uninsured patients had lower use of emergency rooms and hospitalization than either publicly or privately insured patients (Mor et al., 1992; Fleishman and Mor, 1993; Niemcryk et al., 1998; Joyce et al., 1999), suggesting poorer access even at high levels of acuity.

Receipt of Drug Therapies

Adults with HIV infection are more likely to receive effective drug therapies and to receive them earlier in the course of disease if they have health insurance. In an HCSUS analysis with extensive adjustments for sociodemographic and clinical factors, those without health insurance were much less likely to have ever received antiretroviral therapy (OR = 0.35) (Shapiro et al., 1999). Waiting times from diagnosis to the start of therapy with either PIs or nonnucleoside reverse transcriptase inhibitors, were 9.4 months for the privately insured, 12.4 months for Medicaid enrollees, and 13.9 months for uninsured patients (Shapiro et al., 2000).

Overall, many HIV-infected patients abandon recommended drug therapy over time. However, uninsured patients are more likely to stop drug therapy than are those with coverage. At the second follow-up interview of HCSUS respondents in 1997–1998, only half (53 percent) of all HIV-positive patients in care were receiving the recommended combination drug therapy, highly active antiretroviral therapy (HAART), although 71 percent had received HAART at some time in their treatment history (Cunningham et al., 2000). Uninsured patients were significantly less likely than privately insured patients with indemnity coverage (OR = 0.71) to be receiving HAART at the time of follow-up, indicating less appropriate care for uninsured patients with this disease (Cunningham et al., 2000).

Clinical Outcomes and Mortality

Studies of clinical outcomes for HIV patients present an evolving picture of both the efficacy of treatments and the impact of health insurance. A relatively early study of patients hospitalized with Pneumocystis carinii pneumonia (1987– 1990) found that uninsured patients had a higher in-hospital mortality rate than did those with private insurance (OR = 1.49), and Medicaid patients had an even higher in-hospital mortality, relative to private patients (OR = 2.1) (Bennett et al., 1995). Another early and small study (96 patients in one university clinic) found that patients with private insurance had significantly lower CD4 lymphocyte counts (a worse outcome) than either uninsured or Medicaid patients (who had the highest counts), when first treated at the clinic (Katz et al., 1992). The authors hypothesize that some relatively healthy patients with private health insurance coverage may have been reluctant to use it and thus reported their status as uninsured.

More recently, an analysis based on HCSUS examined the mortality experience of insured and uninsured HIV-infected adults and found that having health insurance of any kind reduced the risk of dying within six months of being surveyed between 71 and 85 percent, when severity of illness (measured by CD4 lymphocyte count) and sociodemographic characteristics were controlled (Goldman et al., 2001). The greater reduction in mortality risk (85 percent) was estimated for a surviving subset (2,466 participants) of the original 1996 sample of 2,864 participants a year later, when HAART was in wider use and was reducing mortality among HIV patients who used it. This impact of health insurance on mortality for HIV-infected adults within a short follow-up period, six months, demonstrates how sensitive health outcomes can be to coverage when it facilitates receipt of effective therapy.

Mental Illness

Finding: Health insurance that covers any mental health treatment is associated with the receipt of mental health care and with care consistent with clinical practice guidelines from both general medical and specialty mental health providers.

Mental disorders or illnesses are health conditions that are characterized by changes in thinking, mood, or behavior. They are often chronic conditions but may also occur as single or infrequent episodes over a lifetime. Mental illnesses represent a major source of disability in the United States that is often underestimated by the public and health care professionals alike (USDHHS, 2000). In industrialized economies, mental illness is equivalent to heart disease and cancer in terms of its impact on disability (Murray and Lopez, 1996).

Despite the differential treatment of mental health services in both public and private insurance plans, the studies reviewed by the Committee document a positive association between health insurance coverage and more appropriate care for mental illnesses ( Box 3.9 ). Health insurance plans and programs historically have excluded services related to treatment for mental illness, strictly limited coverage of mental health services, and administered mental health benefits separately from other kinds of medical care. Thus, studies that attempt to measure the effects of health insurance status on health care and outcomes for mental illnesses may be affected by the diversity of health insurance benefits and of cost sharing and administrative requirements for these services and conditions. Variability in benefits among health insurance plans and types of insurance complicates the interpretation of all observational studies of health insurance effects but poses a particular problem vis–a–vis mental health. (See the discussion of measurement bias in Chapter 2 .)

Mental Illness. About 38 million people ages 18 and older are estimated to have a single mental disorder of any severity or both a mental and an addictive disorder in a given year (Narrow et al., 2002). The most common conditions fall into the broad categories (more...)

The use of mental health services in both the general and specialty mental health sectors by adults is positively associated with health insurance coverage (Cooper-Patrick et al., 1999; Wang et al., 2000; Young et al., 2001). Between 1987 and 1997, the overall rate of treatment for depression among American adults under age 65 tripled from 1 person per 100 to 3.2 persons per 100, yet the treatment rate among those without health insurance was half that of the overall population rate in 1997, 1.5 persons treated per 100 population (Olfson et al., 2002). A longitudinal, community-based study in Baltimore, Maryland, between 1981 and 1996 documented increased use of mental health services over this period (Cooper-Patrick et al., 1999). Analyzing the experience of African Americans and whites separately, the authors found that for African Americans specifically, this increase was achieved predominantly with services provided in the general medical sector. For both African Americans and whites, being uninsured reduced the likelihood of receiving any mental health services.

At the same time, insurance coverage for adults with mental illness is less stable than average for those without this condition (Sturm and Wells, 2000; Rabinowitz et al., 2001). In a recent (1998) follow-up survey of participants in the Community Tracking Study, those who reported having symptoms of mental disorders were found to be more likely to lose coverage within a year following their diagnosis than those without a mental disorder (Sturm and Wells, 2000). As discussed below, those with severe mental illness also experience transitions in insurance coverage, frequently ending up with public program coverage (Rabinowitz et al., 2001).

The findings reported below are grouped into those for depression and anxiety disorders and those for severe mental illnesses. Depression and anxiety disorders are often treatable in the general medical sector and primarily require outpatient services. Severe mental illnesses (schizophrenia, other psychoses, and bipolar depression) require the attention of specialty mental health professionals and may require inpatient and other forms of more extensive services (e.g., partial or day hospitalization). Public health insurance, both Medicare and Medicaid, is an important source of coverage for specialty mental health services for those disabled by severe mental illness (SMI).

Depression and Anxiety Disorders

Receipt of appropriate (guideline-concordant) care for depression is associated with improved functional outcomes at two years (Sturm and Wells, 1995). Health insurance coverage specifically for mental health services is associated with an increased likelihood of receiving such care. Two studies support this claim.

The first, a nationally representative study of three prevalent disorders— depression, panic disorder, and generalized anxiety disorder—investigated the contribution of insurance coverage and health care utilization to guideline-con-cordant treatment (Wang et al., 2000). Mental health diagnoses were determined in a structured interview using a well-defined operational definition of mental health care over the previous 12 months. Treatment criteria included the combination of a prescription medication for depression or anxiety from a general medical doctor or a psychiatrist in addition to at least four visits to the same type of provider or, where medication was not prescribed, a minimum of eight visits to either a psychiatrist or a mental health specialist (Wang et al., 2000). A multivariable analysis estimated the effects of sociodemographic characteristics, various measures of clinical status including a measure of mental illness severity, insurance coverage for mental health visits, number and reasons for use of general medical services, other medications, and alternative therapies. Patients diagnosed with depression, panic disorder, or a generalized anxiety disorder who had no health insurance coverage for mental health visits were less likely to receive any mental health services (OR = 0.43). They were also less likely to receive guideline-concordant care in the general medical sector (OR = 0.24) or in the mental health treatment sector (OR = 0.36) (Wang et al., 2000).

A second study of adults with a probable 12-month diagnosis of depression or anxiety examined factors associated with receipt of appropriate care (psychiatric medication and counseling) (Young et al, 2001): 1,636 respondents were identified as having one or more depressive or anxiety disorders based on a structured diagnostic interview. Respondents with a depressive or anxiety disorder who had more education and a greater number of medical disorders were more likely to have had contact with providers than those with less education and fewer medical conditions. Those with no health insurance were less likely to have had any provider contact than were those with any form of health insurance (OR = 0.46). However, for those receiving any care, insurance status was not related to receipt of appropriate care (Young et al., 2001). These findings suggest that health insurance alone may not ensure appropriate mental health care.

Severe Mental Illness

Uninsured adults with severe mental illnesses are less likely to receive appropriate care than are those with coverage and may experience delays in receiving services until they gain public insurance.

In a study using the same sample and survey as that used by Young and colleagues, McAlpine and Mechanic (2001) investigated the association of current insurance coverage and specialty mental health utilization within the past 12 months (i.e., visits to a psychiatrist or psychologist, hospital admission, or emergency room visit for an emotional or substance use problem) for SMI. Two diagnostic indices, including a global measure of mental health, measured the need for care. Potential confounding factors such as physical symptoms and degree of dangerousness and disruptiveness were also measured. One in five respondents identified with an SMI was uninsured. Among persons with SMI, those without health insurance were far less likely to use specialty mental health services than those with Medicare or Medicaid (OR = 0.17) (McAlpine and Mechanic, 2000).

Individuals with SMIs typically lack insurance at the time of hospitalization (Rabinowitz et al., 2001). An important question regarding insurance coverage in this patient population is whether a first hospitalization for SMI results in a change in insurance status and whether such a change influences subsequent mental health care. Rabinowitz and colleagues followed the progress of 443 individuals enrolled in a county mental health project to determine whether changes in coverage followed first admission for psychosis and the association between type of insurance coverage and future care. Overall, the proportion of patients with no insurance 24 months after hospitalization decreased from 42 percent at baseline to 21 percent as a result of enrollment in public insurance programs. Men were more likely to remain uninsured than were women. The total number of days of care received (inpatient, outpatient, day hospital) was significantly higher for the publicly insured group compared to both those with private insurance and those with no insurance during the first 6 months after initial hospitalization and over the entire 24-month period. Uninsured patients with SMI were much less likely to receive outpatient care after hospitalization than patients with Medicaid or Medicare (OR = 0.24) and also less likely than those with private health insurance to receive outpatient care subsequent to hospitalization (OR = 0.56) (Rabinowitz et al., 2001).

An earlier study using the same data also reported an association between health insurance and receipt of mental health services prior to a first admission for psychotic disorder (Rabinowitz et al., 1998). Forty-four percent of patients (n = 525) were uninsured at first admission. Uninsured patients were less likely than those with private insurance to have had

• any mental health treatment prior to admission (OR = 0.53),
• specific psychotherapeutic contact (OR = 0.43),
• voluntary admission (OR = 0.56),
• less than three months between onset of psychosis and admission (OR = 0.56)

and were less likely to have been admitted to a community (versus public) hospital (OR = 0.14) (Rabinowitz et al., 1998). Uninsured patients were also less likely than those with either Medicaid or Medicare to have received antipsychotic medication (OR = 0.4), had voluntary admission (OR = 0.53), and be admitted to a community hospital (OR = 0.33).

• HOSPITAL-BASED CARE

Finding: Uninsured patients who are hospitalized for a range of conditions experience higher rates of death in the hospital, receive fewer services, and are more likely to experience an adverse medical event due to negligence than are insured patients.

Americans assume and expect that hospital-based care for serious and emergency conditions is available to everyone, regardless of health insurance coverage, while recognizing that uninsured patients may be limited to treatment at public or otherwise designated “safety-net” hospitals (IOM, 2001a). Professional and institutional standards of practice grounded in ethics, law, and licensure dictate that the care received by all patients, regardless of financial or insurance status, be of equal and high quality. Yet studies of hospital-based care conducted over the past two decades have documented differences in the services received by insured and uninsured patients, differences in the quality of their care (sometimes but not always related to the site of care), and differences in patient outcomes such as in-hospital mortality rates. 12

One of the most comprehensive of these studies of hospitalization analyzed more than 592,000 hospital discharge abstracts in 1987 (Hadley et al., 1991). The authors report that for adults ages 18–65, uninsured hospital inpatients had a significantly higher risk of dying in the hospital than their privately insured counterparts in 8 of 12 age–sex–race-specific population cohorts (relative risks ranged from 1.1 for black women ages 50–64 to 3.2 for black men ages 35–49). This analysis adjusted for patient condition on admission to the hospital. Uninsured patients were also less likely to receive endoscopic procedures in the hospital than privately insured patients, and when they did receive these diagnostic services, the resultant pathology reports were more likely to be abnormal (OR = 1.56) (Hadley et al., 1991).

This study by Hadley and colleagues also examined the relative resource use (length of stay) of uninsured hospital patients compared to privately insured patients and found that for conditions that afford high discretion in treatment decisions (e.g., tonsillitis, bronchitis, hernia), uninsured patients had significantly shorter lengths of stay (Hadley et al., 1991). However, for diagnoses that afford little discretion in treatment (e.g., gastrointestinal hemorrhage, congestive heart failure), lengths of stay were not significantly different for uninsured and privately insured patients, although uninsured patients tended to have shorter stays. This underscores the possibility that when uninsured patients are found to receive fewer services than insured patients, it may be the result of overtreatment of patients with insurance, rather than undertreatment of those without coverage.

In addition to differences in the resources devoted to the care of insured and uninsured patients, the quality of the care provided may differ. One study of more than 30,000 hospital medical records in 51 hospitals in New York State for 1984 found that the proportion of adverse medical events due to negligence was substantially greater among patients without health insurance than among privately insured patients (OR = 2.35), while the experience of Medicaid patients did not differ significantly from that of the privately insured population (Burstin et al., 1992). This increased risk for uninsured patients was attributable only in part to receiving care more frequently in emergency departments, which generally were found to have higher rates of adverse events.

Because most studies of hospital-based care and outcomes are observational, including only those who literally “show up” for care, and because appropriateness criteria are not available for many conditions, some of the strongest research on health insurance effects involves studies of specific conditions. Studies of certain conditions are less likely to be compromised by nonrandom or unrepresentative samples (selection bias) simply because a larger proportion of the population of interest—namely, acutely ill adults—is likely to be captured in the hospital-based study population. Furthermore, condition-specific studies are more likely to include evidence-based criteria for judging the appropriateness of care.

The following two sections consider research that has examined the effect of health insurance on care and outcomes for patients with (1) emergency conditions and traumatic injuries and (2) cardiovascular disease. For both categories, selection bias among those reaching treatment is minimized, and appropriateness guidelines and outcomes criteria (e.g., mortality) are definitive. Traumatic injuries (specifically automobile accidents), for example, reduce some of the unmeasured differences in propensity to seek care between insured and uninsured patients (Doyle, 2001). Another area of hospital-based services for which there is sufficient professional consensus about appropriate treatment is the use of angiography and revascularization procedures following acute myocardial infarction (AMI) or heart attack, at least for a subset of patients with severe coronary artery disease. 13

Emergency and Trauma Care

Finding: Uninsured persons with traumatic injuries are less likely to be admitted to the hospital, receive fewer services when admitted, and are more likely to die than insured trauma victims.

Two studies based on large, statewide data sets have found substantial and significant differences in the risk of dying for insured and uninsured trauma patients ( Box 3.10 ) who were admitted to hospitals as emergencies. Doyle (2001) analyzed more than 10,000 police reports of auto accidents linked to hospital records maintained by Wisconsin over 1992–1997 to ascertain the care received and the mortality of insured and uninsured crash victims. After controlling for personal, crash, and hospital characteristics, it was found that uninsured accident victims received 20 percent less care, as measured by hospital charges and length of stay, and had a 37 percent higher mortality rate than did privately insured accident victims (5.2 percent versus 3.8 percent, respectively) (Doyle, 2001). The authors conclude that these differences are attributable to provider response to insurance status because extensive patient characteristics were accounted for in the analysis and because unmeasured patient characteristics that might influence these outcomes were unlikely to be related to patients' health insurance status.

Trauma. Throughout the United States in 1997, approximately 34.4 million episodes of injury and poisoning received medical attention and 40.9 million injuries and poisonings were reported as a result (Warner et al., 2000). For injury-related deaths, 43 (more...)

Haas and Goldman (1994) evaluated the treatment experience and mortality of more than 15,000 insured and uninsured trauma patients admitted to hospitals on an emergency basis in Massachusetts in 1990. Adjusting the data for injury severity and comorbidities as well as for age, sex, and race, the authors found that uninsured trauma patients received less care and had higher in-hospital mortality than did patients with private insurance or Medicaid. Uninsured patients were just as likely to receive care in an intensive care unit (ICU) as privately insured trauma patients but were less likely to undergo an operative procedure (OR = 0.68) or to receive physical therapy (OR = 0.61). Uninsured patients were much more likely than privately insured patients to die in the hospital (OR = 2.15) (Haas and Goldman, 1994). The differences in services and mortality experience between Medicaid and privately insured patients were small and were not statistically significant.

Other studies of emergency department use and admissions and care for traumatic injuries shed some light on patient behavior and institutional responses related to health insurance status. Both lacking health insurance and not having a regular source of care have been found in surveys of patients who eventually do arrive at an ED to be related to delays in seeking care (Ell et al., 1994; Rucker et al., 2001). Braveman and colleagues (1994) examined hospital discharge records of more than 91,000 adults diagnosed with acute appendicitis in California hospitals between 1984 and 1989. They found that the risk of a ruptured appendix was 50 percent higher for both uninsured and Medicaid patients, than for privately insured patients in prepaid plans, in an analysis that controlled for age, sex, race, psychiatric diagnoses, diabetes, and hospital characteristics. Admission to a public hospital also was associated with rupture, as were diagnoses of psychiatric illness or diabetes (Braveman et al., 1994). The authors hypothesized that both Medicaid and uninsured patients incurred avoidable delays before seeking care for appendicitis.

Three separate studies that analyzed Medicaid and uninsured trauma patients together report mixed findings regarding patient outcomes and hospital care. Rhee and colleagues (1997) examined patient information for more than 2,800 persons hospitalized at a Level 1 trauma center after a motor vehicle crash in Seattle, Washington, between 1990 and 1993. 14 This study found no significant differences in mortality, hospital charges, or length of stay (LOS) between privately insured patients and those who either had Medicaid coverage or were uninsured, except for patients who ultimately were transferred to a long-term care or rehabilitation facility. In the case of patients awaiting transfer, those with Medicaid or no insurance had an adjusted LOS that was 11 percent longer than privately insured patients (Rhee et al., 1997). The authors speculate that the similarity in treatment and outcomes for patients of different insurance status could be due to the mission of the public, Level 1 trauma center to which they were admitted, which was to serve the entire state population needing that level of care and act as a provider of last resort for uninsured patients. Because this study did not differentiate results for Medicaid and uninsured patients, it provides less information about outcomes for uninsured patients than studies that analyze these groups separately.

Uninsured trauma patients may also be treated differently from insured patients in interhospital transfer decisions. Using Washington State trauma registry information, Nathens and colleagues (2001) identified 2,008 trauma patients between 16 and 64 years of age injured in King County (Seattle) and originally transported to one of seven Level 3 or 4 trauma centers in the county between 1995 and 1999. Adjusting for age, sex, type of injury, and injury severity, they looked at independent predictors of transfer to the Level 1 trauma center in the county—a public, safety-net hospital, and estimated that patients who either had Medicaid or were uninsured were more than twice as likely to be transferred to the higher level facility than were privately insured patients (OR = 2.4) and that many of these transferred patients had low injury severity scores (ISS). 15 The authors conclude that this “payer-based triage” may undermine the effectiveness of Level 1 trauma centers in serving the more critically injured patients by diverting resources to patients who could have been treated appropriately in their original hospital (Nathens et al., 2001).

Finally, the differences found between uninsured and insured patients in highly discretionary cases may reflect overtreatment of those with health insurance rather than undertreatment of uninsured patients. Svenson and Spurlock (2001) evaluated the experience of more than 8,500 patients with head injuries treated in four Kentucky hospitals between 1995 and 1997. For those with less severe head injuries (lacerations, contusion, or concussion), uninsured patients were substantially less likely than privately insured patients to be admitted to the hospital (OR = 0.14 for laceration, 0.38 for contusion or concussion). The likelihood of admission for Medicaid was also substantially lower than for privately insured patients, but not as low as for uninsured patients (ORs = 0.33 and 0.45, respectively). Little difference was found in hospital admissions for more severe head injuries among patients with different insurance status. The authors were unable to determine whether the differences in admissions for less severe head trauma are due to undertreatment of uninsured and Medicaid patients or overtreatment of privately insured patients (Svenson and Spurlock, 2001).

Finding: Uninsured patients with acute cardiovascular disease are less likely to be admitted to a hospital that performs angiography or revascularization procedures, are less likely to receive these diagnostic and treatment procedures, and are more likely to die in the short term.

Finding: Health insurance reduces the disparity in receipt of these services by members of racial and ethnic minority groups.

Health insurance is positively associated with receipt of hospital-based treatments for cardiovascular disease (specifically, coronary artery disease) and with lower patient mortality ( Box 3.11 ). One meta-analysis has credited medical advances in the treatment of cardiovascular disease, including hospital-based care following AMI, with roughly half of the reduction in post-AMI mortality between 1975 and 1995 (with a range of 20 to 85 percent) (Cutler et al., 1998). Some of the most recent studies have used appropriateness criteria to identify when a given procedure is considered necessary according to professional consensus, reducing the chances that differences in rates between uninsured and insured patients are a result of overtreatment of the insured population (i.e., Sada et al.,1998; Leape et al., 1999).

In 2001, an estimated 1.1 million Americans suffered a diagnosed heart attack. An estimated 7.3 million Americans have a history of AMI (American Heart Association, 2001). During 1998, coronary heart disease accounted for about 460,000 deaths; AMI was (more...)

Five studies that examined the mortality experience of patients hospitalized for cardiovascular disease (including AMI, angina, and chest pain) reported higher in-hospital or 30-day posthospitalization mortality for uninsured patients (Young and Cohen, 1991; Blustein et al., 1995; Kreindel et al., 1997; Sada et al., 1998; Canto et al., 2000).

The first study, of about 5,000 patients admitted on an emergency basis for AMI in 1987, found that uninsured patients were more likely to die within 30 days of admission than privately insured patients (OR = 1.5) (Young and Cohen, 1991). In a second study, Blustein and colleagues (1995) examined records for 5,800 patients under 65 who were admitted to California hospitals for AMI in 1991 and found that uninsured patients were more likely to die in the hospital than privately insured patients (OR = 1.9) and still had an increased risk of dying after adjusting for receipt of a revascularization procedure (OR = 1.7). Finally, a study in a single Massachusetts community of 3,700 patients hospitalized for AMI between 1986 and 1993 reported that uninsured patients had a slight, but statistically insignificant greater in-hospital mortality than privately insured patients (OR = 1.2, CI: 0.6–2.4) (Kreindel et al., 1997).

Two larger studies that used more recent data (1994–1996) from the National Registry of Myocardial Infarction reported higher in-hospital mortality for uninsured than for privately insured patients. In the first, Sada and colleagues (1998) reviewed records for 17,600 patients under age 65 who were admitted to hospital for AMI and found that uninsured patients had an in-hospital mortality rate of 5.4 percent, compared with 3.8 percent for private FFS patients and 3.9 percent for private HMO patients. Medicaid patients had the highest in-hospital mortality rate, 8.9 percent. In a model that adjusted for demographic and clinical factors, the likelihood of uninsured patients dying in the hospital was still higher but was not statistically significantly different from that of privately insured patients (OR = 1.2, CI: 0.8–1.6) (Sada et al, 1998). The second national study examined records for more than 332,000 patients admitted with AMI and found that after adjusting for demographics, prior disease history, and clinical characteristics, uninsured patients were more likely to die in the hospital than privately insured FFS patients (OR = 1.29) (Canto et al., 2000). The mortality experience of Medicaid patients was the same as that of uninsured patients.

Only one study, a review of hospital records of 1,556 patients undergoing coronary artery bypass graft surgery in a single Louisiana teaching hospital, found that uninsured patients had better long-term survival than did insured patients (Mancini et al., 2001). However, this study did not control for age or characteristics of the patients. The average age of uninsured patients at the time of surgery was 55, and of insured patients, 65 years. Furthermore, only 7 percent of the insured study population had private insurance, so the population was not representative of the insured population at large.

Coronary Procedures

The body of research on the use of specific procedures to diagnose and treat cardiovascular disease as a function of the insurance status of the patient consistently reports differences in utilization, with uninsured patients generally less likely to receive coronary angiography, CABG, or percutaneous transluminal coronary angioplasty (PTCA) than privately insured patients (Young and Cohen, 1991; Blustein et al., 1995; Kuykendall et al., 1995; Sada et al., 1998; Leape et al., 1999; Canto et al., 2000; Daumit et al., 2000). However, only some of these studies applied appropriateness criteria to identify cases in which the use of these procedures was considered nondiscretionary or necessary. In the studies that examined overall utilization rates, the differences found by insurance status could be attributed to overutilization as well as underutilization.

Angiography (cardiac catheterization) is an invasive diagnostic procedure that provides information to guide decisions about subsequent treatment options, including revascularization procedures. Sada and colleagues (1998) applied the criteria of the American College of Cardiology and American Heart Association Joint Task Force to a national data set of 17,600 myocardial infarction patients under 65 to identify nondiscretionary angiography for revascularization candidates considered to be at high risk. They estimated that in hospitals providing these cardiac procedures, patients with private FFS coverage who were deemed high-risk and for whom angiography was nondiscretionary were more likely than similarly high-risk uninsured patients or Medicaid patients to receive angiography. Among high-risk FFS patients, 84 percent received this service compared to 73 percent of high-risk uninsured patients and 60 percent of similar Medicaid patients (Sada et al., 1998).

Revascularization procedures (either CABG or PTCA) following a heart attack are also more likely to be performed on insured than uninsured patients. In two studies, uninsured patients were less likely to receive revascularization (either CABG or PTCA) than privately insured FFS patients (OR = 0.6 in the 1991 study and 0.8 in the 2000 study) (Young and Cohen, 1991; Canto et al., 2000). Blustein and colleagues (1995) and Kuykendall and colleagues (1995) reported similar comparative findings regarding the revascularization of uninsured and privately insured patients (ORs in these studies ranged from 0.4 to 0.6).

InterHospital Transfers to Receive Services. For patients with AMI, health insurance facilitates access to hospitals that perform angiography and revascularization, whether admission is initial or by means of an interhospital transfer (Blustein et al., 1995; Canto et al., 1999; Leape et al., 1999).

In a study of California hospital admissions for AMI, Blustein and colleagues (1995) found that uninsured patients were less likely than privately insured patients to be admitted initially to a hospital that offered revascularization and much less likely to be transferred if admitted initially to one that did not (ORs = 0.71 and 0.42, respectively).

Leape and colleagues (1999) reviewed 631 records for patients who had received angiography and subsequently met expert panel criteria for necessary revascularization. Overall, 74 percent of patients meeting these criteria received revascularization. Leape et al. found that in hospitals that also performed CABG and PTCA, there were no differences in rates of revascularization for patients with different insurance status. However, for patients initially hospitalized in facilities that did not perform CABG and PTCA, who required a transfer to another hospital to receive revascularization, the rates differed significantly by insurance status: 91 percent of Medicare patients, 82 percent of privately insured patients, 75 percent of Medicaid patients, and just 52 percent of uninsured patients received this indicated surgery (Leape et al., 1999).

Insurance Status and Racial and Gender Disparities. Health insurance has been shown to lessen disparities in the care for cardiovascular disease received by men compared to women and among members of racial and ethnic groups (Carlisle et al., 1997; Daumit et al., 1999, 2000).

An analysis of more than 100,000 hospital discharges with a principal diagnosis of cardiovascular disease in Los Angeles County between 1986 and 1988 revealed significant differences in rates of angiography, CABG, and PTCA between uninsured African-American and white patients but not between members of these ethnic groups who were privately insured (Carlisle et al., 1997). In a multivariate analysis that controlled for demographic and clinical characteristics and hospital procedure volume, the odds ratios for uninsured African Americans to receive one of these services compared with uninsured whites ranged from 0.33 to 0.5 (Carlisle et al., 1997).

A longitudinal study with a seven-year follow-up of a national random sample of patients who initially became eligible for the Medicare ESRD program in 1986 or 1987 found that once uninsured patients qualified for ESRD benefits, pronounced disparities by gender or race in their likelihood of receiving either angiography, CABG, or PTCA were eliminated (Daumit et al., 1999, 2000). In the period prior to qualifying for Medicare, uninsured African Americans were far less likely than uninsured whites to undergo a cardiac procedure (OR = 0.07) (Daumit et al., 1999). Uninsured women were also less likely than uninsured men to receive a cardiac procedure before qualifying for Medicare (OR = 0.4), and uninsured men were much less likely than men with private insurance to receive one (OR = 0.47) (Daumit et al., 2000). In the case of both race and gender, differences in the receipt of these cardiac procedures were eliminated after gaining Medicare ESRD coverage.

• GENERAL HEALTH OUTCOMES

Finding: Longitudinal population-based studies of the mortality of uninsured and privately insured adults reveal a higher risk of dying for those who were uninsured at baseline than for those who initially had private coverage.

Finding: Relatively short (one- to four-year) longitudinal studies document relatively greater decreases in general health status measures for uninsured adults and for those who lost insurance coverage during the period studied than for those with continuous coverage.

This chapter concludes with a review of the studies evaluating the overall health status and mortality experience of insured and uninsured populations. Assessments of general health outcomes such as self-reported health status and mortality or survival rates for uninsured adults under 65 compared to those with some form of health insurance (i.e., employment-sponsored, Medicaid, Medicare, individually purchased policies), present researchers with even greater challenges of analytic adjustment than those encountered in studies of specific health conditions. Not only might health insurance affect health status, but health status can affect health insurance status. Thus, it is difficult to interpret cross-sectional studies of health insurance and health status. However, several well-designed longitudinal studies with extensive analytic adjustments for covariates have found higher mortality and worse overall functional and health status among uninsured adults than among otherwise similar insured adults.

Two studies provide evidence that uninsured adults are more likely to die prematurely than are their privately insured counterparts.

Franks and colleagues (1993a) followed a national cohort of 4,700 adults age 25 or older for 13 to 17 years who, at the baseline interview, were either privately insured or uninsured. At the end of the follow-up period (1987), about twice as many participants who were uninsured at the time of the first interview had died as had those with private health insurance (18.4 percent compared with 9.6 percent). Controlling for sociodemographic characteristics, health examination findings, self-reported health status, and health behaviors, the risk of death for adults who initially were uninsured was 25 percent greater than for those who had private health insurance at the time of the initial interview (mortality hazard ratio = 1.25, CI: 1.00–1.55). The magnitude of this independent health insurance effect on mortality risk was comparable to that of being unemployed, to lacking a high school diploma, or to being in the lowest income category (Franks et al., 1993a). 16 Because insurance status was measured only at the initial interview and thus did not reflect the subjects' cumulative insurance experience over the 13–17 year follow-up period, the difference found in mortality between uninsured and privately insured persons most likely is an underestimate of differences in the mortality experience of those who are continuously uninsured and those who are continuously insured.

A study by Sorlie and colleagues (1994) tracked the mortality experience of 148,000 adults between 25 and 65 years of age until 1987, a two- to five-year follow-up period. After adjusting for age and income, this study found that uninsured white men had a 20 percent higher risk of dying than white men with employment-based health insurance. Uninsured black men and white women each had a 50 percent higher mortality risk than their counterparts with employment-based coverage (Sorlie et al., 1994). Among black women, insurance was not statistically associated with mortality. The authors also examined the mortality experience of insured and uninsured employed white men and women, adjusted for age and income. (Because of small sample size, they did not perform this analysis for black men and women.) Uninsured employed white men had a 30 percent greater risk of dying than their working counterparts with health insurance, and uninsured employed white women had a 20 percent greater risk over two to five years than their counterparts with health insurance (Sorlie et al., 1994).

Loss of Coverage and Changes in Health Status Over Time

Persons who lose health insurance have been found to experience declines in their health status. Longitudinal studies that follow a cohort of individuals over time can provide a “before-and-after” picture of health status, comparing a group that maintained coverage with one that lost it. Such a design helps to minimize the possibility that unmeasured factors that vary along with health insurance status account for differences in health, a competing hypothesis that cannot be eliminated in cross-sectional studies.

Lurie and colleagues (1984, 1986) took advantage of a natural experiment in the mid-1980s when California eliminated Medi-Cal coverage for a group of medically indigent adults. Following matched cohorts of adults seen at an internal medicine practice at a university clinic who either maintained or lost Medi-Cal coverage, the authors found that the patients who lost coverage reported significant decreases in perceived overall health at both six months and a year later, unlike those who maintained coverage. As discussed earlier in this chapter, participants in this study with hypertension who lost coverage also experienced worsening blood pressure control, while those who maintained coverage did not.

Like those with chronic health conditions, adults in late middle age are particularly susceptible to deteriorations of function and health status if they lack or lose health insurance coverage. Baker and colleagues (2001) followed a group of more than 7,500 participants in the longitudinal Health and Retirement Survey (adults ages 51 to 61 at the outset) between 1992 and 1996. The authors compared three groups:

those who were continuously insured over the first two years (measured in 1992 and 1994);

those who were continuously without insurance over that period; and

those who were intermittently uninsured , defined as those who lacked health insurance either in 1992 or in 1994, but not at both times (Baker et al., 2001).

Of those who were continuously uninsured, 22 percent had a major decline 17 in self-reported health, 16 percent of the intermittently uninsured experienced a major decline, and 8 percent of the continuously insured reported a major decline in health. In an analysis that controlled for sociodemographic characteristics, preexisting medical conditions, and health behaviors, the authors estimated a 60 percent greater risk of a major decline in health for continuously uninsured persons and a 40 percent greater risk for intermittently insured persons, as compared with continuously insured persons. Continuously or intermittently uninsured persons also had a 20 to 25 percent greater risk of developing a new difficulty in walking or climbing stairs than did those who were continuously insured (Baker et al., 2001).

Cross-Sectional Studies of Health Status

Cross-sectional studies based on large national population surveys (Medical Expenditure Panel Survey [MEPS], National Medical Expenditure Survey [NMES], and Behavioral Risk Factor Surveillance System, provide snapshots of the subjective or self-reported health status of populations according to insurance status. These surveys report worse health status among those without insurance than among those with coverage. Two large studies with careful and extensive analytic adjustments for covarying personal characteristics are presented here.

Franks and colleagues (1993b) examined the relationship between health insurance status and subjective health across several dimensions, including a general health perceptions scale, physical and role functions, and mental health, for 12,000 adults ages 25 through 64. The authors compared participants who had private health insurance for an entire year with those who had been without health insurance the entire year. In an analysis that controlled for age, sex, race, education, presence of a medical condition, and attitude toward medical care and insurance, uninsured adults had significantly lower subjective health scores across all dimensions. The effect on these measures of health of being uninsured was greater for lower-income persons than for those in families with incomes above 200 percent of the federal poverty level, although the effect persisted in both income groups. For both lower- and higher-income adults, the negative effect on perceived health of being uninsured was greater than that of having minority racial or ethnic status. Overall, the extent to which being uninsured negatively affected subjective health (a decrement of 4 points on a 100-point scale) was greater than that of having either of two diseases, cancer or gall bladder disease, and slightly lower than that for arteriosclerosis (Franks et al., 1993b).

Ayanian and colleagues' (2000) analysis of the 1998 BRFSS compared self-reported health status among adults 18-64 who were uninsured for a year or longer, those uninsured for less than a year, and those with any kind of insurance, public or private. Table 3.1 presents the unadjusted results for the approximately 163,000 adults surveyed. One in five adults uninsured for a year or longer reported being in fair or poor health, compared with one in seven among those uninsured for less than a year, and one in nine for those with health insurance.

Unadjusted Self-Reported Health Status for 18–64 Year-Old Adults, BRFSS, 1998 (percent).

The RAND Health Insurance Experiment

In an experimental study conducted between 1975 and 1982, about 4,000 participants between 14 and 61 years were randomly assigned (in family units) to health insurance plans that differed in the amount of patient cost sharing required, ranging from free care to major deductible plans (95 percent cost sharing, with a maximum of $1,000 per family per year) (Brook et al., 1983; Newhouse et al., 1993). Participants received a lump-sum payment at the beginning of the study to compensate them for their expected out-of-pocket costs if they were in cost-sharing plans. Participants were studied for a three- to five-year period. While persons in plans with any cost sharing had significantly fewer physician visits and hospitalizations than persons in a free-care plan, no difference was found overall between plans with any amount of cost sharing and those with no cost sharing. Free care did result in better outcomes for adults with hypertension, as discussed earlier in this chapter, and in improved visual acuity. This experiment demonstrates both the sensitivity of health care utilization in the general population to cost sharing and the relative insensitivity of short-term (three- to five-year) health outcomes for the general population to cost sharing.

Negative Results

Some studies have reported worse health status for those with health insurance compared to uninsured adults. This result may be attributable to the fact that worse health status may lead to coverage by Medicare or Medicaid, as discussed in Chapter 2 (see Box 2.1 ) and Chapter 4 . However, the competing hypothesis, that health insurance is not associated with overall health status, must also be considered.

Hahn and Flood (1995) used NMES to examine health status by both income level and type and duration of insurance coverage. When SES and demographic characteristics, health behaviors, health care utilization, and Social Security disability status were controlled for in the analysis, self-reported health status was seen to be arrayed from highest to lowest as follows:

• privately insured for the full year,
• privately insured for part of the year and uninsured for part of the year,
• uninsured for the full year,
• publicly insured for part of the year, and
• publicly insured for the full year.

The authors concluded that the likeliest explanation for their results was that the poorer health status of those who qualify for public coverage was not fully accounted for in their analytic model, even though qualification on the basis of disability was considered explicitly (Hahn and Flood, 1995). An alternative (and possibly supplementary) hypothesis was that public insurance—Medicaid specifically—provided enrollees with access and services that were less effective than those provided by private insurance. Neither of these possible explanations can be eliminated based on the research that the Committee has reviewed.

A second study by Ross and Mirowsky (2000) based on the Survey of Aging, Status and the Sense of Control (ASOC) examined the claim that being uninsured contributes to the worse health of persons of lower SES. The ASOC survey included 2,600 adults between ages 18 and 95 at baseline in 1995, 38 percent of whom were 60 years or older. Participants were reinterviewed in 1998 (44 percent were lost to follow-up) (Ross and Mirowsky, 2000). Health status, functional status, and chronic conditions reported by participants at baseline were used to predict health status, functional status, and chronic conditions three years later. Changes in these measures between baseline and follow-up were also included as predictors of health status, functional status, and number of chronic conditions at follow-up in 1998. The authors concluded that privately insured and uninsured persons had similar health status at a three-year follow-up, adjusted for baseline health status, chronic conditions, and sociodemo-graphic characteristics, and that publicly insured persons had worse health status than privately insured and uninsured adults (Ross and Mirowsky, 2000).

The Committee does not find this study convincing in its conclusions because of both the study sample and its analytic design. The sample included a large proportion of persons over 65, all of whom have Medicare, and the substantial fraction of participants lost to follow-up differed systematically from those who were reinterviewed. By including changes in health condition over the study period as independent variables along with health measures at baseline, the authors may have built their findings into the predictive model itself. In addition, Medicare beneficiaries with supplemental health insurance were classified as privately insured; thus, those who counted as publicly insured included only those Medicare beneficiaries without supplemental policies (a lower-income subset of all Medicare beneficiaries) and Medicaid beneficiaries. This atypical classification scheme distorts the comparison between those with public and private health insurance.

This chapter has presented studies examining the impact of health insurance status on general measures of population health, on health care and clinical outcomes for specific conditions, and on the appropriate use of preventive services for the nonelderly adult population in the United States. This body of research yields largely consistent and significant findings about the relationship between health insurance and health-related outcomes. In summary, uninsured adults receive health care services that are less adequate and appropriate than those received by patients who have either public or private health insurance, and they have poorer clinical outcomes and poorer overall health than do adults with private health insurance. The specific findings discussed throughout this chapter are presented in Box 3.12 .

The Committee has assessed the research regarding the effects of health insurance status across a range of health conditions and services affecting adults. In each domain examined—

• preventive care and screening services,
• cancer care and outcomes,
• chronic disease management and patient outcomes,
• acute care services and outcomes for hospitalized adults, and
• overall health status and mortality,

health insurance improved the likelihood of appropriate care and was associated with better health outcomes. Health insurance appears to achieve these positive effects in part through facilitating ongoing care with a regular health care provider and reducing financial barriers to obtaining those services that constitute or contribute to appropriate care, including screening services, prescription drugs, and specialty mental health services.

Chapter 4 specifically addresses the question of the difference that providing health insurance to uninsured individuals and populations would make to their health and health care. The Committee assesses the potential impact of health insurance coverage on those uninsured adults who are most at risk for poor or adverse health-related outcomes, including the chronically ill, adults in late middle age, members of ethnic minorities, and adults in lower-income households. The chapter also reviews the features and characteristics of health insurance that account for its effectiveness in achieving better health outcomes, including both continuity of coverage and scope of benefits.

BOX 4.1 Conclusions

The Committee's conclusions are supported by the evidence and findings presented in Chapter 3 , which are largely based on observational studies.

• Health insurance is associated with better health outcomes for adults and with their receipt of appropriate care across a range of preventive, chronic, and acute care services. Adults without health insurance coverage die sooner and experience greater declines in health status over time than do adults with continuous coverage.
• Adults with chronic conditions, and those in late middle age, are the most likely to realize improved health outcomes as a result of gaining health insurance coverage because of their high probability of needing health care services.
• Population groups that are most at risk of lacking stable health insurance coverage and that have worse health status, including racial and ethnic minorities and lower-income adults, particularly would benefit from increased health insurance coverage. Increased coverage would likely reduce some of the racial and ethnic disparities in the utilization of appropriate health care services and might also reduce disparities in morbidity and mortality among ethnic groups.
• When health insurance affords access to providers and includes preventive and screening services, outpatient prescription drugs, and specialty mental health care, it is more likely to facilitate the receipt of appropriate care than when insurance does not have these features.
• Broad-based health insurance strategies across the entire uninsured population would be more likely to produce the benefits of enhanced health and life expectancy than would “rescue” programs aimed only at the seriously ill.

Chapter 2 discusses the features of observational (nonexperimental) studies that are necessary for methodological soundness. All quantified study results that are presented in this chapter and in Chapter 4 are significant at least at the 95 percent confidence interval. If results do not meet this level of statistical significance, the confidence interval is reported. See “confidence interval” in Appendix C for further discussion.

Earlier studies based on the 1986 Access to Care Survey and the 1982 NHIS had findings consistent with those of the more recent nationally representative sample surveys regarding receipt of preventive and screening services by those without health insurance (Hayward et al., 1988; Woolhandler and Himmelstein, 1988).

Enrollees in private managed care plans is the reference group; however, fee-for-service enrollees did not have significantly different screening rates from those of managed care enrollees. The odds ratio is the relative odds of having an outcome in the uninsured and insured groups. For example, if the odds of receiving a Pap test are 2:1 in a group of uninsured women (i.e., two of every three women or 67 percent receive the test) and the odds are 4:1 in a group of women with insurance (i.e., four of every five women, or 80 percent, receive the test), the odds ratio of uninsured compared to insured women is 0.5 (2:1/4:1). The OR is not a good estimate of the relative risk (the probability of been screened in the uninsured group divided by the probability of being screened in the insured group) because screening is not a rare event. Throughout this report the results of particular studies, if reported as odds ratios or as relative risks, will be presented as the ratio of the uninsured to the insured rates (in this example, as an OR of 0.5).

Comparing results presented in Potosky et al., 1998, and Breen et al., 2001, the gap in screening rates between insured and uninsured adults decreased between 1992 and 1998.

Smoking has been associated with an increased risk of colorectal cancer (Chao et al., 2000).

This hypertension result was an exception to the overall results for the RAND study, which did not find significant differences in outcomes for most conditions and dimensions of health. These results are discussed further in the General Health Outcomes section later in this chapter.

Notably, this same study found that persons with hypertension who had Medicare coverage only (which does not pay for outpatient prescription drugs) did not have a statistically significant difference in their likelihood of receiving antihypertensive medication than uninsured persons, while those who had Medicare plus Medicaid coverage or Medicare with private supplemental insurance were significantly more likely to have received drug therapy than uninsured persons with hypertension.

BRFSS has documented the use of recommended services among insured and uninsured persons with diabetes for two recent years. BRFSS collected information on diabetes management in 1994 in 22 jurisdictions (21 states and the District of Columbia) and in 1998 in 37 jurisdictions, representing 70 percent of the U.S. population (Beckles et al., 1998; Ayanian et al., 2000).

This standard of low hematocrit is below the hematocrit target range of 33–36 percent recommended by the National Kidney Foundation's Dialysis Outcomes Quality Initiative (NKF, 2001).

In this study, “late initiation” is defined as glomerular filtration rate of serum creatinine of <5 ml/min per 1.73 m 2 —a level substantially below both that recommended by the National Kidney Foundation (<10.5 ml/min) and the U.S. mean value at initiation (<7.1 ml/min) (Kausz et al., 2000).

The HIV Cost and Services Utilization Study (HSCUS), conducted by RAND and the Agency for Healthcare Research and Quality, was a probability sample of persons 18 years and older in the contiguous United States known to have HIV infection who had one visit for regular care (except in a military, prison, or emergency treatment facility) within a two-month period in 1996. Three rounds of interviews were conducted over a two-year period, 1996–1998, with between 2,267 and 2,864 subjects (Shapiro et al., 1999). The AIDS Costs and Utilization Survey, a predecessor study to HCSUS, with six waves over 18 months in 1991 and 1992, was not a probability sample (see Box 2.4 for further detail on these surveys).

Older studies that examine hospital-based care and outcomes according to insurance status across a range of diagnoses are summarized in Appendix B . The results of these studies are consistent with the findings discussed in text; however, many are based on hospital records that may be less relevant to the current hospital practice environment.

See Leape et al. (1999) for a description of the RAND methodology for determining appropriateness and its application to developing criteria for revascularization procedures.

The American College of Surgeons designates hospital EDs as trauma centers based on qualifying criteria related to staffing, resources, and services. There are four designations: Level 1, the most stringent requirements, for providing tertiary care on a regional basis; Level 2, similar services to a Level 1 center but without clinical research and prevention activities; Level 3, presence of emergency services, often in a rural area, with fewer specialized services and resources than Level 1 or 2 centers; and Level 4, usually in a rural area, describing hospitals and clinics that serve a triage function (Bonnie et al., 1999).

The authors designated an ISS of <16 as “minimal to moderate injury” and >16 as more severe. Overall, 59 percent of transferred patients had an ISS of <9.

The lowest income category included those with a family income of less than $7,000 at the initial interview (1971–1975).

A “major decline” in health was defined as a change from excellent, very good, or good health in 1992 to fair or poor health in 1996, or from fair health in 1992 to poor health in 1996 (Baker et al., 2001).

• Cite this Page Institute of Medicine (US) Committee on the Consequences of Uninsurance. Care Without Coverage: Too Little, Too Late. Washington (DC): National Academies Press (US); 2002. 3, Effects of Health Insurance on Health.
• PDF version of this title (4.6M)

In this Page

Recent activity.

• Effects of Health Insurance on Health - Care Without Coverage Effects of Health Insurance on Health - Care Without Coverage

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

Connect with NLM

National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894

Web Policies FOIA HHS Vulnerability Disclosure

Help Accessibility Careers

Health insurance sector in India: an analysis of its performance

Vilakshan - XIMB Journal of Management

ISSN : 0973-1954

Article publication date: 30 November 2020

Issue publication date: 16 December 2020

Health insurance is one of the major contributors of growth of general insurance industry in India. It alone accounts for around 29% of total general insurance premium income earned in India. The growth of this sector is important from the perspective of overall growth of general insurance Industry. At the same time, problems in this sector are also many which are affecting its performance.

Design/methodology/approach

The paper provides an understanding on performance of health insurance sector in India. This study attempts to find out how much claims and commission and management expenses it has to incur to earn certain amount of premium. Methodology used for the study is regression analysis to establish relationship between dependent variable (Profit/Loss) and independent variable (Health Insurance Premium earned).

Findings of the study indicate that there is significant relationship between earned premium and underwriting loss. There has been increase of premium earnings which instead of increasing profit for the sector in fact has increased underwriting loss over the years. The earnings of the sector is growing at compounded annual growth rate of 27% still it is unable to earn underwriting profit.

Originality/value

This study is self-driven based on secondary data obtained from insurance regulatory and development authority site.

• Health insurance premium
• Management expenses
• Insurance regulatory and development authority
• Underwriting loss
• Compound annual growth rate

Dutta, M.M. (2020), "Health insurance sector in India: an analysis of its performance", Vilakshan - XIMB Journal of Management , Vol. 17 No. 1/2, pp. 97-109. https://doi.org/10.1108/XJM-07-2020-0021

Emerald Publishing Limited

Copyright © 2020, Madan Mohan Dutta.

Published in Vilakshan - XIMB Journal of Management . Published by Emerald Publishing Limited. This article is published under the Creative Commons Attribution (CC BY 4.0) licence. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial and non-commercial purposes), subject to full attribution to the original publication and authors. The full terms of this licence maybe seen at http://creativecommons.org/licences/by/4.0/legalcode

1. Introduction

1.1 meaning of insurance.

Insurance is a contract between two parties where by one party agrees to undertake the risk of the other in exchange for consideration known as premium and promises to indemnify the party on happening of an uncertain event. The great advantage of insurance is that it spreads the risk of a few people over a large group of people exposed to risk of similar type.

Insurance has been identified as a sunrise sector by the financial planners of India. The insurance industry has lot of potential to grow, penetrate and service the masses of India. Insurance is all about protection. An insured needs two types of protection life and non-life. General insurance industry deals with non-life protection of the insured of which health insurance is a part.

1.2 Meaning of health insurance

Health insurance is a part of general insurance which contributes about 29% of premium amongst all other sectors of general insurance. But problems in this sector are many which is the driving force behind this study. This study will help the insurance companies to understand their performance and the quantum of losses that this sector is making over the years.

A plan that covers or shares the expenses associated with health care can be described as health insurance. These plans fall into commercial health insurance, which is provided by government, private and stand-alone health insurance companies.

Health insurance in India typically pays for only inpatient hospitalization and for treatment at hospitals in India. Outpatient services are not payable under health policies in India. The first health policy in India was Mediclaim Policy. In 2000, the Government of India liberalized insurance and allowed private players into the insurance sector. The advent of private insurers in India saw the introduction of many innovative products like family floater plans, critical illness plans, hospital cash and top-up policies.

Health insurance in India is an emerging insurance sector after life and automobile insurance sector. Rise in middle class, higher hospitalization cost, expensive health care, digitization and increase in awareness level are some important drivers for the growth of health insurance market in India.

Lifestyle diseases are on the rise. A sedentary lifestyle has pervaded our being. There is lower physical labour today than earlier and there is no reason why this would not be the trend going forward. The implication is the advent of lifestyle chronic diseases such as cardiac problems and diabetes.

In the context of the Indian health insurance industry, one could look at it both ways. Mired by low penetration and negative consumer perception about its utility are affecting the prospect of this industry. The flipside though is that we have hardly scratched the surface of the opportunity that lies in the future. It is as if the glass is half full. Much remains to be conquered and even more remains to be accomplished.

Health insurance companies needs to be optimistic and have courage to bring in innovation in the areas of product, services and distribution system. Bring it to the fold as the safety net that smartly covers and craft a health insurance plan befitting the need of the customers.

1.3 Background of health insurance sector in India

India’s tryst with health insurance programme goes back to the late 1940s and early 1950s when the civil servants (Central Government Health Scheme) and formal sector workers (Employees’ State Insurance Scheme) were enrolled into a contributory but heavily subsidized health insurance programmes. As a consequence of liberalization of the economy since the early 1990s, the government opened up private sector (including health insurance) in 1999. This development threw open the possibility for higher income groups to access quality care from private tertiary care facilities. However, India in the past five years (since 2007) has witnessed a plethora of new initiatives, both by the central government and a host of state governments also entering the bandwagon of health insurance. One of the reasons for initiating such programs may be traced to the commitment of the governments in India to scale up public spending in health care.

1.4 The need for health insurance in India

1.4.1 lifestyles have changed..

Indians today suffer from high levels of stress. Long hours at work, little exercise, disregard for a healthy balanced diet and a consequent dependence on junk food have weakened our immune systems and put us at an increased risk of contracting illnesses.

1.4.2 Rare non-communicable diseases are now common.

Obesity, high blood pressure, strokes and heart attacks, which were earlier considered rare, now affect an increasing number of urban Indians.

1.4.3 Medical care is unbelievably expensive.

Medical breakthroughs have resulted in cures for dreaded diseases. These cures however are available only to a select few. This is because of high operating and treatment expenses.

1.4.4 Indirect costs add to the financial burden.

Indirect sources of expense like travel, boarding and lodging, and even temporary loss of income account for as much as 35% of the overall cost of treatment. These facts are overlooked when planning for medical expenses.

1.4.5 Incomplete financial planning.

Most of us have insured our home, vehicle, child’s education and even our retirement years. Ironically however we have not insured our health. We ignore the fact that illnesses strike without warning and seriously impact our finances and eat into our savings in the absence of a good health insurance or medical insurance plan.

1.5 Classification of health insurance plans in India

Health insurance plans in India today can be broadly classified into the following categories:

1.5.1 Hospitalization.

Hospitalization plans are indemnity plans that pay cost of hospitalization and medical costs of the insured subject to the sum insured. There is another type of hospitalization policy called a top-up policy . Top-up policies have a high deductible typically set a level of existing cover.

1.5.2 Family floater health insurance.

Family health insurance plan covers entire family in one health insurance plan. It works under assumption that not all member of a family will suffer from illness in one time.

1.5.3 Pre-existing disease cover plans.

It offers covers against disease that policyholder had before buying health policy. Pre-existing disease cover plans offers cover against pre-existing disease, e.g. diabetes, kidney failure and many more. After waiting for two to four years, it gives covers to the insured.

1.5.4 Senior citizen health insurance.

This type of health insurance plan is for older people in the family. It provides covers and protection from health issues during old age.

1.5.5 Maternity Health insurance.

Maternity health insurance ensures coverage for maternity and other additional expenses.

1.5.6 Hospital daily cash benefit plans.

Daily cash benefits are a defined benefit policy that pays a defined sum of money for every day of hospitalization.

1.5.7 Critical illness plans.

These are benefit-based policies which pay a lump sum amount on certain critical illnesses, e.g. heart attack, cancer and stroke.

1.5.8 Disease-specific special plans.

Some companies offer specially designed disease-specific plans such as Dengue Care and Corona Kavach policy.

1.6 Strength, weakness, opportunity and threat analysis of health insurance sector (SWOT analysis)

The strengths, weaknesses, opportunities and threats (SWOT) is a study undertaken to identify internal strengths and weaknesses as well as external opportunities and threats of the health insurance sector.

1.6.1 Strengths.

The growth trend of the health insurance sector is likely to be high due to rise in per capita income and emerging middle-income group in India. New products are being launched in this sector by different insurance companies which will help to satisfy customers need. Customers will be hugely benefited when cash less facility will be provided to all across the country by all the insurance companies.

1.6.2 Weaknesses.

The financial condition of this sector is weak due to low investment in this sector. The public sector insurance companies are still dominating this industry due to their greater infrastructure facilities. This sector is prone to high claim ratio and many false claims are also made.

1.6.3 Opportunities.

The possibility of future growth of this sector is high, as penetration in the rural sector is low. The improvement of technology and the use of internet facility are helping this sector to grow in magnitude and move towards environment-friendly paperless regime.

1.6.4 Threats.

The biggest threat of this sector lies in the change in the government regulations. The profitability of this sector is affected due to increasing expenses and claims. The economic slowdown and recession in the economy can affect growth of this sector adversely. The increasing losses and need for insurance might reach a point of no return where insurance companies may be compelled to decline an insurance policy.

1.7 Political economic socio cultural and technological analysis of health insurance sector (PEST analysis)

This analysis describes a framework of macro-environmental factors used as strategic tool for understanding business position, growth potential and direction for operations.

1.7.1 Political factors.

Service tax on premium on insurance policies is being increased by the government for past few years during budget. Government monopoly in this sector came to an end after insurance companies were opened up for private participation in the year 2000. Foreign players were allowed to enter into joint venture with their Indian counterpart with 26% holding and which was further increased to 49% in the year 2015.

1.7.2 Economic factors.

The gross savings of people in India have increased significantly thereby encouraging people to buy insurance policy to cover their risks. Insurance companies are fast becoming prominent players in the security market. As these companies have huge disposable income which they are investing in the security market.

1.7.3 Socio-cultural factors.

Increase in insurance knowledge is helping people to increase their awareness about the risk to be covered through insurance. Change in lifestyle is leading to increase in risk thereby giving an opportunity to insurance companies to innovate newer products. Societal benefit is derived by transfer of risk through insurance due to improved socio-cultural environment.

1.7.4 Technological factors.

Insurance companies deals in large database and maintaining it by the application of latest technology is huge gain for this sector. Technological advancement has helped insurance companies to sale their products through their electronic portals. This has made their task of providing service to the customers easier and faster.

2. Review of literature

After opening up of the insurance industry health insurance sector has become significant both from economic and social point of view and researchers have explored and probed these aspects.

Ellis et al. (2000) reviewed a variety of health insurance systems in India. It was revealed that there is a need for a competitive environment which can only happen with the opening up of the insurance sector. Aubu (2014) conducted a comparative study on public and private companies towards marketing of health insurance policies. Study revealed that private sector services evoked better response than that of public sector because of new strategies and technologies adopted by them. Nair (2019) has made a comparative study of the satisfaction level of health insurance claimants of public and private sector general insurance companies. It was revealed that majority of the respondents had claim of reimbursement nature through third party administrator. Satisfaction with respect to settlement of claim was found relatively higher for public sector than private sector. Devadasan et al. (2004) studied community health insurance to be an important intermediate step in the evolution of an equitable health financing mechanism in Europe and Japan. It was concluded that community health insurance programmes in India offer valuable lessons for its policy makers. Kumar (2009) examined the role of insurance in financing health care in India. It was found that insurance can be an important means of mobilizing resources, providing risk protection and health insurance facilities. But for this to happen, it will require systemic reforms of this sector from the end of the Government of India. Dror et al. (2006) studied about willingness among rural and poor persons in India to pay for their health insurance. Study revealed that insured persons were more willing to pay for their insurance than the uninsured persons. Jayaprakash (2007) examined to understand the hurdles preventing the people to purchase health insurance policies in the country and methods to reduce claims ratio in this sector. Yadav and Sudhakar (2017) studied personal factors influencing purchase decision of health insurance policies in India. It was found that factors such as awareness, tax benefit, financial security and risk coverage has significant influence on purchase decision of health insurance policy holders. Thomas (2017) examined health insurance in India from the perspective of consumer insights. It was found that consumers consider various aspects before choosing a health insurer like presence of a good hospital network, policy coverage and firm with wide product choice and responsive employees. Savita (2014) studied the reason for the decline of membership of micro health insurance in Karnataka. Major reason for this decline was lack of money, lack of clarity on the scheme and intra house-hold factors. However designing the scheme according to the need of the customer is the main challenge of the micro insurance sector. Shah (2017) analysed health insurance sector post liberalization in India. It was found that significant relationship exists between premiums collected and claims paid and demographic variables impacted policy holding status of the respondents. Binny and Gupta (2017) examined opportunities and challenges of health insurance in India. These opportunities are facilitating market players to expand their business and competitiveness in the market. But there are some structural problems faced by the companies such as high claim ratio and changing need of the customers which entails companies to innovate products for the satisfaction of the customers. Chatterjee et al. (2018) have studied health insurance sector in India. The premise of this paper was to study the current situation of the health-care insurance industry in India. It was observed that India is focusing more on short-term care of its citizens and must move from short-term to long-term care. Gambhir et al. (2019) studied out-patient coverage of private sector insurance in India. It was revealed that the share of the private health insurance companies has increased considerably, despite of the fact that health insurance is not a good deal. Chauhan (2019) examined medical underwriting and rating modalities in health insurance sector. It was revealed that while underwriting a health policy one has to keep in mind the various aspects of insured including lifestyle, occupation, health condition and habits. There have been substantial studies on health insurance done in India and abroad. But there has not been any work on performance of health insurance sector based on underwriting profit or loss.

3. Research gap

After extensive review of literature it is understood that there has not been substantial study on the performance of health insurance sector taking underwriting profit or loss into consideration. In spite of high rate of growth of earned premium, this sector is unable to make underwriting profit. This is mainly because growth of premium is more than compensated by claims incurred and commission and other expenses paid. Thereby leading to growth of underwriting loss over the years across the different insurance companies covered under both public and private sector. This unique feature of negative performance of this sector has not been studied so far in India.

4. Objectives

review health insurance scenario in India; and

study the performance of health insurance sector in India with respect to underwriting profit or loss by the application of regression analysis.

5. Research methodology

The study is based on secondary data sourced from the annual reports of Insurance Regulatory Development Authority (IRDA), various journals, research articles and websites. An attempt has been made to evaluate the performance of the health insurance sector in India. Appropriate research tools have been used as per the need and type of the study. The information so collected has been classified, tabulated and analysed as per the objectives of the study.

The data is based on a time period of 12 years ranging from 2006–2007 to 2018–2019.

Secondary data analysis has been done using regression of the form: Y =   a   +   b X

The research has used SPSS statistics software package for carrying out regression and for the various graphs Microsoft Excel software has been used.

5.1 The problem statement

It is taken to be a general assumption that whenever the premium increases the profit also increases. This determines that profits are actually dependent on the premium income. Hence, whenever the premium tends to increase, the profit made also supposed to increase.

The aim of the study is to find out whether the underwriting profit of the health insurance sector is increasing or there is an underwriting loss.

The problem statement is resolved by applying regression analysis between the premium earned and underwriting profit or loss incurred. It is assumed that if the underwriting profit increases along with the premium received, then the pattern forms a normal distribution and alternate hypothesis can be accepted and if this pattern of dependability is not found then the null hypothesis will be accepted stating that there is no relation between the premium and the underwriting loss or the underwriting profit by the sector. But what is happening in this sector is the increase in premium is leading to increase in underwriting loss. So premium is negatively impacting underwriting profit which is astonishing thing to happen and is the crux of the problem of this sector.

5.1.1 Underwriting profit/loss = net premium earned – (claim settled + commission and management expenses incurred).

Underwriting profit is a term used in the insurance industry to indicate earned premium remaining after claims have been settled and commission and administrative expenses have been paid. It excludes income from investment earned on premium held by the company. It is the profit generated by the insurance company in the normal course of its business.

5.2 Data analysis

Table 1 shows that health insurance premium increased from Rs.1910 crores in 2006–2007 to Rs. 33011 crores in 2018–2019. But claims incurred together with commission and management expenses have grown from Rs. 3349 crores to Rs. 40076 crores during the same period. So the claims and management expenses incurred together is more than the health insurance premium earned in all the years of our study thereby leading to underwriting loss.

Claim incurred shown above is the outcome of the risk covered against which premium is received and commission and management expenses are incurred to obtain contract of insurance. Both these expenses are important for insurance companies to generate new business as stiff competition exists in this sector since it was opened up in the year 2000.

Figure 1 depicts the relationship between health insurance premium earned and claims and management expenses incurred by the insurance companies of the health insurance sector for the period 2006–2007 to 2018–2019.

Bar chart between premiums earned and claims and management expenses incurred show that claims and management expenses together is higher than premium earned in all the years of the study thereby leading to losses. Claims, commission and management expenses are important factors leading to the sale of insurance policies thereby earning revenue for the insurance companies in the form of premium. But proper management of claims and commission and management expenses will help this sector to improve its performance.

Table 2 provides insight into the performance of health insurance sector in India. The growth of health insurance in India has been from Rs.1909 crores for the financial year 2006–2007 to Rs. 33011crores for the financial year 2018–2019. The growth percentage is 1629% i.e. growing at an average rate of 135% per annum. Compounded Annual Growth Rate (CAGR) is working out to be 27%.

From the same table, it can be inferred that health insurance sector is making underwriting loss in all the financial years. There is no specific trend can be seen, it has increased in some years and decreased in some other years. Here underwriting loss is calculated by deducting claims and commission and management expenses incurred from health insurance premium earned during these periods.

With every unit of increase in premium income the claims incurred together with commission and management expenses paid increased more than a unit. Thereby up setting the bottom line. So instead of earning profit due to better business through higher premium income, it has incurred losses.

Underwriting principles needs to be streamlined so that proper scrutiny of each policy is carried out so that performance of this sector improves.

It is seen from Figure 2 that there is stiff rise in premium earned over the years but claims and commission and management expenses incurred have also grown equally and together surpassed earned premium. So the net impact resulted in loss to this sector which can also be seen in the figure. It is also seen that loss is increasing over the years. So, increase in earnings of revenue in the form of premium is leading to increase in losses in this sector which is normally not seen in any other sectors.

But a time will come when commission and management expenses will stabilize through market forces to minimize underwriting losses. On the other hand, it will also require proper management of claims so that health insurance sector can come of this unprofitable period.

5.3 Interpretation of regression analysis

5.3.1 regression model..

Where Y = Dependent variable

X = Independent variablea = Intercept of the lineb = Slope of the line

5.3.2 Regression fit.

Here, Y is dependent variable (Underwriting Profit or Loss) which is to be predicted, X is the known independent variable (Health Insurance Premium earned) on which predictions are to be based and a and b are parameters, the value of which are to be determined ( Table 3 ). Y =   − 1028.737 − 0.226   X

5.3.3 Predictive ability of the model.

The value of R 2 = 0.866 which explains 86.6% relationship between health insurance premium earned and loss made by this sector ( Table 4 ). In other words, 13.4% of the total variation of the relationship has remained unexplained.

4.1 Regression coefficients ( Table 5 ).

H1.1 : β = 0 (No influence of Health Insurance Premium earned on Underwriting Profit or Loss made)

5.4.1.2 Alternative hypothesis.

H1.2 : β ≠ 0 (Health Insurance Premium earned influences underwriting Profit or Loss made by this sector)

The computed p -value at 95% confidence level is 0.000 which is less than 0.05. This is the confidence with which the alternative hypothesis is accepted and the null hypothesis is rejected. Thus regression equation shows that there is influence of health insurance premium earned on loss incurred by this sector.

The outcome obtained in this analysis is not what happens normally in the industry. With the increase of revenue income in the form of premium, it may lead to either profit or loss. But what is happening surprisingly here is that increase of revenue income is leading to increase of losses. So growth of premium income instead of influencing profit is actually influencing growth of losses.

6.1 Findings

The finding from the analysis is listed below:

The average growth of net premium for the health insurance has been around 135% per annum even then this sector is unable to earn underwriting profit.

The CAGR works out to around 27%. CAGR of 27% for insurance sector is considered to be very good rate of growth by any standard.

Along with high growth of premium, claims and commission and management expenses incurred in this sector have also grown substantially and together it surpassed in all the years of the study.

Thus, growth of claims and commission and management expenses incurred has more than compensated high rate of growth of health insurance premium earned. This resulted into underwriting loss that this sector is consistently making.

Astonishing findings has been higher rate of increase of premium earnings leading to higher rate of underwriting loss incurred over the years. Even though the sector is showing promise in terms of its revenue collection, but it is not enough to earn underwriting profit.

6.2 Recommendations

COVID 19 outbreak in India has led to a spike in health-care costs in the country. So, upward revision of premium charges must be considered to see bottom line improvement in this sector.

Immediate investigation of the claim is required. This will enable the insurers to curb unfair practice and dishonest means of making a claim which is rampant in this sector.

Health insurance market is not able to attract younger generation of the society. So entry age-based pricing might attract this group of customers. An individual insured at the age 30 and after 10 years of continuous coverage the premium will be less than the other individual buying a policy at the age of 40 for the first time.

6.3 Limitations and scope of future studies

The analysis of performance of health insurance sector in India taking underwriting profit into consideration is the only study of its kind in this sector. As a result, adequate literature on the subject was not available.

Health insurance and health care are part of medical care industry and are inter dependent with each other. So performance of health insurance sector can be better understood by taking health-care industry into consideration which is beyond the scope of the study.

This sector is consistently incurring losses. So, new ideas need to be incorporated to reduce losses if not making profits.

Opportunity of the insurance companies in this sector lies in establishing innovative product, services and distribution channels. So, continuous modification by the application of research is required to be undertaken.

Health insurance sector will take a massive hit, as tax benefit is going to be optional from this financial year. This can be a subject of study for the future.

6.4 Conclusion

This sector is prone to claims and its bottom line is always under tremendous pressure. In recent times, IRDA has taken bold step by increasing the premium rate of health insurance products. This will help in the growth of this sector.

With better technological expertise coming in from the foreign partners and involvement by the IRDA the health insurance sector in India must turn around and start to earn profit.

The COVID-19 pandemic is a challenge for the health insurance industry on various fronts at the same time it provides an opportunity to the insurers to fetch in new customers.

The main reason for high commission and management expense being cut-throat competition brought in after opening up of the insurance sector in the year 2000. So, new companies are offering higher incentives to the agents and brokers to penetrate into the market. This trend needs to be arrested as indirectly it is affecting profitability of this sector.

The study will richly contribute to the existing literature and help insurance companies to know about their performance and take necessary measures to rectify the situation.

Chart on health insurance premium earned and claims and management expenses paid

Chart on performance of health insurance sector in India

Data showing health insurance premium earned and claims and management expenses paid

. Dependent variable: Underwriting profit or loss;

. Predictors: (Constant), Health insurance premium earned

Aubu , R. ( 2014 ), “ Marketing of health insurance policies: a comparative study on public and private insurance companies in Chennai city ”, UGC Thesis, Shodgganga.inflibnet.ac.in .

Chatterjee , S. , Giri , A. and Bandyopadhyay , S.N. ( 2018 ), “ Health insurance sector in India: a study ”, Tech Vistas , Vol. 1 , pp. 105 - 115 .

Chauhan , V. ( 2019 ), “ Medical underwriting and rating modalities in health insurance ”, The Journal of Inssurance Institute of India , Vol. VI , pp. 14 - 18 .

Devadasan , N. , Ranson , K. , Damme , W.V. and Criel , B. ( 2004 ), “ Community health insurance in India: an overview ”, Health Policy , Vol. 29 No. 2 , pp. 133 - 172 .

Dror , D.M. , Radermacher , R. and Koren , R. ( 2006 ), “ Willingness to pay for health insurance among rural and poor persons: Field evidence form seven micro health insurance units in India ”, Health Policy , pp. 1 - 16 .

Ellis , R.P. , Alam , M. and Gupta , I. ( 2000 ), “ Health insurance in India: Prognosis and prospectus ”, Economic and Political Weekly , Vol. 35 No. 4 , pp. 207 - 217 .

Gambhir , R.S. , Malhi , R. , Khosla , S. , Singh , R. , Bhardwaj , A. and Kumar , M. ( 2019 ), “ Out-patient coverage: Private sector insurance in India ”, Journal of Family Medicine and Primary Care , Vol. 8 No. 3 , pp. 788 - 792 .

Gupta , D. and Gupta , M.B. ( 2017 ), “ Health insurance in India-Opportunities and challenges ”, International Journal of Latest Technology in Engineering, Management and Applied Science , Vol. 6 , pp. 36 - 43 .

Hand book on India Insurance Statistics revisited ( 2020 ), “ Insurance regulatory and development authority website ”, available at: www.irda.gov.in ( accessed 2 July 2020 ).

Jayaprakash , S. ( 2007 ), “ An explorative study on health insurance industry in India ”, UGC Thesis, Shodgganga.inflibnet.ac.in .

Kumar , A. ( 2009 ), “ Health insurance in India: is it the way forward? ”, World Health Statistics (WHO) , pp. 1 - 25 .

Nair , S. ( 2019 ), “ A comparative study of the satisfaction level of health insurance claimants of public and private sector general insurance companies ”, The Journal of Insurance Institute of India) , Vol. VI , pp. 33 - 42 .

Savita ( 2014 ), “ A qualitative analysis of declining membership in micro health insurance in Karmataka ”, SIES Journal of Management , Vol. 10 , pp. 12 - 21 .

Shah , A.Y.C. ( 2017 ), “ Analysis of health insurance sector post liberalisation in India ”, UGC Thesis, Shodgganga.inflibnet.ac.in .

Thomas , K.T. ( 2017 ), “ Health insurance in India: a study on consumer insight ”, IRDAI Journal , Vol. XV , pp. 25 - 31 .

Yadav , S.C. and Sudhakar , A. ( 2017 ), “ Personal factors influencing purchase decision making: a study of health insurance sector in India ”, BIMAQUEST , Vol. 17 , pp. 48 - 59 .

Further reading

Beri , G.C. ( 2010 ), Marketing Research , TATA McGraw Hill Education Private , New Delhi, ND .

Dutta , M.M. and Mitra , G. ( 2017 ), “ Performance of Indian automobile insurance sector ”, KINDLER , Vol. 17 , pp. 160 - 168 .

Majumdar , P.I. and Diwan , M.G. ( 2001 ), Principals of Insurance , Insurance Institute of India , Mumbai, MM .

Pai , V.A. and Diwan , M.G. ( 2001 ), “ Practice of general insurance ”, Insurance Institute of India , Mumbai, MM .

Shahi , A.K. and Gill , H.S. ( 2013 ), “ Origin, growth, pattern and trends: a study of Indian health insurance sector ”, IOSR Journal of Humanities and Social Science , Vol. 12 , pp. 1 - 9 .

Corresponding author

Related articles, all feedback is valuable.

Please share your general feedback

Report an issue or find answers to frequently asked questions

Contact Customer Support

• Open access
• Published: 23 October 2024

Artificial intelligence in healthcare: a scoping review of perceived threats to patient rights and safety

• Nkosi Nkosi Botha   ORCID: orcid.org/0000-0001-7879-3459 1 , 7 ,
• Cynthia E. Segbedzi 1 ,
• Victor K. Dumahasi 2 ,
• Samuel Maneen 1 ,
• Ruby V. Kodom 4 ,
• Ivy S. Tsedze 3 ,
• Lucy A. Akoto 7 ,
• Fortune S. Atsu 6 ,
• Obed U. Lasim 5 &
• Edward W. Ansah 1  

Archives of Public Health volume  82 , Article number:  188 ( 2024 ) Cite this article

Metrics details

The global health system remains determined to leverage on every workable opportunity, including artificial intelligence (AI) to provide care that is consistent with patients’ needs. Unfortunately, while AI models generally return high accuracy within the trials in which they are trained, their ability to predict and recommend the best course of care for prospective patients is left to chance.

This review maps evidence between January 1, 2010 to December 31, 2023, on the perceived threats posed by the usage of AI tools in healthcare on patients’ rights and safety.

We deployed the guidelines of Tricco et al. to conduct a comprehensive search of current literature from Nature, PubMed, Scopus, ScienceDirect, Dimensions AI, Web of Science, Ebsco Host, ProQuest, JStore, Semantic Scholar, Taylor & Francis, Emeralds, World Health Organisation, and Google Scholar. In all, 80 peer reviewed articles qualified and were included in this study.

We report that there is a real chance of unpredictable errors, inadequate policy and regulatory regime in the use of AI technologies in healthcare. Moreover, medical paternalism, increased healthcare cost and disparities in insurance coverage, data security and privacy concerns, and bias and discriminatory services are imminent in the use of AI tools in healthcare.

Conclusions

Our findings have some critical implications for achieving the Sustainable Development Goals (SDGs) 3.8, 11.7, and 16. We recommend that national governments should lead in the roll-out of AI tools in their healthcare systems. Also, other key actors in the healthcare industry should contribute to developing policies on the use of AI in healthcare systems.

Peer Review reports

Introduction

The global health system is facing unprecedented pressures due to the changing demographics, emerging diseases, administrative demands, dwindling and large migration of workforce, increasing mortality and morbidity, and changing demands and expectations in information technology [ 1 , 2 ]. Meanwhile, the needs and expectations of patients are increasing and getting ever complicated [ 1 , 3 ]. The global health system is thus, forced to leverage every opportunity, including the use of artificial intelligence (AI), to provide care that is consistent with patients’ needs and values [ 4 , 5 ]. As expected, AI has become an obvious and central theme in the global narrative due to its enormous potential positive impacts on the healthcare system. AI, in this context, should be construed as capability of computers to perform tasks similar to those performed by human professionals even in healthcare [ 6 , 7 ]. This includes the ability to reason, discover and extrapolate meanings, or learn from previous experiences to achieve healthcare goals artificially [ 4 ].

The term AI, a term credited to Sir John McCarthy since 1956, is vast, and it seems there is no consensus yet on what truly constitutes AI [ 8 , 9 ]. AI is not a single type of technology, but many different types of computerised systems (hardware and software) that require large datasets to realise their full potential [ 10 , 11 ]. AI tools are transforming the state of healthcare globally giving hope to patients with conditions that appear to defy traditional treatment techniques [ 1 , 2 , 3 ]. In clinical decision-making for instance, AI tools have improved diagnosis, reduced medical errors, stimulated prompt detection of medical emergencies, reduced healthcare cost, improved patient health outcomes, and facilitated public health interventions [ 3 , 4 ]. Additionally, AI tools have facilitated workflow, improved turnaround time for patients, and also improved the accuracy and reliability of patients’ data.

The successes of the use of AI in healthcare seems promising, if not great already, but there is the need for caution. There is the need for moderation in the celebrations and expectations of the capabilities of AI tools in healthcare, because these tools also present threats yet to be fully understood and appreciated [ 6 , 12 13 ]. So far, there are serious concerns that AI tools could threaten the privacy and autonomy of patients [ 2 , 11 ]. Moreover, widespread adoption and use of AI tools in healthcare could be confounded by factors such as lack of standardised patients’ data, inadequate curated datasets, and lack of robust legal regimes that clearly define standards for professional practice using AI tools [ 11 ]. Additionally, socio-cultural differences, lack of government commitment, proliferation of AI-savvy persons with malicious intents, irregular supply of electric power, and poverty (especially in the global south) are but a few of the many factors that may work against the potentials of AI tools in healthcare [ 14 ]. For instance, algorithms on which AI tools operate can be weaponised to perpetuate discrimination based on race, age, gender, sexual identity, socio-cultural background, social status, and political identity [ 15 , 16 ]. Notwithstanding their immense capabilities, AI tools are but a means to an end and not an end in themselves.

There is also a growing concern over how AI tools could facilitate and perpetuate unprecedented “infodemic” of misinformation via online social media networks that threaten global public health efforts [ 17 , 18 , 19 , 20 ]. In fact, the pandemic of disinformation has led to the coining of the term “infodemiology”, now acknowledged by WHO and other public health organisations globally as an important scientific field and critical area of practice especially during major disease outbreaks [ 17 , 18 , 19 , 20 ]. Recognising the consequences of disinformation to patients’ rights and safety and the potential of AI tools in facilitating same, public health experts have suggested a tighter control over patients’ information, and advocated for eHealth literacy and science and technology literacy [ 17 , 18 , 19 , 20 ]. Additionally, the experts also suggested the need to encourage peer review and fact checking systems to help improve the knowledge and quality of information regarding patient care [ 17 , 18 , 19 , 20 ]. Furthermore, there is the need to eliminate delays in the translation and transmission of knowledge in healthcare to mitigate distorting factors such as political, commercial, or malicious influences, as was widely reported during the SARS-CoV-2 outbreak [ 17 , 18 , 19 , 20 ].

Moreover, it is difficult to demonstrate how the deployment of AI tools in healthcare is contributing to the realisation of the Sustainable Development Goals (SGDs) 3.8, 11.7, and 16. For instance, SDG 11.7 provides for universal access to safe, inclusive and accessible public spaces, especially for women and children, older persons and persons with disabilities [ 24 ]. Moreover, SDG 3.8 calls for the realisation of universal health coverage, including access to quality essential healthcare services and essential medicines and vaccines for all. SDG 16 advocates for peaceful, inclusive, and just societies for all and building effective, accountable and inclusive institutions at all levels [ 24 ]. Thus, to achieve these and many others, there are many questions to be answered.

For instance, will the usage of AI tools in their present situations help achieve these SGDs by 2030? What constitutes professional negligence of AI tools in healthcare? Who takes responsibility for the commissions and omissions of AI tools in healthcare? What remedies accrue to patients who suffer serious adverse events from care provided by AI tools? What are the implications of using AI tools in healthcare on insurance policies of patients? To what extent is an AI tool developer liable for the actions and inactions of these intelligent tools? What constitutes informed consent when AI tools provide care to patients? In the event of conflicting decisions between AI tools and human clinicians, which would hold sway? Obviously, a lot more research, including reviews, are needed to clearly and confidently respond to these and several other nagging questions. Despite considerable research globally on AI, majority of these research have been done in non-clinical settings [ 22 , 23 ]. For instance, randomised controlled studies, the gold standard in medicine, are yet to provide further and better evidence on how AI adversely impacts patients [ 23 ]. Therefore, the objective of this review is to map current existing evidence on the perceived threats by AI tools in healthcare on patients’ rights and safety.

Considering the social implications, this review is envisaged to positively impact the development, deployment, and utilisation of AI tools in patient care services [ 3 , 25 , 26 , 27 , 28 , 29 ]. This is anticipated as the review to interrogate the main concerns of the patients and the general public regarding the use of these intelligent machines. The preposition is that these tools have the possibility for unpredictable errors, couple with inadequate policy and regulatory regime, may increase healthcare cost and create disparities in insurance coverage, breach privacy and data security of patients, and provide bias and discriminatory services which can be worrying [ 2 , 7 , 10 , 25 ]. Therefore, the review envisaged that manufacturers of AI tools will pay attention and factor these concerns into the production of more responsible and patient-friendly AI tools and software. Additionally, medical facilities would subject newly procured IA tools and software to a more rigorous machine learning regime that would allay the concerns of patients and guarantee their rights and safety [ 25 , 26 , 27 ]. Moreover, the review may trigger the formulation and review of existing policies at the national and medical facility levels, which would provide adequate promotion and protection of the rights and safety of patients from the adverse effects of AI tools [ 26 , 27 , 28 ].

Furthermore, there are practical implications of this review to the deployment and application of AI tools in patient care. For instance, this review would remind healthcare managers of the need to conduct rigorous machine learning and simulation exercises for AI tools before deploying them in the care process [ 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 27 , 28 , 29 ]. Moreover, medical professionals would have to scrutinise decisions of the AI tools before making final judgements on patients’ conditions. Again, healthcare professionals would find a way to make patients active participants in the care process. Finally, the review would draw attention of researchers to the issues that could undermine the acceptance of AI tools in patients care services [ 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 ]. For instance, this review may inform future research direction that explores potential threats posed by AI tools to patients’ rights and safety.

Several reviews are published recently (between January 1, 2022 and June 25, 2024) on the application of AI tools and software use in healthcare [ 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 ] (See Table  1 ). Almost halve (9 articles) of these recent reviews [ 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 ] explored the positives impacts of AI tools on healthcare services while almost halve (9 articles) [ 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 ] also examined both the positive and potential threats. Of these recent reviews, only one articles [ 48 ] studied the challenges pertaining to the adoption of AI tools in healthcare. Thus far, the current review provided a more focused and comprehensive perspectives to the threats posed by AI tools to patients’ rights and safety. The current review specifically interrogates the diverse and collates rich evidence from the perspectives of patients, healthcare workers, and the general public regarding the perceived threats posed by AI tools to patients’ rights and safety.

We scrutinised, synthesised, and analysed peer review articles according to the guidelines by Tricco et al. [ 49 ]. Thus, (1) definition and examination of study purpose, (2) revision and thorough examination of study questions, (3) identification and discussion of search terms, (4) identification and exploration of relevant databases/search engines and download of articles, (5) data mining, (6) data summarisation and synthetisation of result, and (7) consultation.

Research questions

Six study questions guided this review. They are: (1) What are the implications of AI tools on medical errors? (2) What are the ethicolegal implications of AI tools to patient care? (3) What are the implications of AI tools on patients-provider relationship? (4) What are the implications of AI tools on the cost of healthcare and insurance coverage? (5) What are the potential threats of AI tools on patients’ rights and data security? And (6) What are the perceived implications of AI tools on discrimination and bias in healthcare?

Search strategy

We mapped evidence on the topic using the Preferred Reporting Items for Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) [ 49 , 50 ]. We searched the following databases/search engines for peer review articles: Nature, PubMed, Scopus, ScienceDirect, Dimensions, Web of Science, Ebsco Host, ProQuest, JStore, Semantic Scholar, Taylor & Francis, Emeralds, World Health Organisation, and Google Scholar (see Fig.  1 ; Table  2 ). To ensure the search process was rigorous and detailed, we first searched in the PubMed using Medical Subject Headings (MeSH) terms on the topic (see Table  2 ). The search was conducted at two levels based on the search terms. First, the search terms “Confidentiality” OR “Artificial Intelligence” produced 4,262 articles. Second, the search was guided using 30 MeSH terms and controlled vocabularies which also yielded 1,320 articles (see Fig.  1 ; Table  2 ).

PRISMA flow diagram of articles used in the current review conducted from January 1, 2010 to December 31, 2023

The search covered studies conducted between January 1, 2010 and December 31, 2023, because the use of AI in healthcare is generally new and mostly unknown to people in the past three decades. Moreover, we conducted the study between January 1 and December 31, 2023. Through a comprehensive data screening process, we separated all duplicate articles into a folder, which were later removed. These articles also included those that were inconsistent with the inclusion threshold (see Table  2 ). The initial screening was conducted by authors 4, 5, 6, 7, 8, and 9, but where the qualification of an article was in doubt, that article was referred to authors 1, 3, 4 and 10 for further assessment until consensus was reached. Moreover, 1 and 10 further reviewed the data. To enhance comprehension and rigour in the search process, citation chaining was conducted on all full-text articles that met the inclusion threshold to identify additional relevant articles for further assessment. Table  2 presents inclusion and exclusion criteria used in selecting relevant articles for this review.

Quality rating

We conducted a quality rating of all selected full-text articles based on the guideline prescribed by Tricco et al. [ 49 ]. Thus, the reviewed article must provide a research background, purpose, context, suitable method, sampling, data collection and analysis, reflectivity, value of research, and ethics. We assessed and scored all selected articles based on the set criteria [ 49 ]. Thus, articles which scored “A” had few or no limitation, “B” had some limitations, “C” had substantial limitations but possess value, and “D” carry substantial flaws that could compromise the study as a whole. Therefore, articles scoring “D” were removed from the review [ 49 ].

Data extraction and thematic analysis

All authors independently extracted the data. Authors 5, 6, 7, 8, and 9, extracted data on “authors, purpose, methods, and country”, while authors 1, 2, 3, 4, and 10 extracted data on “perceived threats and conclusions” (see Table  3 ). Leveraging on Cypress [ 51 ], Morse [ 52 ], qualitative thematic analysis was conducted by authors 1, 2, 3, 4, and 10. Data were coded and themes emerged directly from the data consistent with study questions [ 53 , 54 ]. Specifically, the analysis included repeated reading of the articles to gain deep insight into the data. We further created initial candidate codes, identified and examined emerging themes. Additionally, candidate themes were reviewed, properly defined and named, and extensively discussed until a consensus was reached. Finally, we composed a report and extensively reviewed it to ensure internal and external cohesion of the themes (see Table  4 ).

This scoping review covered 2010 to 2023 on the perceived threats of AI use in healthcare on the rights and safety of patients. We screened 1,320, of which 519(39%) studied AI application in healthcare, but only 80(15%) met the inclusion threshold, passed the quality rating and were included in this review. From the 80 articles, 48(60%) applied quantitative approach, 23(29%) qualitative, and 9(11%) mixed method. The 80 articles covered 2023–1(1.25%), 2022–7(8.75%), 2021–24(30%), 2020–21(26.25%), 2019–9(11.25%), 2018–7(8.75%), 2017–7(8.75%), 2016–1(1.25%), 2015–2(2.5%), and 2014–1(1.25%). This shows that the years 2020 and 2021 alone accounted for majority (56.25%) of the articles under review. Furthermore, 26(32.5%) of the articles came from Asia alone, 22(27.5%) from only North America, 18(22.5%) from only Europe, 5(6.25%) from only Australia, 5(6.25%) from only South America, 2(2.5%) from only Africa, 1(1.25%) from North America and Asia and 1(1.25%) from North America and Europe (see Fig.  2 below).

Geographical distribution of articles used in the current review

Perceived unpredictable errors

We report that majority of the articles reviewed revealed a widespread concern over the possibility of unpredictable errors associated with the use of AI tools in patient care. Of the 80 articles reviewed, 56(70%) [ 2 , 55 , 56 , 57 , 58 , 59 , 60 , 61 , 62 , 63 , 64 , 65 , 66 , 67 , 68 , 69 , 70 , 71 , 72 , 73 , 74 , 75 , 76 , 77 , 78 , 79 , 80 , 81 , 82 , 83 , 84 , 85 , 86 , 87 , 88 , 89 , 90 , 91 , 92 , 93 , 94 , 95 , 96 , 97 , 98 , 99 , 100 , 101 , 102 , 103 , 104 , 105 , 106 , 107 ] reported the concern of AI tools committing unintended errors during care. Consistent with the operations of all machines, be it intelligent or not, AI tools could commit errors with potentially immeasurable consequences to patients [ 60 , 61 , 62 , 63 , 64 , 65 , 100 , 103 , 106 ]. This has triggered some level of hesitation and suspicion for AI applications in healthcare [ 2 , 57 , 63 , 70 ]. Perhaps, because the use of AI tools in healthcare is largely new and still emerging, the uncertainties and suspicions about their abilities and safety are largely in doubt [ 1 , 3 , 6 , 25 , 26 , 27 , 28 , 29 ]. Moreover, there are centuries of personal and documented accounts of medical errors (avoidable or not) within the healthcare industry, but it is doubtful who becomes responsible or liable if such AI tools commute errors (see Figs.  3 and 4 ).

Inadequate policy and regulatory regime

The public was also seriously concerned about lack of adequate policies and regulations, specifically on AI use in healthcare, that define the legal and ethical standards of practice. This is evident in 29(36%) of the articles [ 56 , 58 , 59 , 60 , 78 , 79 , 89 , 72 , 73 , 74 , 75 , 76 , 77 , 78 , 79 , 80 , 81 , 82 , 83 , 84 , 85 , 86 , 87 , 88 , 89 , 90 , 91 , 92 , 93 , 94 , 96 , 97 , 101 , 108 , 109 , 110 , 111 , 112 , 113 , 114 , 115 , 116 , 117 , 118 , 119 , 120 , 121 , 122 , 123 , 124 ] reviewed in this study. As with all machines, AI tools could get it wrong [ 56 , 78 , 79 , 94 , 97 , 101 ], through malfunction, with potentially terrible consequences to the health and well-being of patients. Thus, where lies the burden of liability in case of breach of duty of care, privacy, trespass, or even negligence? There were no specific regulations on AI use in healthcare to respond to the scope and direction of liability for ‘professional misconducts’ of intelligent [ 59 , 60 , 78 , 79 , 96 , 97 , 99 ], or unintelligent conducts of the machines. This finding is anticipated because the healthcare sector is already characterised by disputes between patients and the medical facilities (including their agents) [ 12 , 22 , 23 , 48 ]. Generally, patients want to be clear on what remedies accrue to them when there is a breach in duty of care. Moreover, the healthcare professionals on their part want to be clear on who takes responsibility when AI tools provide care that is sub-optimal [ 12 , 22 , 48 ]. Somebody must be responsible, is it the AI tool, manufacturer, healthcare facility or who?

Perceived medical paternalism

The application of AI tools could also interfere with the traditional patient-doctor interactions and potentially undermine patient satisfaction and the overall quality of care. This was reported by 22(27%) of the articles reviewed [ 2 , 55 , 60 , 79 , 84 , 92 , 96 , 97 , 99 , 100 , 101 , 102 , 116 , 118 , 125 , 126 , 127 , 128 , 129 , 130 , 131 , 132 ]. We argued that AI tools lacked adequate humanity required in patient care. Though AI tools may have the ability to better predict the moods of patients, they may not be trusted to competently provide very personal and private services, such as psychological and counselling care [ 2 , 55 , 84 , 92 , 101 , 102 , 116 , 118 ]. Thus, the personal and human touch that define the relationship between patients and human clinicians may not be guaranteed through AI applications [ 2 , 97 , 99 , 100 , 118 , 125 , 129 , 132 ]. It is highly expected that patients will fear losing the opportunity to interact directly with human caregivers (through verbal and non-verbal cues) [ 11 , 12 , 13 , 14 , 15 , 23 ]. The question is, is the use of AI tools sending patient care back to the application of biomedical model in healthcare? Therefore, the traditional human-to-human interactions of patients and the medics may be lost when machine clinicians replace human clinicians in patient care [ 11 , 12 , 13 , 14 , 15 , 23 ].

Increased healthcare cost and disparities in insurance coverage

Evidence also showed the public is concerned that the use of AI tools will increase the cost of healthcare and insurance coverage 7(9%) [ 2 , 76 , 77 , 119 , 122 , 133 ]. Given that adoption of AI tools in healthcare could be capital-intensive and potentially inflate operational cost of care, patients are likely to be forced to pay far more for services beyond their economic capabilities. Moreover, most health insurance policies have not yet cover services provided by AI tools leading to dispute in the payment of bills for services relating to AI applications [ 2 , 77 , 119 , 133 ]. Already, healthcare cost is one major concern for patients globally [ 7 , 11 , 16 , 27 ]. Therefore, it is legitimate for patients and the public to become anxious about the possibility of AI tools worsening the rising cost of healthcare and triggering disparities in health insurance coverage. Cost of machines learning, cost of maintenance, cost of data, cost of electricity, cost of security and safety of AI tools and software, cost of training and retraining of healthcare professionals in the use of AI tools, and many other related costs could escalate the overhead cost of providing and receiving essential healthcare services [ 7 , 11 , 16 , 27 ].

Breach of privacy and data security

We report that the public is concerned about the breach of patient privacy and data security by AI tools. As reported by 5(7%) of the articles [ 2 , 55 , 79 , 81 , 119 , 123 ] reviewed, AI tools have the potential to gather large volumes of patient data in a split of a second, sometimes at the blind side of the patients or their legal agents. As argued by Morgenstern et al. [ 79 ] and Richardson et al. [ 2 ], given their sheer complexity and automated abilities, it will be difficult to foretell when and how a specific patient data are acquired and used by AI tools, a tuition the presents a ‘black box’ for patients. Thus, apart from what the patient may be aware of, there was no surety of what else these machine clinicians could procure, albeit unlawfully, about the patient. Furthermore, it is unclear how patient data are indemnified against wrongful use and manipulation [ 2 , 119 , 123 ]. These AI tools could, wittingly or unwittingly disclose privileged information about a patient with potentially dire consequences for the privacy and security of patients. It is expected that patients would be apprehensive about the privacy and security of their personal information stored by AI tools [ 5 , 8 , 10 , 11 , 12 , 13 , 14 , 15 , 16 ]. Given that these AI tools could act independently, patients would naturally be worried about what happens to their personal information.

Potential for bias and discriminatory services

The results further suggest that there is potential for discrimination and bias on a large scale when AI tools are used in healthcare. As reported by 5(6%) of the articles [ 2 , 57 , 79 , 89 , 112 ] we reviewed, the utility of AI is a function of its design and the quality of training provided [ 2 , 57 , 112 ]. In effect, if the data used for training these machines discriminate against a population or group, this could be perpetuated and potentially be escalated when AI tools are deployed on a wider scale to provide care [ 2 , 57 , 79 ]. Thus, AI tools could perpetuate and escalate pre-exiting biases and discrimination, leaving affected populations more marginalised than ever [ 2 , 89 , 112 ]. A common feature in healthcare globally is the issues of bias and discrimination in the patient care [ 8 , 13 , 37 ]. Therefore, the fear that AI tools could be setup to provide bias and discriminatory care is both real and legitimate, because their actions and inactions are based on the data and machine learning provided [ 8 , 13 , 37 ].

There is a steady growth in AI research across diverse disciplines and activities globally [ 1 , 134 ]. However, previous studies [ 4 , 23 ] raised concerns about the paucity of empirical data on AI use in healthcare. For instance, Khan et al. [ 23 ] argued that majority of studies on AI usage in healthcare are unevenly distributed across the world and many are also conducted in non-clinical environments. Consistent with these findings, the current review showed that there is inadequate empirical evidence on the perceived threats of AI use in healthcare. Of the 519 articles on AI use in healthcare, only 80(15%) met the inclusion threshold of our study. Moreover, affirming findings from the previous studies [ 21 , 135 ], we found uneven distribution of these selected articles across the continents, with majority ( n  = 66; 82.5%) coming from three continents; Asia ( n  = 26; 32.5%), North America – ( n  = 22; 27.5%), and Europe – ( n  = 18; 22.5%). We discussed our review findings under perceived unpredictable errors, inadequate policy and regulatory regime, perceived medical paternalism, increased healthcare cost and disparities in insurance coverage, perceived breach of privacy and data security, and potential for bias and discriminatory services.

There is little contention of the capacity of AI tools to significantly reduce diagnostic and therapeutic errors in healthcare [ 10 , 138 , 139 , 140 ]. For instance, the huge data processing capacity and novel epidemiological features of modern AI tools are very effective in the fight against complex infectious diseases such as the SARS-CoV-2 and a game-changer in epidemiological research [ 140 ]. However, previous studies [ 1 , 12 , 22 ] found that AI tools are limited by factors that could undermine their efficacy and produce adverse outcomes on patients. For instance, power surges, poor internet connectivity, flawed data and faulty algorithms, and hacking could confound the efficacy of IA applications in healthcare. Indeed, hacking and internet failure could constitute the most dangerous threats to the use of AI tools in healthcare especially in resource-limited countries where internet speed and penetration are very poor [ 8 , 9 , 10 , 11 , 12 , 13 ]. Furthermore, we found that fear of unintended harm on patients by AI tools was widely reported by the articles (70%) we reviewed. For instance, potential for unpredictable errors were raised in a study that investigated perspectives about AI use in healthcare [ 99 ]. Similarly, Meehan et al. [ 141 ] argued that the generalizability and clinical utility of most AI applications are yet to be formally proven. Besides, concerns over AI related errors featured in a study on diagnostic performance, feasibility, and end-user experiences of AI assisted diabetic retinopathy [ 88 ]. Also, in the application of an AI-based Decision Support System (DSS) in the emergency department [ 81 ], such error concerns were raised.

The evidence is that, patients were in fear of being told “we do not know what went wrong”, when AI tools produce adverse outcomes [ 22 ]. This is because errors of commission or omission are associated with all machines, including these machines clinicians, whether intelligent or not [ 12 , 22 ]. Therefore, there is merit in the argument that AI tools should be closely monitored and supervised to avoid or at least minimise the impact of unintended harms to patients [ 138 ]. We are of the view that the attainment of universal health coverage, including access to quality essential healthcare services, medicines and vaccines for all by 2030 (SDG 3.8) could be accelerated through evidence-based application of AI tools in healthcare provision [ 20 ]. Thus, given that the use of AI tools in healthcare is generally new and still emerging [ 7 , 9 , 15 , 25 , 26 , 27 , 28 , 29 ], the uncertainties and suspicions about the trustworthiness of such tools (that is their capabilities and safety) are natural reactions that should be expected from patients and the general public. However, these concerns could ultimately slowdown the achievement of the SDG 3.8. Moreover, there are a lot of occurrence of medical errors (avoidable or not) within the healthcare industry with dire consequences to patients [ 13 , 29 , 30 , 37 ]. Thus, the finding comes as no surprise because medical care has always been characterised by uncertainties and unpredictable outcomes with dire consequences to patients, families, facilities and the health system [ 4 , 9 , 28 , 31 ].

The fragility of human life requires that those in the healthcare business are held to the highest standards of practice and accountability [ 13 , 24 , 137 ]. Previous studies [ 10 , 22 , 136 ] argued that healthcare must be delivered consistent with ethicolegal and professional standards that uphold the sanctity of life and respect for individuals. In keeping with this, our review showed that the public is worried about the lack of adequate protection against perceived infractions, deliberate or not, by AI tools in healthcare. Concerns over the lack of a clear policy regime to regulate the use of AI applications in patient care featured in a study that integrated a deep learning sepsis detection and management platform, sepsis watch, into routine clinical care [ 92 ]. Similar concerns were raised in a study that evaluated consecutive patients for suspected Acute Coronary Syndrome [ 11 ]. Moreover, another evidence that used the Neural Network (NN) algorithm-based models to improve the Global Registry of Acute Coronary Events (GRACE) score performance to predict in-hospital mortality and acute Coronary Syndrome [ 10 ] similar concerns were found.

The contention is that, existing policy and legal frameworks are not adequate and clear enough on what remedies accrue to patients who suffer adverse events during AI care. Our view is that patients may be at risk of a new form of discrimination, especially targeted at minority groups, persons with disabilities, and sexual minorities [ 14 ]. The need for a robust policy and regulatory regime is urgent and apparent to protect patients from potential exploitation by AI tools. This finding is not strange, because the healthcare sector is already being regulated with policies which covering the various services [ 11 , 30 , 31 , 35 ]. Moreover, because patients are normally the vulnerable parties in the patient-healthcare provider relationship [ 30 , 32 , 33 , 34 , 35 , 36 ], we argue that patients would seek adequate protection from the actions and inactions of AI tools, but unfortunately, these machine tools may not have the capabilities. Moreover, human clinicians should be equally concerned about who takes responsibility for infractions of these machine clinicians during patents care [ 8 , 24 , 29 , 35 ]. Therefore, there is the need for policy that clearly define and meaning to the scope and nature of liability of the relationship between humans and machine clinicians during patient care.

Intelligent machines hold tremendous prospects for healthcare, but human interaction is still invaluable [ 3 , 21 , 141 , 142 , 143 , 144 ]. According to Checkround et al. [ 145 ], the overriding strength of AI models in healthcare is their super-abilities to leverage large datasets to foretell and prescribe the most suitable course of intervention for prospective patients. Unfortunately, the ability of AI models to predict treatment outcomes in Schizophrenia, for example, are highly context-dependent and have limited generalizability. Our review revealed that the public is equally worried that AI tools could limit the quality of interaction between patients and human clinicians. So, through empathy and compassion, human clinicians are better able to procure effective patient participation in the care process and reach decisions that best serve the personal-cultural values, norm and perspectives of the patients [ 143 ].

We found that as AI tools provide various services and care, human clinicians may end up losing some essential skills and professional autonomy [ 24 ]. For example, concerns over reduction in critical thinking and professional autonomy was raised in some studies, including a study that used socio-technical system to implement a computer-aided diagnosis [ 97 ], adherence to antimicrobial prescribing guidelines and Computerised Decision Support Systems (CDSSs) adoption [ 12 ] and barriers and facilitators to the uptake of an evidence-based Computerised Decision Support Systems (CDSS) [ 64 ]. Thus, human medics need to take a lead role in the care process and cease every opportunity to continually practice and improve their skills. We believe that because patients normally would want to interact directly with human clinicians (through verbal and non-verbal cues) and be convinced that the conditions of the patient are well understood by human beings [ 2 , 3 , 16 , 31 ]. Typically, patients want to build cordial relationship that is based on trust with their human clinicians and other human healthcare professionals. However, this may not be feasible when AI clinicians are involved in the care process [ 11 , 16 , 26 ], especially dosing so independently. Therefore, the traditional human-to-human interactions between the patients and the human medics may be lost when machine clinicians takeover patient care.

Globally, the cost of healthcare seems to be too high for the average person [ 24 ], but the usage of AI tools could reverse this and make things better [ 10 , 23 , 144 ]. A large body of literature [ 1 , 10 , 12 , 23 , 144 ] showed that deploying AI tools in healthcare could actually reduce the cost of care for providers and patients. However, we found that the public was of the opinion that AI tools could escalate the cost of healthcare [ 2 , 76 , 77 , 119 , 122 , 133 ], especially for those in the developing world such as Africa. The reason is that healthcare facilities would have to procure, operate and maintained, where the cost is certainly going to be shifted to the patients [ 2 ]. For instance, in addition to the concerns over cost of care, limited insurance coverage was a concern raised in the use of AI-based Computer-Assisted Diagnosis (CADx) in training healthcare workers [ 67 ]. Similar concerns featured in a study that explored costs and yield from systematic HIV-TB screening, including computer-aided digital chest X-Ray test [ 68 ]. Similar concerns were found in another study involving the use of a medical-grade wireless monitoring system based on wearable and AI technology [ 103 ].

Furthermore, some of our reviewed articles [ 2 , 12 ] reported that most health insurance companies were yet to incorporate AI medical services into their policies. This situation has implications for health equity and universal health coverage. We contend that the promotion of inclusive and just societies for all and building effective, accountable, and inclusive institutions at all levels by 2030 (SGD 16) may not be achieved without affordable and accessible healthcare, including the use of advanced technology like AI in health [ 24 ]. Thus, governments need to financially support healthcare facilities, especially governments in the developing world, to implement AI services and ensure that costs do not increase health disparities and rather reduce health inequalities. The cost of healthcare is one of the major barriers to access to quality healthcare services globally [ 7 , 9 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 ]. Therefore, patients and the public are anxious about how the use of AI tools in patient care may further control the cost of healthcare services. The cost of machine learning, cost of maintenance, cost of data, cost of electricity, cost of security and safety for the AI tools and software, cost of training and retraining of healthcare professionals in the use of AI tools, and many other related costs could escalate the overhead cost of providing and receiving essential healthcare services [ 7 , 16 , 25 ], but disproportionately precarious in resourced-limited societies.

Perceived breach of privacy and data security

The fundamental obligation of a healthcare system is to provide reasonable privacy for all patients and ensure adequate protection of patients’ data from malicious use [ 9 , 11 , 16 ]. Some studies [ 12 , 136 ] suggested that AI tools in healthcare guarantee better protection for patients’ privacy and data. Contrary to this, our review found that the public is worried that AI tools may undermine patient privacy and data security. This is because the existing structures for upholding patient privacy and data integrity are grossly inadequate [ 2 , 7 ]. For example, patients’ privacy and data security concerns were raised in studies that investigated the interactions between healthcare robots and older patients [ 8 ]. Similarly concerns were raised investigated AI in healthcare [ 23 ], and public perception and knowledge of AI use in healthcare, therapy, and diagnosis [ 102 ].

There seems to be merit in these fears because of the paucity of evidence to the contrary. Moreover, the current review found that AI tools could wittingly or unwittingly disclose privileged information about patient. Such a situation has potential for dire consequences to patients, including job loss, stigma, discrimination, isolation, and breakdown of relationships, trust and result in legal battles [ 11 ]. It is our view that because the use of AI tools in patient care is still emerging most patients are not very familiar with these tools and are also certain about their trustworthiness of these machine clinicians [ 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 ]. Therefore, it is very natural that patients would be apprehensive about the privacy and security of their information procured and stored by non-human medics that could not be questioned. These concerns are widespread because of the capacity AI tools to act independently [ 11 , 16 ].

Algorithms based on flawed or limited data could trigger prejudices of racial, cultural, gender, or social status [ 2 , 4 , 11 , 15 , 24 ]. For instance, previous studies [ 3 , 12 , 15 , 24 ] reported that pre-existing and new forms of biases and discrimination against underrepresented groups could be worsened in the absence of responsible AI tools. We found that the public is concerned about the potential of AI tools discriminating against specific groups. For instance, such fears were raised in a study that assessed consecutive patients for suspected Acute Coronary Syndrome. Similar concerns were found in a study that determined the impact of AI on public health practices [ 72 ], and another study that explored the views of patients about various AI applications in healthcare [ 87 ]. Thus, the public strongly advocates for effective human oversight and governance to deflate potential excesses of AI tools during patient care [ 2 , 4 , 15 , 24 ]. Thus, we believe that algorithms employed by AI tools should not absolve medics and their facilities from responsibility. We further contend that until the necessary steps are taken, AI usage in healthcare could undermine the SDG 11.7, for universal access to safe, inclusive, and accessible public spaces for all by 2030 [ 24 ]. The evidence is that patients and the public are generally aware of bias and discriminatory services at many medical facilities [ 4 , 11 , 15 ]. Therefore, the fear that AI tools could be deliberately setup to provide biased and racialised care that may compromise rather than improve health outcomes [ 4 , 15 ].

Limitations

Notwithstanding the contributions of this study to the body of knowledge and practice, there are some limitations noteworthy. First, the use of only primary studies written in the English language may limit the literature sampled. Therefore, future research direction may resolve this by broadening the literature search beyond English Language. Therefore, future research needs to broaden the literature beyond the scope of the current review. Additionally, future research direction may have to leverage software that could translate articles written in other languages into the English Language to make future reviews far more representative than the current review. Besides, articles that have failed the inclusion criteria may have contained very useful information on the topic, so revising the inclusion and exclusion criteria could help increase the article base of future reviews. Moreover, we recognise that the current review may have inherited some weaknesses and biases from the included articles. Therefore, we acknowledge that the interpretation of some findings of this review, for instance the perceived medical paternalism, disparities in insurance coverage, bias and discriminatory services, may differ across the globe. Thus, future research direction may have to reflect carefully over the context of the candidate articles before drawing conclusions on the findings. Additionally, it is proposed that future research direction carefully examine the limitations reported in the included articles to shape the discussion and conclusions reached. This would help improve the overall reliability of the findings and conclusions reached by future reviews.

Possible future research direction

Comparing the previous and later approaches and interventions at addressing the challenges in patient care, AI tools are emerging as arguably the most promising technology for better health outcomes for patients. While AI tools have so far made noteworthy impacts on the healthcare industry, key actors (such as the healthcare professionals, patients, and the general public) have expressed concerns which need further and better interrogation. Therefore, it would be appropriate for future researchers to lead and shape the debate on the potential threats of the use of AI tools in healthcare and ways to address such threats. For instance, future research can focus on how AI tools compromise the total quality of care to sexual minorities, especially, in Africa and the developing world in general. This is necessary, given that this group remains largely marginalised from accessing basic healthcare services. Additionally, future research direction may deliberately and comprehensively examine how AI tools promote racialised healthcare services and make proposals for redress.

Furthermore, a future research may probe the challenges and quality of machine learning, especially, in Africa and the developing world in general. Also, future research direction could examine existing legal and policy frameworks (by comparing the situation across continents) regarding the use of AI tools in patient care. Additionally, a future research direction could look at how AI tools may contribute to the realisation of the health-related SDGs. Findings from such future research could be leveraged to improve and make AI tools more efficient, acceptable, safer, accessible, culturally sensitive, and cost effective for all. Finally, a future research direction may investigate how AI tools are contributing to disinformation which could be undermining patients’ rights and safety. This is importance given how “infodemic” of false information undermined the global fight against the SAR-CoV-2 pandemic [ 17 , 18 , 19 , 20 ]. This will help guarantee more effective and efficient approaches to upholding patients’ rights and safety during crisis such as pandemics and epidemics.

Contribution to body of knowledge

Several reviews explored the use of AI tools in healthcare [ 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 ]. While acknowledging the significant contributions of previous reviews to the field, the current review provides some novelty. The current review provided a more detailed and comprehensive outlook to the subject by focusing specifically on the potential threats intelligent tools pose to patient care. This is significant because most of the previous studies explored both the prospects and threats of AI use in healthcare. Even the few previous studies that focused on the potential threats of AI use in healthcare, are limited in scope and depth. Furthermore, while the previous reviews either considered only patients, healthcare providers with fewer articles, the current study examined AI use in health from diverse perspectives, including patients, healthcare professionals, and the general public and others using a large volume of data (80 articles) in this fast pace AI revolution. While no single study could exhaustively address all issues on the subject because of the explosion of the literature in the AI tools, the current review emphasised the need to pay attention to issues that matter to both the patients and experts in the field of patients care. Certainly, producers and designers of AI machines and software, experts in AI machine learning, medics, and governments across the world would find that findings of the current review useful in to make AI tools and software safer, efficient, cost effective, user friendly, and culturally sensitive.

Suggestions to addressing potential threats by AI tools in patients care

Healthcare professionals, manufacturers and designers of AI tools and software, and policy makers may benefit from the following suggestions to improve and make AI tools and allied devices safer, efficient, cost effective, culturally sensitive, and more accessible to all.

To ensure greater efficiency and fully optimise AI tools and software, healthcare managers need to graduate the deployment and use of the machines. Therefore, the AI tools and software should be subjected to rigorous machine learning regime using rich and robust data. The machine learning could start with a small dataset and later increased to large dataset with diverse characteristics.

Manufacturers and designers of AI tools and related machines need to collaborate with healthcare experts and researchers, coalition and experts in patient rights, and experts in medicolegal issues to ensure responsible usage of AI tools and software in healthcare.

Governments need to commission a team composed of healthcare experts and researchers, coalition and experts in patient rights, manufacturers and designers, and experts in medicolegal issues to develop policies for AI use in healthcare.

Healthcare managers could commission a team (composed of medical experts and managers) to verify decisions of AI tools during patient care. This would help ensure that patients are protected from ill decisions of AI tools during care.

We report that the use of AI tools is fast emerging in the global healthcare systems. While these tools hold enormous prospects for global health, including patient care, they present potential threats that are worthy of note. For instance, there is potential for breach of patients’ privacy and AI tools could trigger prejudices against race, culture, gender, or social status. Moreover, AI tools could commit errors that may harm or compromise patient’s quality of health, or health outcomes. Additionally, AI tools could also limit active patient participation in the care process resulting in a machine-centred care and deprive patients of psycho-emotional aspects of care. Furthermore, AI tools could potentially increase the cost of care and may even result in dispute between patients and insurance companies, generating different dimension of legal disputes. Unfortunately, there are inadequate policies and regulations that define ethicolegal and professional standards for the use of AI tools in healthcare. Clearly, these issues could undermine our quest towards the realisation of the SDGs 3.8, 11.7, and 16. To change the narrative, governments should commit to the development and deployment, and responsible use of AI tools in healthcare.

To ensure greater efficiency and fully optimise AI tools and software, healthcare managers could subject AI tools and software to rigorous machine learning regimes using rich and robust data. Also, manufacturers and designers of AI tools need to collaborate with other key stakeholders in healthcare to ensure responsible use of AI tools and software in patient care. Additionally, governments need to commission a team of AI and health experts to develop policies on AI use in healthcare.

Summary of key findings from previous reviews on the use of artificial intelligent tools in healthcare

Summary of key findings in current review on the use of artificial intelligent tools in healthcare

Data availability

No datasets were generated or analysed during the current study.

Abbreviations

• Artificial intelligence

Sustainable Development Goals

Reviews and Meta Analyses extension for Scoping Reviews–

Medical Subject Headings

Reddy S, Fox J, Purohit MP. Artificial intelligence-enabled healthcare delivery. J R Soc Med. 2019. https://doi.org/10.1177/0141076818815510 .

Article   PubMed   Google Scholar  

Richardson JP, Smith C, Curtis S, Watson S, Zhu X, Barry B, et al. Patient apprehensions about the use of artificial intelligence in healthcare. Npj Digit Med. 2021. https://doi.org/10.1038/s41746-021-00509-1 .

Article   PubMed   PubMed Central   Google Scholar  

Kerasidou A. Artificial intelligence and the ongoing need for empathy, compassion and trust in healthcare. Bull World Health Organisation. 2020. https://doi.org/10.2471/BLT.19.237198 .

Article   Google Scholar  

Rubeis G. iHealth: the ethics of artificial intelligence and big data in mental healthcare. Internet Interventions. 2022. https://doi.org/10.1016/j.invent.2022.100518 .

Solanki P, Grundy J, Hussain W. Operationalising ethics in artificial intelligence for healthcare: a framework for AI developers. AI Ethics. 2023. https://doi.org/10.1007/s43681-022-00195-z .

Chen C, Ding S, Wang J. Digital health for aging populations. Nat Med. 2023. https://doi.org/10.1038/s41591-023-02391-8 .

Naik N, Hameed BMZ, Shetty DK, Swain D, Shah M, Paul R, et al. Legal and ethical consideration in artificial intelligence in healthcare: who takes responsibility? Front Surg. 2022. https://doi.org/10.3389/fsurg.2022.862322 .

Bahl AK. Artificial intelligence and healthcare. J Clin Diagn Res. 2022. https://doi.org/10.7860/jcdr/2022/56148.17020 .

Khalid N, Qayyum A, Bilal M, Al-Fuqaha A, Qadir J. Privacy-preserving artificial intelligence in healthcare: techniques and applications. Comput Biol Med. 2023. https://doi.org/10.1016/j.compbiomed.2023.106848 .

Radanliev P, De Roure D. Advancing the cybersecurity of the healthcare system with self-optimising and self-adaptative artificial intelligence (part 2). Health Technol. 2022. https://doi.org/10.1007/s12553-022-00691-6 .

Wang Y, Chen TT, Chiu M. A systematic approach to enhance the explainability of artificial intelligence in healthcare with application to diagnosis of diabetes. Healthc Analytics. 2023. https://doi.org/10.1016/j.health.2023.100183 .

Horgan D, Romao M, Morré SA, Kalra D. Artificial intelligence: power for civilisation - and for Better Healthcare. Public Health Genomics. 2019. https://doi.org/10.1159/000504785 .

Lord R, Roseen D. Why should we care? In do no harm. New America. 2019; http://www.jstor.org/stable/resrep19972.6 . Accessed 13 Jun 2023.

Center of Intellectual Property and Technology Law (CIPTL). State of AI in Africa 2023. Nairobi, Kenya: Author. 2023; https://creativecommons.org/licenses/by-nc-sa/4.0 . Accessed 13 Jun 2023.

Cataleta MS. Humane artificial intelligence: The fragility of human rights facing AI. East-West Center. 2020; http://www.jstor.org/stable/resrep25514 . Accessed 13 Jun 2023.

Davenport TH, Kalakota R. The potential for artificial intelligence in healthcare. Future Healthc J. 2019. https://doi.org/10.7861/futurehosp.6-2-94 .

Zarocostas J. How to fight an infodemic. Lancet. 2020. https://doi.org/10.1016/S0140-6736(20)30461-X .

Eysenbach G. How to fight an infodemic: the four pillars of infodemic management. J Med Internet Res. 2020. https://doi.org/10.2196/21820 .

Hang CH, Yu P-D, Chen S, Tan CW, Chen G. MEGA: machine learning-enhanced graph analytics for infodemic risk management. IEEE J Biomedical Health Inf. 2023. https://doi.org/10.1109/JBHI.2023.3314632 .

Gallotti R, Valle F, Castaldo N, Sacco P, De Domenico M. Assessing the risk of infodemic in response to COVID-19 epidemics. Nat Hum Behav. 2020. https://doi.org/10.1038/s41562-020-00994-6 .

Manso JA, Ferrer RT, Pidevall I, Ballester J, Martin-Fumadó C. Use of photography in dermatology: ethical and legal implications. 2020; https://doi.org/10.1016/j.adengl.2019.04.020

Alami H, Lehoux P, Denis J-L, Motulsky A, Petitgand C, Savoldelli M, et al. Organisational readiness for artificial intelligence in health care: insights for decision-making and practice. J Health Organisation Manage. 2021. https://doi.org/10.1108/JHOM-03-2020-0074 .

Khan B, Fatima H, Qureshi A, Kumar S, Hanan A, Hussain J, et al. Drawbacks of artificial intelligence and their potential solutions in the healthcare sector. Biomedical Mater Devices (New York N Y). 2023. https://doi.org/10.1007/s44174-023-00063-2 .

World Health Organisation. Ethical use of artificial intelligence: Principles, guidelines, frameworks and human rights standards. In WHO consultation towards the development of guidance on ethics and governance of artificial intelligence for health: Meeting report. Geneva, Switzerland: World Health Organisation; 2021a; http://www.jstor.org/stable/resrep35680.8 . Accessed 13 Jun 2023.

Gupta P, Maharaj T, Weiss M, Rahaman N, Alsdurf H, Minoyan N, et al. Proactive contact tracing. PLOS Digit Health. 2023. https://doi.org/10.1371/journal.pdig.0000199 .

Hang C-N, Tsai Y-Z, Yu P-D, Chen J, Tan C-W. Privacy-enhancing digital contact tracing with machine learning for pandemic response: a comprehensive review. Big Data Cogn Comput. 2023. https://doi.org/10.3390/bdcc7020108 .

International Labour Organisation. World employment and social outlook. CH-1211, Geneva 22, Switzerland: International Labour Office. 2024; https://doi.org/10.54394/HQAE1085

Shaheen MY. AI in Healthcare: medical and socio-economic benefits and challenges. Preprint. 2021; https://doi.org/10.14293/S2199-1006.1.SOR-PPRQNI1.v1

Shaheen MY. Application of artificial intelligence (AI) in healthcare: a review. Preprint. 2021. https://doi.org/10.14293/S2199-1006.1.SOR-PPRQNI1.v1 .

Al Kuwaiti A, Nazer K, Al-Reedy A, Al-Shehri S, Al-Muhanna A, Subbarayalu AV, Al Muhanna D, Al-Muhanna FA. A review of the role of artificial intelligence in healthcare. J Pers Med. 2023. https://doi.org/10.3390/jpm13060951 .

Alnasser B. A review of literature on the economic implications of implementing artificial intelligence in healthcare. E-Health Telecommunication Syst Networks. 2023. https://doi.org/10.4236/etsn.2023.123003 .

Botha NN, Ansah EW, Segbedzi CE, Dumahasi VK, Maneen S, Kodom RV, Tsedze IS, Akoto LA, Atsu FS. Artificial intelligent tools: evidence–mapping on the perceived positive effects on patient–care and confidentiality. BMC Digit Health. 2024. https://doi.org/10.1186/s44247-024-00091-y .

Kitsios F, Kamariotou M, Syngelakis AI, Talias MA. Recent advances of artificial intelligence in healthcare: a systematic literature review. Appl Sci. 2023. https://doi.org/10.3390/app13137479 .

Krishnan G, Singh S, Pathania M, Gosavi S, Abhishek S, Parchani A, Dhar M. Artificial intelligence in clinical medicine: catalyzing a sustainable global healthcare paradigm. Front Artif Intell. 2023. https://doi.org/10.3389/frai.2023.1227091 .

Lambert SI, Madi M, Sopka S, Lenes A, Stange H, Buszello C, Stephan A. An integrative review on the acceptance of artificial intelligence among healthcare professionals in hospitals. Npj Digit Med. 2023. https://doi.org/10.1038/s41746-023-00852-5 .

Tucci V, Saary J, Doyle TE. Factors influencing trust in medical artificial intelligence for healthcare professionals: a narrative review. J Med Artif Intell. 2022. https://doi.org/10.21037/jmai-21-25 .

World Health Organisation. Global review of the role of artificial intelligence and machine learning in health-care financing for UHC. Geneva, Switzerland: World Health Organisation; 2023; http://creativecommons.org/lincenses/by-nc-sa/3.0/igo

Wu H, Lu X, Wang H. The application of artificial intelligence in health care resource allocation before and during the COVID-19 pandemic: scoping review. JMIR. 2023. https://ai.jmir.org/2023/1/e38397

Ahsan MM, Luna SA, Siddique Z. Machine-learning-based disease diagnosis: a comprehensive review. Healthcare. 2022. https://doi.org/10.3390/healthcare10030541 .

Ali O, Abdelbaki W, Shrestha A, Elbasi E, Alryalat MAA, Dwivedi YK. A systematic literature review of artificial intelligence in the healthcare sector: benefits, challenges, methodologies, and functionalities. J Innov Knowl. 2023. https://doi.org/10.1016/j.jik.2023.100333 .

Alowais SA, Alghamdi SS, Alsuhebany N, Alqahtani T, Alshaya AI, Almohareb SN, Aldairem A, Alrashed M, Saleh KB, Badreldin HA, Al Yami MS, Al Harbi S, Albekairy AM. Revolutionizing healthcare: the role of artificial intelligence in clinical practice. BMC Med Educ. 2023. https://doi.org/10.1186/s12909-023-04698-z .

Kooli C, Al Muftah H. Artificial intelligence in healthcare: a comprehensive review of its ethical concerns. Technological Sustain. 2022. https://doi.org/10.1108/TECHS-12-2021-0029 .

Kumar P, Chauhan S, Awasthi KL. Artificial intelligence in healthcare: review, ethics, trust challenges & future research directions. Eng Appl Artif Intell. 2023. https://doi.org/10.1016/j.engappai.2023.105894 .

Lindroth H, Nalaie K, Raghu R, Ayala IN, Busch C, Bhattacharyya A, Moreno Franco P, Diedrich DA, Pickering BW, Herasevich V. Applied artificial intelligence in healthcare: a review of computer vision technology application in hospital settings. J Imaging. 2024. https://doi.org/10.3390/jimaging10040081 .

Mohamed Fahim J. A review paper on artificial intelligence in healthcare. Int J Eng Manage Humanit (IJEMH). 2022. https://doi.org/10.13140/RG.2.2.25981.23529 .

Olawade DB, Wada OJ, David-Olawade AC, Kunonga E, Abaire O, Ling J. Using artificial intelligence to improve public health: a narrative review. Front Public Health. 2023. https://doi.org/10.3389/fpubh.2023.1196397 .

Rubaiyat M, Mondal H, Podder P, Bharati S. A review on explainable artificial intelligence for healthcare: why, how, and when? Med Comput Sci. 2023. https://doi.org/10.1109/TAI.2023.3266418 .

Aldwean A, Tenney D. Artificial intelligence in healthcare sector: a literature review of the adoption challenges. Open J Bus Manage. 2024. https://doi.org/10.4236/ojbm.2024.121009 .

Tricco AC, Lillie E, Zarin W, O’Brien KK, Colquhoun H, Levac D, et al. PRISMA extension for scoping reviews (PRISMAScR): Checklist and explanation. Ann Intern Med. 2018. https://doi.org/10.7326/M18-0850 .

Munn Z, Peters MDJ, Stern C, Tufanaru C, McArthur A, Aromataris E. Systematic review or scoping review? Guidance for authors when choosing between systematic and scoping review approach. BMC Med Res Methodol. 2018. https://doi.org/10.1186/s128018-0611-x .

Cypress BS. Rigor or reliability and validity in qualitative research, perspectives, strategies, reconceptualisation and recommendations. Dimens Crit Care Nurs. 2017. https://doi.org/10.1097/DCC.0000000000000253 .

Morse JM. Critical analysis of strategies for determining rigor in qualitative inquiry. Qual Health Res. 2015. https://doi.org/10.1177/1049732315588501 .

Sundler AJ, Lindberg E, Nilsson C, Plamer L. Qualitative thematic analysis based on descriptive phenomenology. Nurs Open. 2019. https://doi.org/10.1002/nop2.275 .

Van Wijngaarden E, Meide HV, Dahlberg K. Researching health care as a meaningful practice: towards a nondualistic view on evidence for qualitative research. Qual Health Res. 2017. https://doi.org/10.1177/1049732317711133 .

Fritsch SJ, Blankenheim A, Wahl A, Hetfeld P, Maassen O, Deffge S, et al. Attitudes and perception of artificial intelligence in healthcare: a cross-sectional survey among patients. Digit Health. 2022. https://doi.org/10.1177/20552076221116772 .

Al’Aref SJ, Singh G, van Rosendael AR, et al. Determinants of in-hospital mortality after percutaneous coronary intervention: a machine learning approach. J Am Heart Association. 2019;8:5. e011160.

Google Scholar  

Al’Aref SJ, Singh G, Choi JW, et al. A boosted ensemble algorithm for determination of plaque stability in high-risk patients on coronary CTA. J Am Coll Cardiology: Cardiovasc Imaging. 2020;13(10):2162–73.

Aljarboa S, Shah M, Kerr D. Asia Pacific Decision Sciences Institute,. Perceptions of the adoption of clinical decision support systems in the Saudi healthcare sector. In: Blake J, Miah SJ, Houghton L, Kerr D (eds). Proc. 24th Asia-Pacific Decision Science Institute International Conference, pp. 40–53; 2019.

Borracci RA, Higa CC, Ciambrone G, Gambarte J. Treatment of individual predictors with neural network algorithms improves global registry of acute coronary events score discrimination. Arch De Cardiolog´ıa De M´exico. 2021;91(1):58–65. https://doi.org/10.24875/ACM.20000011 .

Catho G, et al. Factors determining the adherence to antimicrobial guidelines and the adoption of computerised decision support systems by physicians: a qualitative study in three European hospitals. Int J Med Inf. 2020;141:104233.

Dogan MV, Beach S, Simons R, Lendasse A, Penaluna B, Philibert R. Blood-based biomarkers for predicting the risk for 4ve-year incident coronary heart disease in the Framingham Heart Study via machine learning. Genes. 2018;9:12.

Fan X et al. Utilization of self-diagnosis health chatbots in real-world settings: case study. J Med Internet Res. 2021;23:e19928.

Golpour P, Ghayour-Mobarhan M, Saki A, et al. Comparison of support vector machine, na¨ıve bayes and logistic regression for assessing the necessity for coronary angiography. Int J Environ Res Public Health. 2020;17(18):6449–50.

Horsfall HL, et al. Attitudes of the surgical team toward artificial intelligence in neurosurgery: International 2-stage cross-sectional survey. World Neurosurg. 2021;146:e724–30.

Hu D, Dong W, Lu X, Duan H, He K, Huang Z. Evidential MACE prediction of acute coronary syndrome using electronic health records. BMC Med Inf Decis Mak. 2019;19:S2.

Jauk S, et al. Technology acceptance of a machine learning algorithm predicting delirium in a clinical setting: a mixed-methods study. J Med Syst. 2021;45:48.

Joloudari JH, Hassannataj Joloudari E, Saadatfar H, et al. Coronary artery disease diagnosis; ranking the signi4cant features using a random trees model. Int J Environ Res Public Health. 2020;17(3):731.

Kanagasundaram NS, et al. Computerized clinical decision support for the early recognition and management of acute kidney injury: a qualitative evaluation of end-user experience. Clin Kidney J. 2016;9:57–62.

Kayvanpour E, Gi WT, Sedaghat-Hamedani F, et al. MicroRNA neural networks improve diagnosis of acute coronary syndrome (ACS). J Mol Cell Cardiol. 2021;151:155–62.

Article   CAS   PubMed   Google Scholar  

Khong PCB, Hoi SY, Holroyd E, Wang W. Nurses’ clinical decision making on adopting a wound clinical decision support system. Comput Inf Nurs. 2015;33:295–305.

Kim JK, Kang S. Neural network-based coronary heart disease risk prediction using feature correlation analysis. J Healthc Eng. 2017;13. https://doi.org/10.1155/2017/2780501 .

Kitzmiller RR, et al. Diffusing an innovation: clinician perceptions of continuous predictive analytics monitoring in intensive care. Appl Clin Inf. 2019;10:295–306.

Krittanawong C, Virk HUH, Kumar A, et al. Machine learning and deep learning to predict mortality in patients with spontaneous coronary artery dissection. Scienti*c Rep. 2021;11:1.

Li D, Xiong G, Zeng H, Zhou Q, Jiang J, Guo X. Machine learning-aided risk strati4cation system for the prediction of coronary artery disease. Int J Cardiol. 2021;326:30–4.

Liu X, Jiang J, Wei L, et al. Prediction of all-cause mortality in coronary artery disease patients with atrial 4brillation based on machine learning models. BMC Cardiovasc Disord. 2021;21(499):1–12. https://doi.org/10.1186/s12872-021-02314-w .

Article   CAS   Google Scholar  

Love SM, et al. Palpable breast lump triage by minimally trained operators in Mexico using computer-assisted diagnosis and low-cost ultrasound. J Glob Oncol. 2018. https://doi.org/10.1200/JGO.17.00222 .

McBride KE, Steffens D, Duncan K, Bannon PG, Solomon MJ. Knowledge and attitudes of theatre staff prior to the implementation of robotic-assisted surgery in the public sector. PLoS ONE. 2019;14:e0213840.

Mehta N, Harish V, Bilimoria K, et al. Knowledge and attitudes on artificial intelligence in healthcare: a provincial survey study of medical students. MedEd Publish. 2021. https://doi.org/10.15694/mep.2021.000075.1 .

Morgenstern JD, Rosella LC, Daley MJ, Goel V, Schünemann HJ, Piggott T. AI’s gonna have an impact on everything in society, so it has to have an impact on public health: a fundamental qualitative descriptive study of the implications of artificial intelligence for public health. BMC Public Health. 2021. https://doi.org/10.1186/s12889-020-10030-x .

Motwani M, Dey D, Berman DS, et al. Machine learning for prediction of all-cause mortality in patients with suspected coronary artery disease: a 5-year multicentre prospective registry analysis. Eur Heart J. 2017;38(7):500–7.

PubMed   Google Scholar  

Betriana F, Tanioka T, Osaka K, Kawai C, Yasuhara Y, Locsin RC. Improving the delivery of palliative care through predictive modeling and healthcare informatics. J Am Med Inf Assoc. 2021a;28:1065–73.

Naushad SM, Hussain T, Indumathi B, Samreen K, Alrokayan SA, Kutala VK. Machine learning algorithm-based risk prediction model of coronary artery disease. Mol Biol Rep. 2018;45(5):901–10.

Nydert P, Vég A, Bastholm-Rahmner P, Lindemalm S. Pediatricians’ understanding and experiences of an electronic clinical-decision-support-system. Online J Public Health Inf. 2017;9:e200.

Omar A, Ellenius J, Lindemalm S. Evaluation of electronic prescribing decision support system at a tertiary care pediatric hospital: the user acceptance perspective. Stud Health Technol Inf. 2017;234:256–61.

Orlenko A, Kofink D, Lyytik¨ainen LP, et al. Model selection for metabolomics: Predicting diagnosis of coronary artery disease using automated machine learning. Bioinformatics. 2020;36(6):1772–8.

Panicker RO, Sabu MK. Factors influencing the adoption of computerized medical diagnosing system for tuberculosis. Int J Inf Technol. 2020;12:503–12.

Petitgand C, Motulsky A, Denis J-L, Régis C. Investigating the barriers to physician adoption of an artificial intelligence-based decision support system in emergency care: an interpretative qualitative study. Digital personalized health and medicine. Amsterdam. The Netherlands: IOS; 2020. pp. 1001–5.

Pieszko K. Predicting long-term mortality after acute coronary syndrome using machine learning techniques and hematological markers. Disease Markers. 2019;2019:9.

Ploug T, Sundby A, Moeslund TB, Holm S. Population preferences for performance and explainability of artificial intelligence in health care: choice-based conjoint survey. J Med Internet Res. 2021;e26611. https://doi.org/10.2196/26611 .

Polero LD. A machine learning algorithm for risk prediction of acute coronary syndrome (angina). Revista Argentina De Cardiolog´ıa. 2020;88:9–13.

Romero-Brufau S, Wyatt KD, Boyum P, Mickelson M, Moore M, Cognetta-Rieke C. Implementation of artificial intelligencebased clinical decision support to reduce hospital readmissions at a regional hospital. Appl Clin Inf. 2020;11:570–7.

Sarwar S, Dent A, Faust K, Richer M, Djuric U, Ommeren RV, et al. Physician perspectives on integration of artificial intelligence into diagnostic pathology. Npj Digit Med. 2021. https://doi.org/10.1038/s41746-019-0106-0 .

Scheetz J, Koca D, McGuinness M, Holloway E, Tan Z, Zhu Z, et al. Real-world artificial intelligence-based opportunistic screening for diabetic retinopathy in endocrinology and indigenous healthcare settings in Australia. Sci Rep. 2021. https://doi.org/10.1038/s41598-021-94178-5 .

Schuh C, de Bruin JS, Seeling W. Clinical decision support systems at the Vienna General Hospital using Arden Syntax: design, implementation, and integration. Artif Intell Med. 2018;92:24–33.

Sherazi SWA, Jeong YJ, Jae MH, Bae JW, Lee JY. A machine learning–based 1-year mortality prediction model after hospital discharge for clinical patients with acute coronary syndrome. Health Inf J. 2020;26(2):1289–304.

Sujan M, White S, Habli I, Reynolds N. Stakeholder perceptions of the safety and assurance of artificial intelligence in healthcare. SSRN Electron J. 2022. https://doi.org/10.2139/ssrn.4000675 .

Terry AL, Kueper JK, Beleno R, Brown JB, Cejic S, Dang J, et al. Is primary health care ready for artificial intelligence? What do primary health care stakeholders say? BMC Med Inf Decis Mak. 2022. https://doi.org/10.1186/s12911-022-01984-6 .

Tscholl DW, Weiss M, Handschin L, Spahn DR, Nöthiger CB. User perceptions of avatar-based patient monitoring: a mixed qualitative and quantitative study. BMC Anesthesiol. 2018;18:188.

Ugarte-Gil C, et al. Implementing a socio-technical system for computer-aided tuberculosis diagnosis in Peru: a field trial among health professionals in resource-constraint settings. Health Inf J. 2020;26:2762–75.

Van der Zander QEW, van der Ende-van Loon MCM, Janssen JMM, Winkens B, van der Sommen F, Masclee AAM, et al. Artificial intelligence in (gastrointestinal) healthcare: patients’ and physicians’ perspectives. Sci Rep. 2022. https://doi.org/10.1038/s41598-022-20958-2 .

Visram S, Leyden D, Annesley O, et al. Engaging children and young people on the potential role of artificial intelligence in medicine. Pediatr Res. 2023;93:440–4. https://doi.org/10.1038/s41390-022-02053-4 .

Wang D et al. Brilliant AI Doctor in rural clinics: challenges in AI-powered clinical decision support system deployment. In: Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems; 2021, pp. 1–18.

Xu H, Li P, Yang Z, Liu X, Wang Z, Yan W, He M, Chu W, She Y, Li Y, et al. Construction and application of a medical-grade wireless monitoring system for physiological signals at general wards. J Med Syst. 2020;44:1–15.

Zhai H, et al. Radiation oncologists’ perceptions of adopting an artificial intelligence-assisted contouring technology: model development and questionnaire study. J Med Internet Res. 2021;23:1–16.

Zhang H, Wang X, Liu C, et al. Detection of coronary artery disease using multi-modal feature fusion and hybrid feature selection. Physiol Meas. 2020;41(11):115007.

Zhou N, et al. Concordance study between IBM watson for oncology and clinical practice for patients with cancer in China. Oncologist. 2019;24:812–9.

Zhou LY, Yin W, Wang J, et al. A novel laboratory-based model to predict the presence of obstructive coronary artery disease comparison to coronary artery disease consortium ½ score, duke clinical score and diamond-forrester score in China. Int Heart J. 2020;61(3):437–46.

Alumran A, et al. Utilization of an electronic triage system by emergency department nurses. J Multidiscip Healthc. 2020;13:339–44.

Ayatollahi H, Gholamhosseini L, Salehi M. Predicting coronary artery disease: a comparison between two data mining algorithms. BMC Public Health. 2019;19(1):448. https://doi.org/10.1186/s12889-019-6721-5 .

Baskaran L. Machine learning insight into the role of imaging and clinical variables for the prediction of obstructive coronary artery disease and revascularization: an exploratory analysis of the CONSERVE study. PLoS ONE. 2020;15:6. e0233791.

Betriana F, Tanioka T, Osaka K, Kawai C, Yasuhara Y, Locsin RC. Interactions between healthcare robots and older people in Japan: a qualitative descriptive analysis study. Jpn J Nurs Sci. 2021;18:e12409.

Bouzid Z, Faramand Z, Gregg RE, et al. In search of an optimal subset of ecg features to augment the diagnosis of acute coronary syndrome at the emergency department. J Am Heart Association. 2021;10:3. e017871.

Davari Dolatabadi A, Khadem SEZ, Asl BM. Automated diagnosis of coronary artery disease (CAD) patients using optimized SVM. Comput Methods Programs Biomed. 2017;138:117–26.

Du Z, Yang Y, Zheng J, et al. Accurate prediction of coronary heart disease for patients with hypertension from electronic health records with big data and machine-learning methods: model development and performance evaluation. JMIR Med Inf. 2020;8:7. e17257.

Gonçalves LS, Amaro MLM, Romero ALM, Schamne FK, Fressatto JL, Bezerra CW. Implementation of an artificial intelligence algorithm for sepsis detection. Rev Bras Enferm. 2020;73:e20180421.

Isbanner S, Pauline O, Steel D, Wilcock S, Carter S. The adoption of artificial intelligence in health care and social services in Australia: findings from a methodologically innovative national survey of values and attitudes (the AVA-AI study). J Med Internet Res. 2022. https://doi.org/10.2196/37611 .

Lee EK, Atallah HY, Wright MD, Post ET, Thomas CIV, Wu DT, Haley LL. Transforming hospital emergency department workflow and patient care. Interfaces. 2015;45:58–82.

Liberati EG, et al. What hinders the uptake of computerized decision support systems in hospitals? A qualitative study and framework for implementation. Implement Sci. 2017;12:1–13.

Petersson L, Larsson I, Nygren JM, Nilsen P, Neher M, Reed JE, et al. Challenges to implementing artificial intelligence in healthcare: a qualitative interview study with healthcare leaders in Sweden. BMC Health Serv Res. 2022. https://doi.org/10.1186/s12913-022-08215-8 .

Prakash A, Das S. Intelligent conversational agents in mental healthcare ser­vices: a thematic analysis of user perceptions. Pac Asia J Assoc Inf Syst. 2020;12(2):1–34. https://doi.org/10.17705/1pais.1201 .

Pumplun L, Fecho M, Wahl N, Peters F, Buxmann P. Adoption of machine learning systems for medical diagnostics in clinics: qualitative interview study. J Med Internet Res. 2021;23:e29301.

Sendak MP, Ratliff W, Sarro D, Alderton E, Futoma J, Gao M. Real-world integration of a sepsis deep learning technology into routine clinical care: implementation study. JMIR Med Inf. 2020;8:e15182.

Wittal CG, Hammer D, Klein F, Rittchen J. Perception and knowledge of artificial intelligence in healthcare, therapy and diagnostics: A population-representative survey. 2022. https://doi.org/10.1101/2022.12.01.22282960

Zheng B, et al. Attitudes of medical workers in China toward artificial intelligence in ophthalmology: a comparative survey. BMC Health Serv Res. 2021;21:1067.

Blanco N, et al. Health care worker perceptions toward computerized clinical decision support tools for Clostridium difficile infection reduction: a qualitative study at 2 hospitals. Am J Infect Control. 2018;46:1160–6.

Elahi C, et al. An attitude survey and assessment of the feasibility, acceptability, and usability of a traumatic brain injury decision support tool in Uganda. World Neurosurg. 2020;139:495–504.

Fan W, Liu J, Zhu S, Pardalos PM. Investigating the impacting factors for the healthcare professionals to adopt artificial intelligence-based medical diagnosis support system (AIMDSS). Ann Oper Res. 2020;294:567–92.

Garzon-Chavez D et al. Adapting for the COVID-19 pandemic in Ecuador, a characterization of hospital strategies and patients. PLoS ONE. 2021;16:e0251295.

Grau LE, Weiss J, O’Leary TK, Camenga D, Bernstein SL. Electronic decision support for treatment of hospitalized smokers: a qualitative analysis of physicians’ knowledge, attitudes, and practices. Drug Alcohol Depend. 2019;194:296–301.

McCoy A, Das R. Reducing patient mortality, length of stay and readmissions through machine learning-based sepsis prediction in the emergency department, intensive care unit and hospital floor units. BMJ Open Qual. 2017;6:e000158.

O’Leary P, Carroll N, Richardson I. The practitioner’s perspective on clinical pathway support systems. In IEEE International Conference on Healthcare Informatics. 2014;194–201.

Van der Heijden AA, Abramoff MD, Verbraak F, van Hecke MV, Liem A, Nijpels G. Validation of automated screening for referable diabetic retinopathy with the IDx-DR device in the hoorn diabetes care system. Acta Ophthalmol. 2018;96:63–8.

MacPherson P et al. Computer-aided X-ray screening for tuberculosis and HIV testing among adults with cough in Malawi (the PROSPECT study): a randomised trial and cost-effectiveness analysis. PLoS Med. 2021;18:e1003752.

Hitti E, Hadid D, Melki J, Kaddoura R, Alameddine M. Mobile device use among emergency department healthcare professionals: prevalence, utilisation and attitudes. Sci Rep. 2021. https://doi.org/10.1038/s41598-021-81278-5 .

Arakpogun EO, Elsahn Z, Olan F, Elsahn F. Artificial intelligence in Africa: challenges and opportunities. In: Hamdan A, Hassanien AE, Razzaque A, Alareeni B (eds). Entrepreneurship, innovation and strategy, marketing, operations and systems. Cham: Switzerland; 2022, pp. 375–88. https://doi.org/10.1007/978-3-030-62796-6_22 .

Chapter   Google Scholar  

Leenes RE, Palmerini E, Koops B, Bertolini A, Salvini P, Lucivero F. Regulatory challenges of robotics: some guidelines for addressing legal and ethical issues. Law Innov Technol; 2017. https://doi.org/10.1080/17579961.2017.1304921 .

World Health Organisation. Addressing challenges to ethics and governance. In WHO consultation towards the development of guidance on ethics and governance of artificial intelligence for health: Meeting report. Geneva, Switzerland: World Health Organisation. 2021b; http://www.jstor.org/stable/resrep35680.10 . Accessed 21 Jul 2023.

Jiang F, Jiang Y, Zhi H, Dong Y, Li H, Ma S, et al. Artificial intelligence in healthcare: past, present and future. Stroke Vascular Neurol. 2017. https://doi.org/10.1136/svn-2017-000101 .

Coiera E, Liu S. Evidence synthesis, digital scribes, and translational challenges for artificial intelligence in healthcare. Cell Rep Med. 2022;2022. https://doi.org/10.1016/j.xcrm.2022.100860 .

Fei Z, Ryeznik Y, Sverdlov O, Tan CW, Wong WK. An overview of healthcare data analytics with applications to the COVID-19 pandemic. IEEE Trans Big Data. 2022. https://doi.org/10.1109/TBDATA.2021.3103458 .

Meehan AJ, Lewis SJ, Fazel S, Fusar-Poli P, Steyerberg EW, Stahl D, Danese A. Clinical prediction models in psychiatry: a systematic review of two decades of progress and challenges. Mol Psychiatry. 2022;27:27000–2708. https://doi.org/10.1038/s41380.022-01528-4 .

Krumholz HM. In the US, patient data privacy is an illusion. BMJ (Clinical Res ed). 2023. https://doi.org/10.1136/bmj.p1225 .

Rentmeester C. Heeding humanity in an age of electronic health records: Heidegger, Levinas, and healthcare. Nurs Philos. 2018. https://doi.org/10.1111/nup.12214 .

Silva W, Sacramento CQ, Silva E, Garcia AC, Ferreira SB. Health information, human factors and privacy issues in mobile health applications. Hawaii Int Conf Syst Sci. 2020. https://doi.org/10.24251/hicss.2020.420 .

Checkround AM, Hawrilenko M, Loho H, Bondar J, Gueorguieva R, Hasan A, Kambeitz J, Corlett PR, Koutsouleris N, Krumholz HM, Krystal JH, Paulus M. Illusory generalizability of clinical prediction models. Science. 2024;383(6679):164–7. https://doi.org/10.1126/science.adg8538 .

Download references

Acknowledgements

We are grateful to Lieutenant Commander (Ghana Navy) Candice FLEISCHER-DJOLETO of 37 Military Hospital, Ghana Armed Forces Medical Services, for proofreading the draft manuscript.

No author received funding for a part or the whole of the study.

Author information

Authors and affiliations.

Department of Health, Physical Education and Recreation, University of Cape Coast, Cape Coast, Ghana

Nkosi Nkosi Botha, Cynthia E. Segbedzi, Samuel Maneen & Edward W. Ansah

Institute of Environmental and Sanitation Studies, Environmental Science, College of Basic and Applied Sciences, University of Ghana, Legon, Ghana

Victor K. Dumahasi

Department of Adult Health, School of Nursing and Midwifery, University of Cape Coast, Cape Coast, Ghana

Ivy S. Tsedze

Department of Health Services Management/Distance Education, University of Ghana, Legon, Ghana

Ruby V. Kodom

Department of Health Information Management, School of Allied Health Sciences, University of Cape Coast, Cape Coast, Ghana

Obed U. Lasim

E. P. College of Education, Amedzofe, Ghana

Fortune S. Atsu

Air Force Medical Centre, Armed Forces Medical Services, Air Force Base, Takoradi, Ghana

Nkosi Nkosi Botha & Lucy A. Akoto

You can also search for this author in PubMed   Google Scholar

Contributions

NNB, EWA, CES, SM, and VKD Conceptualised and Designed the Review Protocols. EWA, VKD, CES, FSA, RVK, IST, LAA, SM, OUL, and NNB Conducted Data Collection and Acquisition. EWA, VKD, CES, FSA, IST, LAA, SM, OUL, RVK, and NNB carried out extensive data processing and management. EWA, CES, NNB developed the initial manuscript. All authors edited and considerably reviewed the manuscript, proofread for intellectual content and consented to its publication.

Corresponding author

Correspondence to Nkosi Nkosi Botha .

Ethics declarations

Ethics approval and consent to participate.

Not applicable.

Consent for publication

All authors consented to publish this paper.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ .

Reprints and permissions

About this article

Cite this article.

Botha, N.N., Segbedzi, C.E., Dumahasi, V.K. et al. Artificial intelligence in healthcare: a scoping review of perceived threats to patient rights and safety. Arch Public Health 82 , 188 (2024). https://doi.org/10.1186/s13690-024-01414-1

Download citation

Received : 25 February 2024

Accepted : 01 October 2024

Published : 23 October 2024

DOI : https://doi.org/10.1186/s13690-024-01414-1

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Computational intelligence
• Computer vision systems
• Confidentiality
• Knowledge representation
• Privacy of patient data
• Privileged communications

Archives of Public Health

ISSN: 2049-3258

• General enquiries: [email protected]