national decision model case study

Browse Theses

January 2021

• Canterbury Christ Church University (United Kingdom)

This researched explored the under-researched area of how the National Decision Model (NDM) is utilised, focusing on how professionals (n=15) from one Police Service in England and Wales applied the model while responding to incidents and during criminal investigations. The NDM was introduced in 2012 for police officers to apply to all decisions, regardless of location, rank, or situation. While the ethical principles of the NDM have been explored in research (Adams, 2014; Lax, 2014), its practical application for decisions has not been evaluated. This research provides an exploratory study intended to fill this gap in knowledge. Based on a review of literature, four research questions were identified: (1) What are practitioner perceptions of the NDM? (2) Are practitioners using the NDM intuitively or consciously? (3) How does practitioner utilisation of the NDM vary between different levels of the Professionalising Investigations Programme (PIP)? (4) How does practitioner utilisation of the NDM vary between investigations and response? Naturalistic Decision-Making was used as a theoretical framework, exploring the use of the Recognition Primed Decision Model (RPDM) in police decisionmaking. Qualitative data was gathered through semi-structured interviews and thematically analysed using the software platform, NVivo. The findings identified three themes: Intuitive Thinking versus Conscious Thinking, An Effective Model, and Investigation versus Response. This research concluded with recommendations for the College of Policing, the participating Police Service, and further research, in the hopes of creating a foundation for Evidence-Based Practice. This research was limited by its small sample size and the effects of COVID-19 restrictions impeding access to participants. There was difficulty in obtaining an equal number of participants from each PIP level due to fewer more experienced officers and due to officers being redeployed in different departments due to COVID-19.

Save to Binder

• Publication Years 2021 - 2021
• Publication counts 1
• Citation count 0
• Available for Download 0
• Downloads (cumulative) 0
• Downloads (12 months) 0
• Downloads (6 weeks) 0
• Average Downloads per Article 0
• Average Citation per Article 0
• Publication Years
• Publication counts 0

Index Terms

Applied computing

Computers in other domains

Computing in government

E-government

Computing methodologies

Human-centered computing

Information systems

Information systems applications

Social and professional topics

Computing / technology policy

Professional topics

Recommendations

Advanced criminal investigations and intelligence operations: tradecraft methods, practices, tactics, and techniques, researcher-practitioner interaction.

This workshop explores whether problems exist between HCI researchers and the practitioners who are consumers of research - and, if so, will endeavor to identify the dimensions of the problems and propose possible solutions. On the one hand, the ...

Probing some theoretical aspects of practitioner-based inquiry research in information systems

Information and Communication Technologies (ICT) practitioners are often located in the social sciences branch of Information Systems (IS). Practitioner research comprises of research that practitioners undertake with a view to the advancement of their ...

Export Citations

• Please download or close your previous search result export first before starting a new bulk export. Preview is not available. By clicking download, a status dialog will open to start the export process. The process may take a few minutes but once it finishes a file will be downloadable from your browser. You may continue to browse the DL while the export process is in progress. Download
• Download citation
• Copy citation

We are preparing your search results for download ...

We will inform you here when the file is ready.

Your file of search results citations is now ready.

Your search export query has expired. Please try again.

• Executive viewpoint
• Industry spotlight
• Learning Hub
• Operational
• Research & Policy
• Sponsored content
• Staff Officer Stitchley
• Training & Development
• Excellence in Policing
• Case Studies
• Crime & Intelligence Analysis
• Investigative Practice
• SHARING PRACTICE
• EXCELLENCE IN POLICINdG
• DIGITAL EDITTIONS

Please login or subscribe to access the Police Professional website

Access to this content is only available to subscribers. If you have an active subscription, please enter your details to log in. If you would like subscribe to Police Professional, please click here .

Group Subscription

If your force, organisation or institution has a group subscription and you have not yet registered your details on the site, please click here .

Username or Email Address

Remember Me

To get the most from our network, tell us a little about your interests. Job and news alerts can be sent according to your interest and other users will be able to see your selection. You can skip this section and sign up in your account at a later date.

• Administration
• Administration, Finance and HR
• Analysis and Intelligence
• Corporate Development and Communications
• Crime Investigation
• Finance and HR
• Government funding
• Green policing
• Information Communications Technology (ICT)
• Officer welfare
• Operational Support and Firearms
• Pre-recorded evidence
• Professional Standards
• Remote jury trials
• Response and Neighbourhood
• Scientific and Forensics
• Strategic Leadership
• Training and Development
• The Police Professional Network allows you to connect to other users of the site. If you prefer not be seen in the Network please tick here

ACPO launches decision model for all policing

Chief constables in England and Wales have adopted a new decision-making approach for the police service that puts values at its heart.

The National Decision Model (NDM) will replace all existing decision models in policing. It is part of a concerted drive to focus on delivering the mission of policing while acting in accordance with values, enhancing the use of discretion and professional judgement, reducing risk-aversion and in so doing helping to strike the balance between demand for police services and increasingly limited resources.

The development of the NDM has been led by Chief Constable Adrian Lee and Chief Constable Brian Moore – who chair the Association of Chief Police Officers (ACPO) Ethics and Risk Co-ordination portfolios respectively – ensuring that the way the police service approaches risk has a sound ethical grounding.

Mr Lee, chief constable of Northamptonshire Police, said: “The NDM is a deliberately simple model and it shares common ingredients with the three models for decision-making it replaces. The model is for use by the whole of the police service, not just for operational decisions, and provides a practical tool for police officers and police staff making difficult decisions in challenging circumstances.

“The most significant shift for the new NDM is the central touchstone that focuses on the policing mission, the values we share, risks and protection of human rights. These factors must be considered at every stage because the decisions officers and staff make need to be both technically and ethically right.”

“The elements of the model are very familiar to people but I think rather than saying ‘I think this is how I did it’, you can say ‘there’s a model. Here are the five stages. I went through them and at each stage I was trying to implement what I understand to be the mission and values of the police service. If as a result of that some harms come that was completely avoidable then that’s how I defend my decision-making’. It’s about helping staff to properly record their rationale.”

Surrounding this central core of the NDM are five stages, each of which encourage a considered approach to intelligence, risk, and applicable powers and policy as precursors to taking action – interspersed with prompts to review action taken and develop a working strategy based on intelligence, threat and risk.

Mr Moore, chief constable of Wiltshire Police, said: “Understanding and managing risks, both those to the general public and those to the police service itself, are fundamental to effective policing.

“The adoption of the NDM is a significant development in this regard and will also contribute to achieving reductions in bureaucracy as well as developments in safeguarding and public protection.”

Chief constables across the country have agreed that the NDM will officially replace the Conflict Management Model (CMM), Scanning, Analysing and Responding to and Assessing Problems (SARA) and the Values-Based Decision-Making model with immediate effect.

Work has already begun on the implementation of the model, led by the ACPO Business areas and the National Policing Improvement Agency (NPIA). An e-learning package is being developed to provide quick and readily available training, which can be brief and simple, like the model itself.

And Mr Moore said the NDM will change the way officers are recruited to the force.

Related News

Select vacancies, superintendent, superintendent or chief inspector on promotion to superintendent.

Copyright © 2024 Police Professional

• Terms of Use
• About Police Professional
• Privacy Policy

Decision-Making in Law Enforcement

• First Online: 02 September 2021

Cite this chapter

• Mark Roycroft 3  

892 Accesses

The greater accountability of police decision makers is unprecedented in all areas, including the management of critical incidents. The justifiable emphasis for particular scrutiny is on the effectiveness of the decisions made in major enquires. The police service has seen a vast increase in the transparency of police decision-making and the need for police officers of all ranks to justify, record and explain their decision-making. This chapter focuses upon the challenges and responsibilities faced by police officers. This chapter reviews and explores the naturalistic decision-making practices of officers and how they prioritise risk.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Subscribe and save.

• Get 10 units per month
• Download Article/Chapter or eBook
• 1 Unit = 1 Article or 1 Chapter
• Cancel anytime
• Available as PDF
• Read on any device
• Instant download
• Own it forever
• Available as EPUB and PDF
• Compact, lightweight edition
• Dispatched in 3 to 5 business days
• Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

Similar content being viewed by others

Why Understanding Police Investigative Decision Making Is Important

History of decision-making, police officer decision-making in dispute encounters: digging deeper into the ‘black box’.

ACPO. (2006). Murder investigation manual . Wyboston: NPIA.

Google Scholar  

Ansoff, H. I. (1965). Corporate strategy . New York: McGraw Hill.

College of Policing. (2014). Authorised professional practice: National decision model . [online]. https://www.app.college.police.uk/app-content/national-decision-model/the-national-decision-model/ . Accessed 01 Sep 2018.

Cook, T., & Tattersall, A. (Eds.). (2016). Blackstone’s senior investigating officers handbook (4th ed.). Oxford: University Press.

Copple, C. K., & Copple, J. E. (2018). Risk Management in Law Enforcement: Discussions on Identifying and Mitigating Risk for Officers, Departments, and the Public . Washington, DC: Office of Community Oriented Policing Services.

Crego, J., & Harris, C. (2000). Do NDM and CDM theories mark the ends of a continuum? Or Does a hybrid approach better describe team-based decision-making problem solving processes with the management of critical incidents? Draft Paper July 2000. [online]. http://www.calt.demon.co.uk/critical%20incident%20management.htm . Accessed 06 Jan 2019.

Flin, F., & Arbuthnot, K. (2020). Ashgate Incident Command: Tales from the Hot Seat.

Irvine, B., & Dunningham, C. (1993). Human factors in the quality control of CID investigations research (Royal Commission on Criminal Justice, Research Study Number 21) . London: Home Office

Kind, S. (1987). The Scientific Investigation of Crime John Sherrat and Son Manchester.

Klein, H. J. (1989). An integrated control theory model of work motivation. The Academy of Management Review, 14 (2), 150–172. https://doi.org/10.2307/258414

Mark, R. (1978). In the office of constable . London: Collins.

Quinton, P., Bland, N., & Miller, J. (2000). Police stops, decision-making and practice (Police Research Series Paper 130) . London: Home Office

Rossmo, D. K. (2009). Criminal investigative failures . Boca Raton: CRC Press.

Roycroft, M. (2019). Decision Making in Police Enquiries and Critical Incidents . Palgrave Macmillan.

Roycroft, M., & Roach, J. (Eds.). (2019). Decision making in police enquires and critical incidents: What really works . London: Palgrave.

Schon, D. (1983). The reflective practitioner: How professionals think in action . New York: Basic Books.

Scraton, P. (2009). Hillsborough – The truth . Edinburgh: Mainstream Publishing UK.

Smith, N., & Flanagan, C. (2000). The effective detective: Identifying the skills of an effective SIO . London: Home Office.

Stelfox, P. (2009). Criminal investigation: An introduction to principles and practice . Oxon: Willan.

Turner, R. (2016). Hillsborough disaster: Five key mistakes . [online]. https://www.bbc.co.uk/news/uk-england-merseyside-35462767 . Accessed 20 Dec 2018.

Download references

Author information

Authors and affiliations.

Rabdan Academy, Al Bhustan, Abu Dhabi, United Arab Emirates

Mark Roycroft

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Mark Roycroft .

Editor information

Editors and affiliations.

Lindsey Brine

Rights and permissions

Reprints and permissions

Copyright information

© 2021 The Author(s), under exclusive license to Springer Nature Switzerland AG

About this chapter

Roycroft, M. (2021). Decision-Making in Law Enforcement. In: Roycroft, M., Brine, L. (eds) Modern Police Leadership. Palgrave Macmillan, Cham. https://doi.org/10.1007/978-3-030-63930-3_29

Download citation

DOI : https://doi.org/10.1007/978-3-030-63930-3_29

Published : 02 September 2021

Publisher Name : Palgrave Macmillan, Cham

Print ISBN : 978-3-030-63929-7

Online ISBN : 978-3-030-63930-3

eBook Packages : Law and Criminology Law and Criminology (R0)

Share this chapter

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

• Publish with us

Policies and ethics

• Find a journal
• Track your research

• The Open University
• Accessibility hub
• Guest user / Sign out
• Study with The Open University

My OpenLearn Profile

Personalise your OpenLearn profile, save your favourite content and get recognition for your learning

About this free course

Become an ou student, download this course, share this free course.

Start this free course now. Just create an account and sign in. Enrol and complete the course for a free statement of participation or digital badge if available.

4 Identifying potential solutions

4.1 decision making in policing.

In the following video, Police Officer Ben Hargreaves discusses the challenges around decision making for community safety.

While this is a powerful framework for problem solving (and has been adopted by police forces across the UK) it cannot be seen in isolation from the underlying social and psychological perspectives. Your own personal biases and preferences will have an impact on how you understand and interpret data and the way you relate to policies and ‘objective’ guidelines. Much of this will also be impacted by culture and the unspoken rules evident in organisations.

The National Decision Model

The NDM is a police framework designed to make the decision-making process easier and standardised. It should be used by all officers, decision makers and assessors who are involved in the whole decision process. Not only is it used for making decisions but to assess and judge those decisions. It can also be used to improve future decisions and help to create techniques and methods for many different situations.

The NDM is based around the police force mission statement and the Code of Ethics, which should be considered when completing each of the stages. You should ask yourself whether the action you are considering is consistent with the Code of Ethics, what the police service would expect, and what the community and the public as a whole would expect of you.

The NDM stages are:

Stage 1 Gather information about the problem in hand. Not only should you work out what you do know, but what you do not know. You will use the information gathered in stage 1 throughout the rest of the process and also when your decisions are being assessed and judged after the event.

Stage 2 Determine the threat, its nature and extent so that you can assess the situation and make the right decisions. Ask yourself, do you need to take the necessary action straight away or is this an ongoing problem? What is the most likely outcome and what would be the implications? Are the police the most appropriate people to deal with the problem, and are you best equipped to help resolve the problem at hand or would somebody else be better?

Stage 3 Knowing what the problem is, you will need to determine what powers you and the police have to combat the problem. Ask yourself which powers will be needed and if the required powers and policies need any additional or specialist assistance to be instigated and introduced. Is there any legislation that covers the process?

Stage 4 Armed with all of the information regarding the problem and any policies and other legislations that may exist, you are in a position to draw up a list of options. You should also use this opportunity to develop a contingency plan or a series of contingencies that can provide you with a backup plan if things do not go exactly to plan.

Stage 5 Once you have determined the most appropriate action, it is time to put this in place. Perform the most desirable action and, if necessary, begin the process again to get the best results possible. Review the process and determine whether or not you could have done things better and what you would do in the future if you were faced with a similar, or the same, problem.

Pentagon with text at the centre surrounded by five ovals with text. Block arrows pointing from central pentagon to each oval. Further block arrows pointing from each oval to the next in a clock-wise direction, indicating a cycle sequence.

Central pentagon text is ‘Code of Ethics’

Oval text as follows moving clockwise:

12 o’clock ‘Gather information and intelligence’

2 o’clock ‘Assess threat and risk and develop a working strategy’

5 ‘clock ‘Consider powers and policy’

7 o’clock ‘Identify options and contingencies’

10 o’clock ‘Take action and review what happened’

When it comes to policing, there are many standard decision making models to refer to such as OSARA and the National Decision Model. What they share is an attempt to put a common, objective framework on decision making efforts so that decisions can be objectively supported and justified, can be explained to colleagues and can be clearly analysed for lessons and understanding after the fact

A nation must think before it acts.

• America and the West
• Middle East
• National Security
• Central Asia
• China & Taiwan
• Expert Commentary
• Conversations
• Intern Corner
• Newsletters
• Press Contact
• Upcoming Events
• People, Politics, and Prose
• Briefings, Booktalks, and Conversations
• The Benjamin Franklin Award
• Event and Lecture Archive
• Ties That Bind: NATO at 75 and Beyond
• Chain Reaction
• Bear Market Brief
• Baltic Ways
• Report in Short
• Our Mission
• Board of Trustees
• Board of Advisors
• Research Programs
• Audited Financials
• PA Certificate of Charitable Registration
• Corporate Partnership

How National Security Decisions are Made

• Nikolas K. Gvosdev
• October 14, 2016
• National Security Program

National Security Council Meeting

Most Americans believe that they have an understanding of how the President makes important national security decisions, influenced by the depictions they see in Hollywood movies, from Dr. Strangelove to Independence Day . Experts send options and information upward, and the President convenes Cabinet officials and senior military officers who cluster in the Oval Office or the Situation Room to thrash out appropriate solutions, which are then transmitted to the responsible agencies of the U.S. government for implementation. This perspective aligns with what Robert Cutler, President Dwight D. Eisenhower’s first National Security Advisor, dubbed “Policy Hill”—that in making policy, “recommendations travel upward … where they are thrashed out and submitted to the President. When the President has approved a policy recommendation, it travels down the other side of Policy Hill to the departments and agencies responsible for its execution.” [1]

The “Policy Hill” analogy, however, rests on two unstated assumptions: first, that the President is personally interested and involved in the decision-making; and second, that the key advisors surrounding the President happen to be the heads of the relevant agencies and departments. Over the last several administrations, these assumptions have been increasingly challenged by the actual realities of how policy is made.

Even though all U.S. government decisions are made in the name of the President, in actuality, however, the President only has time and energy to make a fraction of the national security decisions in a personal capacity. As Ben Heineman points out, “a President can have only five to 10 top priorities on which he makes virtually all decisions. He can set general direction for perhaps 25 secondary priorities.” Yet, the Constitution makes no provision for a collective Presidency or for Presidential deputies to share the burden of Executive authority. At the same time, as another former National Security Advisor, Robert C. “Bud” McFarlane has noted, there is a strong preference not to have to bring “minor matters”—especially the whole host of day-to-day, routine issues—to the President for resolution and authorization. [2] The challenge, as yet a third former National Security Advisor, Jim Jones , has observed, is that there must also be a way to determine whether a matter needs to be brought personally to the President for his or her intervention, since “not everything has to go to the President for a decision” while, at the same time, ensuring a high degree of fidelity “with the President’s wishes” in terms of policy outcomes.

So far, so good. There is no intrinsic contradiction between a pyramidal organizational chart that stipulates that all national security decisions are made in the name of the President and a system that allows for the Chief Executive to transfer authority for routine decision-making further down the duly-constituted chain of command. Where the question gets far murkier is what happens when a President opts for what is termed by Robert Durant and Paul Diehl a “counterbureaucratic” solution—that is, assigning policymaking functions to units within the White House staff rather than the permanent departments and agencies of the executive branch. [3] Increasingly, over the last several decades, the “counterbureaucratic” approach has steadily gained in popularity in successive administrations. Since the Presidential staff, in theory, can coordinate policy across the government and thus not be captured by any one agency or department’s interests, instead will focus on what the President wants to achieve. [4] Presidents also like the appeal of centralizing national security oversight within the White House staff because, as Susan Rice , President Barack Obama’s current National Security Advisor, has complained, leaving matters to the bureaucracy meant that policy “wasn’t working until we sucked it into the White House and the President put his personal muscle behind it.”

But what this results in is an increasing shift away from the formal officers charged with overseeing the various parts of the executive branch entrusted with national security matters towards a White House circle of advisors and staff. It is no longer automatic that the Secretaries of State and Defense will be “in the room” when the President makes a final decision, and there have been growing concerns that recent administrations have kept senior Cabinet officials, as well as their departments, in the dark about critical national security initiatives started by the White House and managed by the President’s personal staff.

Thus, when considering what sort of Presidency a future Hillary Clinton or Donald Trump administration will produce, it is becoming more important to introduce what we have termed the “ palace politics ” perspective. This places a great deal of emphasis on understanding who has the most influence with the President and therefore focuses on who has access (and who does not); who serves as a gate-keeper (and which views are kept out); whose judgment and perspective the Chief Executive trusts (and whose he or she does not).   In the past, those individuals might have been the Senate-confirmed officials at the helm of the Departments of State and Defense, but in more recent years, individuals who have little or no statutory authority or bureaucratic base have emerged as the principal members of the immediate entourage around the President.  Indeed, the role of key figures in the White House serving as Presidential “gatekeepers” can be a subject of concern for Cabinet officers because West Wing staff can block access to the President and prevent those officers from offering advice and counsel on policy. The “palace politics” perspective cautions that formal titles and places on an organization chart may not carry as much weight in understanding who influences policy decisions. Former Secretary of Defense Leon Panetta offers this observation: “You have a lot more staff people running around at the White House on these foreign policy issues. And very frankly, proximity is everything, when you’re operating at the White House. The person that’s closest to the president has greater influence than even a cabinet member, who may be located elsewhere in a department.”

The Presidential staff occupies an ambiguous position with regards to national security decision making. They enjoy the virtues of proximity to the “sole organ”—the President—but are not part of the chain of command. White House staff requests are transmitted in the name of the President (even if done so without the President’s own explicit knowledge or approval) by people not in the formal chain yet who are steps away from the Chief Executive. Cabinet officers may fume at this—and even, as Don Rumsfeld or Bob Gates did, issue orders to their subordinates to ignore tasks from the White House staff—but nevertheless, senior officials increasingly must deal with the reality, expressed by the unnamed senior sub-cabinet State Department official who fumed to Washington Post reporter Karen de Young , that “any little twerp from the NSC can call a meeting and set the agenda” without regard to rank or experience—and thus control and influence how policy is shaped.

There is nothing to indicate that either of the two Presidential candidates will change this dynamic. For different reasons, both Clinton and Trump will have strong predilections to continue to weight and prioritize the influence of the “palace” over that of the bureaucracy. As I noted in a recent observation for the National Interest :

For Trump, the #NeverTrump movement forces the candidate both to search elsewhere for expertise and to distrust the “establishment” as uncommitted to his agenda. A President Trump and his close associates could very easily and eagerly fall into the temptation that former national security advisor Brent Scrowcroft warns against: “not to replace departments. . . . That’s always the instinctive thing—well, ‘These guys aren’t doing a good job on something, we’ll just do it ourselves.’” This tendency could be further reinforced if a President Trump finds that a hostile or suspicious Senate is not prepared to confirm his preferred nominees for key positions. If officials are “forced” upon him in this way, why wouldn’t Trump and his team simply continue the trend to downgrade the importance of the cabinet and subcabinet in favor of retaining a larger White House operation?

Hillary Clinton is in a different position, with thousands of Democratic staff and experts (not to mention some Republicans as well) all eager to offer their services in a Clinton administration. Yet Clinton has consistently shown a preference for managing via a small circle of dedicated loyalists who, in turn, act as her gatekeepers and enforcers. The consistent reconstruction of the “Hillaryland” structure—from the White House in the 1990s to her Senate office in the 2000s to Foggy Bottom in the first term of the Obama administration—all strongly suggests that no matter the formal setup of National Security Council and cabinet meetings, the real focal point of decisionmaking will remain that inner circle.

Either Clinton or Trump will inherit a large Presidential apparatus, given the dramatic increase in the size of the National Security staff, from the some one hundred personnel that Bill Clinton relied upon in growing the size of the leaner George H. W. Bush staff of fifty, to the two hundred slots bequeathed by George W. Bush to Barack Obama, to the more than four hundred today (even with some recent trimming by National Security Advisor Susan Rice).  This increase in staff size has occurred because of the suspicions held by previous Chief Executives that the national security bureaucracy “is doing too much and going beyond what the president has decided, or that it is doing too little and not fulfilling what the president wants done.” [5]  

One way forward may be for a President Trump or Clinton to continue to delegate authorities to the staff as a way to develop new linkages between the White House and the bureaucracy. Early in the first term of his administration, President Obama decided that White House/NSC staff, not departmental officials, would chair the interagency policy committees (IPCs) where the “meat and potatoes” of much of U.S. national security and foreign policy is made, thus giving the staff formal rights to convene meetings and request briefing materials and position papers from the different departmental and agency representatives. [6] Expansion of the special envoy/White House coordinator/“czar” model may also hold promise for reconciling a Presidential desire to maintain control over the policy process on key issues in the White House by creating formal channels to connect the White House staff to the permanent bureaucracy and officers in the field. This model ensures that both the vertical reporting chain (within a particular organization) and the horizontal one (connecting to the White House) are engaged at the same time. [7] This might avoid the issues that Rumsfeld and Gates raised about information being solicited from DoD by the President’s staff without their consent or knowledge of what was being shared.

Faced with the reality of the rivalry between the West Wing and the departments as well as staff politics within the White House, it is easy to yearn for a return to the simplicity of the George Washington or even the Dwight D. Eisenhower administration by radically paring down the size of the White House staff and returning to a reliance on the Cabinet officers who are in the chain of command to serve as the principal advisers and staff associates to a President.

Two decades ago, Samuel Kernell, in assessing these claims, argued why they were impractical: “To assume that the governability of the modern presidency simply required reducing the staff and returning it to its original size and import would be to ignore the pervasive environmental forces that have transformed presidential leadership.” The end result, he argued, would be to leave the President “incapacitated.” [8] Even if subsequent administrations make some reduction in the size of the White House operation and fight against the urge to centralize the national security policy process, Dan Drezner warns, “This is a problem that will not go away.” The President will remain in charge—but the struggle over who acts as the President’s agent and who has the President’s ear will continue.

[1] Robert Cutler, “The Development of the National Security Council,” Foreign Affairs 34:3 (April 1956), 448.

[2] Robert C. McFarlane with Richard Saunders and Thomas C. Shull, “The National Security Council: Organization for Policy-Making,” in The Presidency and National Security Policy , ed. R. Gordon Hoxie (New York: Center for the Study of the Presidency, 1984), 266.

[3] Robert F. Durant and Paul F. Diehl, “Agendas, Alternatives and Public Policy: Lessons from the U.S. Foreign Policy Agenda,” Journal of Public Policy 9:2 (April 1989), 183.

[4] William W. Newmann, “Reorganizing for National Security and Homeland Security,” Public Administration Review 62:S1 (September 2002), 130.

[5] Derek Chollet, The Long Game: How Obama Defied Washington and Redefined America’s Role in the World (New York: PublicAffairs, 2016), 83.

[6] Paul D. Miller, “Organizing the National Security Council: I Like Ike’s,” Presidential Studies Quarterly 43:3 (September 2013), 595, see also Alan Whittaker, Shannon Brown, Frederick Smith and Elizabeth McCune, The National Security Policy Process: The National Security Council and Interagency System (Washington, D.C.: Industrial College of the Armed Forces, National Defense University, 2011), 35.

[7] See, for instance, some of the comments on the operation of the Special Representative for Afghanistan and Pakistan, in United States Department of State and the Broadcasting Board of Governors, Office of Inspector General, Inspection of the Office of the United States Special Representative for Afghanistan and Pakistan, Report Number ISP-I-11-48 (June 2011), 1, 6, 17. 

[8] Samuel Kernell, “The Evolution of the White House Staff,” Can the Government Govern? eds. John E. Chubb and Paul Peterson (Washington, DC: The Brookings Institution, 1989), 235.

You May Also be Interested in

The presidential candidates on democracy promotion in the middle east.

The past two presidents have been deeply engaged in promoting democracy in the Middle East. After the terrorist...

When Border Walls Work (and Don’t): An Open Memorandum to Mr. Trump

    “A nation without borders is not a nation. There must be a wall across the southern border.”Statement...

Will China decide to reduce tension in the South China Sea?

On May 19, The Straits Times published an article written by Xu Bu, China’s Ambassador to Asean, that criticises...

• Search Menu
• Sign in through your institution
• Author Guidelines
• Why Publish
• Submission Site
• Reviewer guidelines
• Open Access
• About Policing
• Editorial Board
• Advertising and Corporate Services
• Journals Career Network
• Dispatch Dates
• Journals on Oxford Academic
• Books on Oxford Academic

Article Contents

Introduction, discussion: autonomy, professionalism, and phronetic policing.

• < Previous

Professional policing and the role of autonomy and discretion in decision-making: A qualitative study from a virtue ethical perspective

• Article contents
• Figures & tables
• Supplementary Data

Andrew Maile, Sarah Ritzenthaler, Aidan Thompson, Kristján Kristjánsson, Professional policing and the role of autonomy and discretion in decision-making: A qualitative study from a virtue ethical perspective, Policing: A Journal of Policy and Practice , Volume 17, 2023, paac086, https://doi.org/10.1093/police/paac086

• Permissions Icon Permissions

Within an overarching identified construct of ‘autonomy and discretion’, this paper uses four themes to discuss how 30 interviewed UK police officers respond to challenging, ethically charged situations and what frameworks guide and structure their decision-making. These themes, elicited through qualitative thematic analysis, are personal judgement, doing the right thing, effective and adaptive communication, and emotional regulation. The relevance of these themes is discussed in the context of police professionalization and the Code of Ethics in England and Wales. The findings illustrate that the professional autonomy of police officers, when grounded in virtue ethics rather than more formulaic deontology, supports morally good policing and safeguards police legitimacy.

Considering the recent challenges that police forces have been facing to their legitimacy and socio-moral standing, the overall aim of the qualitative study reported in this paper was to gauge the extent to which professional character enables police officers to carry out their job effectively and with a sense of purpose. The qualitative study set out to explore what features predominated in the decision-making processes of 30 UK 1 police officers. A qualitative analysis of extended interviews with these police officers was conducted as part of a larger study of character virtues in UK policing that took place in 2021, with the aim of better understanding the moral practice of policing ( Kristjánsson et al ., 2021 ). To place this study in a theoretical context, something must be said first about the perennial legitimacy concern, hanging historically like the Sword of Damocles over the practice of policing.

Policing, as a public-facing occupation, has long been held to require public consent for the execution of its powers and responsibility: something which Bottoms and Tankebe (2012) refer to as ‘audience-legitimacy’, as understood within the normative model of procedural justice defined by Tyler (2006) . The role of ‘self-legitimacy’, whereby police officers seek to understand the moral validity of their power, and deploy this power in a way that is morally right, has also been identified as fundamental to the establishment and maintenance of police legitimacy ( Bottoms and Tankebe, 2013 ; Tankebe, 2019 ). In this view, both individuals’ ‘personal morality’, namely the general set of beliefs about how one should act, and their perceptions of police ‘legitimacy’, understood as the authority of police decision-making, need to be satisfied for procedural justice and police legitimacy to be secured. As Tyler concludes, ‘people comply with the law not so much because they fear punishment as because they feel that legal authorities are legitimate and that their actions are generally fair’ (1990, cited in Bottoms and Tankebe, 2012 :120).

As such, it can thus be argued that policing requires the trust and confidence of the public, a case which has again been made in a recent independent review of the Metropolitan Police Service ( Independent Office for Police Conduct, 2022 ). One way in which public trust and confidence can be gained (or maintained) by the police is through the execution of their roles in a moral and professional manner, displaying a moral alignment with the views and values of the general public ( Manning, 2019 ). Yet the evidence of unethical practice and professional scandals appear far too frequently in the media, inviting members of the public to scrutinize negatively, and continue to doubt, the legitimacy of the police.

One possible response to such morally questionable and unprofessional behaviour is to implement stricter rules and sanctions. However, if policing in England and Wales is to live up to the expectations placed upon it by the recent professionalization agenda, alternatives to increased formal limitations on how police officers can go about their work ought to be considered. In this regard, Neyroud and Beckley (2001 :220) have proposed a much-needed ‘renewal of the contract between the police officer and the citizen’ in 21 st -century policing, which, in turn, requires ‘a new commitment to ethics at the core of policing’. This commitment to placing ethics at the core of policing manifests itself, at least in part, with the professionalization agenda of policing in England and Wales.

The current study forms part of a series of studies exploring the role of characterological and virtue ethical conditions in shaping good ethical decision-making and a sense of professional purpose in various UK professions ( Arthur and Earl, 2020 ). Previous studies have indicated that these characterological conditions are best exemplified through the development and exhibition of intellectual, moral, civic, and performance strengths of character (also known as virtues) that enable the professional to act wisely in a given situation by making good decisions for the right reasons and practise one’s profession in an ethically salient manner. A core finding in previous studies has been about the importance of developing the meta-virtue of phronesis (practical wisdom) in one’s pre- and in-service professional training. A more general concern motivating all these studies, including the current one, has been the recent turn in professional ethics, including police ethics ( Morrell and Bradford, 2019 ; Wood, 2020 ), away from a compliance-based, rule-following ethical outlook (deontology) towards an outlook that foregrounds individual and collective moral character (virtue ethics). We say more about those outlooks below.

The designation of the term ‘professional’ to policing practice is particularly contested, dating back to at least the middle of the 20th century ( Critchley, 1967 ; Mark, 1977 ). Spurred on by the report on the training and leadership of police officers by Neyroud (2011) , the professionalization agenda progressed with the establishment of the College of Policing in 2013—which serves as the professional body for policing in England and Wales—seeking inter alia to tackle corruption and address low levels of public trust and confidence in the police through the setting of standards for good professional practice ( Neyroud, 2011 ). Additional processes were implemented, including changes to the training and recruitment of police officers via the newly established Police Education Qualification Framework (PEQF), replacing vocational training programmes with undergraduate degrees and undergraduate and postgraduate degree apprenticeships ( Hunter and May, 2019 ; Williams et al. , 2019 ).

The professionalization of policing in England and Wales is a topic that has received much attention and debate, with some suggesting it should be considered ‘re-professionalization’, while others remain unconvinced of its attribution to the role of a police officer ( Holdaway, 2017 ; Neyroud, 2011 ; Williams et al. , 2019 ). Typically referring to the process by which an occupation gains recognition as a profession, ‘professionalization’ also refers to the subsequent improvement of service delivery—which, based on recent policing controversies, is something many would support ( Her Majesty’s Inspectorate of Constabulary, 2022 ; Independent Office for Police Conduct, 2022 ). From a sociological perspective, the term ‘profession’ is typically applied to a particular group of traditional occupations that are associated with a certain status and privilege. The term ‘professional’ can also define the aspects of an occupation that mark it as a moral practice , with the term ‘unprofessional’ signifying a failure to meet the moral standards of behaviour and conduct expected of an occupation ( Jubilee Centre, 2016 ).

In contrast to the term ‘professional’, policing has also been described as a ‘craft’ or ‘artisan’ role, distinguishing it from scientific and authoritative knowledge associated with other, often elitist, professions and reaffirming its ‘citizen in uniform’ conceptualization ( Neyroud, 2011 ). Rowe and colleagues (2016) unpack the ‘craft’ nature of policing, and suggest that while it is possible to identify ‘craft’ aspects of policing, this does not mean it cannot be considered a profession. Rather, they draw into focus the applied and practical aspects of policing that are typically beyond the scope of a degree programme ( Rowe et al ., 2016 : 284). As will be seen throughout the following, reference to policing as a profession does not preclude the ‘craft’ aspects of the role. Accordingly, the importance of peer learning, which may be considered a ‘craft’ aspect, appears in the empirical findings below.

The present research builds on the Jubilee Centre for Character and Virtues (2016) neo-Aristotelian conception of moral professional practice, in the virtue ethics tradition, seeing good professional decision-making as rooted in wise moral reasoning, where decisions are made for the right reasons, utilizing one’s own strengths of character, whilst seeking to benefit those around us. This differs from the other two main moral theories of deontology, emphasizing rules-based reasoning, and utilitarianism or consequentialism, emphasizing good-outcome-maximizing reasoning, as further explained below.

While debates about whether it is appropriate for policing to be regarded as a profession remain, for the purpose of this paper this designation of policing is accepted. This is supported both by the ongoing professionalization agenda, spearheaded by the College of Policing, and by how the following five generally accepted features of a profession are (at least in part) met by the police ( Carr, 1999 ):

First, an ethical dimension of the profession is captured by a formalized code of practice, which for policing occurred with the publication of the Code of Ethics 2 [hereafter: the Code ] by the College of Policing (2014) . 3

Second, the profession requires both theoretically and practically grounded expertise. It could be argued that the publication of the PEQF and the subsequent degree-level requirements for entrance into policing meet this criterion, outlining the required practical and theoretical learning for accredited universities to confer policing degrees. The PEQF also presents the opportunity for policing to regain professional authority with a better understanding of its role in society, and a chance to respond to societal demands of police officers in ways that are thoughtful, meaningful, and appropriate ( Wood, 2020 : 376). Additionally, the College of Policing is the formal body of policing expertise, offering guidance through its ‘Authorized Professional Practice’ (see https://www.college.police.uk/app ).

Third, the profession provides an important public service. This characterization is met in policing, which is viewed as both a ‘customer-based’ and ‘public’ service ( Neyroud, 2006 ).

Fourth, there are organizational regulations for the recruitment and discipline of police officers. This criterion is met in both the PEQF requirements for policing, and the Standards of Professional Behaviour for police officers, which includes guidance on dealing with misconduct and the Police Appeals Tribunal that are issued as part of the 1996 Police Act ( Home Office, 2014 ).

The fifth and final criterion is that professions require a high degree of judgement and autonomy in the execution of their roles. It is this criterion, in particular, that the findings from the in-depth qualitative work presented in this paper will be addressing.

The five professional criteria considered above indicate that while professionals require a foundation in theoretical and practical expertise, they must necessarily function as autonomous individuals who act upon their own decisions, informed by theory and experience. In order to fully understand the extent to which autonomy is conducive to the role in question, it is necessary to unpack the different approaches to professional policing ethics, that is, virtue ethics, deontology, and utilitarianism. The present research adopts a virtue ethical stance in seeing policing as a moral practice, requiring certain moral, civic, intellectual, and performative virtues of its practitioners. It views the Code through such a lens and explores what the Code promotes in support of a moral approach to policing.

To rehearse briefly the theories in question, virtue ethics is one of the three leading ethical theories competing for allegiance within both ethical theorizing in general and professional ethics in particular. Virtue ethics defines moral rightness according to the effect it has on the agent, in terms of the extent to which it supports the agent to be virtuous and lead a well-rounded flourishing life ( Aristotle, 1985 ). Consequentialism considers the moral rightness of an action in terms of the consequences it has for the maximization of overall human happiness ( Mill, 1863 ). In contrast, moral rightness from a deontological perspective is defined in terms of adherence to universal, rationally grounded principles, or the maxims (in the form of rules or codes) derived from them ( Kant, 1964 ). To give a simple example, virtue ethics would justify the value of procedural justice in terms of its manifestation of the generally acknowledged virtue of fairness and its intrinsic value for a well-rounded life. Consequentialism would justify procedural justice in terms of its extrinsic/instrumental long-term benefits for the whole of humankind; and deontology would consider justice a rational principle, manifested in written laws, the violation of which counts both as a breach of practical reason and of institutionally sanctioned rules.

Different features of police ethics are amenable to consequentialist as well as deontological interpretations; hence, the enduring appeal of these approaches within police ethics. For example, to the extent that policing involves enforcing particular laws in society—as well as upholding the ‘rule of law’ in an abstract sense—this requires the enactment of principles, which takes us back to the deontological territory. Yet concerns have been raised regarding an understanding of police ethics that equates it to the mere adherence to principles and rules. First, considering the complexity of judgements required of professionals, it is almost impossible to find general rules that are applicable in all situations. Second, when professional conduct is defined in terms of rules and protocols, this can result in a mindset that is unreflective or conformist, discouraging the kind of autonomy and initiative expected from professionals. Further, this may even serve as a disincentive for moral reasoning, as professionals discern the inevitable ‘shades of grey’ in complex situations. Third, a carrot-and-stick method of increasing rules and sanctions within practices has proven to be ineffective in professional work contexts ( Schwartz and Sharpe, 2010 ; Jubilee Centre, 2016; Wood, 2020 ).

This explains why, in recent theorizing about those features within police ethics, the pendulum has been swinging away from deontology and towards virtue ethics ( Manning, 2019 ; Wood, 2020 ). Moreover, virtue ethics is more sensitive to the motivational make-up of the ethical agent/professional. Thus, according to virtue ethics, someone may be judged unfit for professional practice if their possession of theoretical and practical expertise is not complemented with appropriate character traits, attitudes, and motivations ( Carr, 2000 ).

Modern policing involves much more than law enforcement, application of police-force policies or apprehending criminals for it to be considered good policing ( Wood, 2020 : 4). Indeed, a considerable portion of the work requires ‘proactive policing’, which involves mediation and quelling of conflicts before they escalate ( Wood, 2020 : 39). Oftentimes, this does not qualify as a reaction to any legal violations, as there has been no breach of the law. Rather, it requires creating in individuals conditions of character that are conducive to maintaining the ‘Queen’s Peace’. In that sense, some aspects of police work, at least, already have a strong virtue ethical dimension built into them.

Stemming from this virtue ethical tradition is the neo-Aristotelian conception of phronesis , the intellectual meta-virtue and meta-cognitive capacity for contextual, holistic, integrative adjudication and practical reflection about moral issues, which leads the professional (in this case, police officer) towards moral action ( Kristjánsson, 2015 ). Importantly, phronesis enables the individual to engage in reflection and adjudication about moral issues in a way that is non-codifiable: it cannot be prescribed by rules, nor is it amenable to an algorithmic style of decision-making. Rather, it requires discernment, discretion, and deliberation, and relies upon expert knowledge and experience. Wood (2020 : 5) describes ‘the ability of an officer to make sense of competing demands within different contexts, as opposed to following ethical rules of one kind or another’, and it is this ‘ability’, which in the virtue ethical literature is typically referred to as (professional) phronesis .

Virtue ethics and the police Code of Ethics

Returning to the issue of different understandings of procedural justice, Hough (2021) warns that—for some police officers—the principles of procedural justice may simply be viewed as a method, or set of tactics, for ensuring that the public complies with the law. In this way, such principles may merely be used ‘as a series of behavioural tricks—or social skills—which need to be acquired in order to deliver effective policing’ ( Hough, 2021 : 87). It is precisely here where a virtue ethical approach, serving as a foundation of the moral actions of a police officer, would be distinct from one informed by deontology or consequentialism. The recent ethical re-focus of policing in England and Wales promoted by Neyroud and Beckley (2001) , and evidenced by changes such as the publication of the Code, ought to be commended for the emphasis it places on ethics sitting at the centre of police decisions and actions.

Developed by the College of Policing, the Code was published in 2014 and is one of the important outcomes of the professionalization agenda, and the ethical focus of policing. Codes of this kind have usually been inspired historically by the ethical theory of deontology: adherence to discrete rules and principles. This Code , however, is unique when compared to other professional codes of ethics (with the exception of the civil service) in apparently being informed by insights from virtue ethics. The Code clearly indicates that police officers’ ethical behaviour ought to stem from personal values, reflected in the Code , that guide their conduct. The nine values in the Code— referred to as ‘principles’ (indicating a historical deontological provenance)—are ‘selflessness’, ‘honesty’, ‘openness’, ‘respect’, ‘accountability’, ‘fairness’, ‘objectivity’, ‘integrity’, and ‘leadership’. With the exception of ‘accountability’, these values are, from a virtue ethical perspective, considered to be intellectual and moral virtues : namely strengths of character, rather than simply principles or values. The Code is supplemented by the National Decision Model (NDM), a five-step flexible guide for navigating ethical decision-making in policing.

It is possible to be either sympathetic or unsympathetic when interpreting the Code from a virtue ethical perspective. Two of three recent works promoting a virtue ethical take on UK police ethics take the more critical, unsympathetic view ( Manning, 2019 ; Wood, 2020 ). Morrell and Bradford (2019) do not discuss the Code , however. Wood (2020) worries that the Code reduces ethical decision-making to models and mnemonics, giving officers the impression that ethical decision-making can be learned simply by following the given principles. There is, thus, ‘a danger that it fosters unthinking compliance rather than intelligent, ethical reasoning’ (2020: 25), revealing concern that the Code reflects more of a deontological alignment, and suggesting that the promotion of professional autonomy needs to be made more explicit. Manning’s verdicts on the Code are even harsher. He worries, like Wood, that the Code seems to ‘imply that moral conduct will naturally follow’ from its publication (2019: 16); and he rues the absence of any explicit ethical theoretical grounding (2019: 118). In this respect, Manning (in particular, and corroborated by Williams, 2021 and Wood, 2020 ) indicates dissatisfaction with how the Code was implemented—an important consideration for future iterations of the Code, its implementation and subsequent reception among the police. He further argues ‘that the publication of the [Code] falls short of providing an exemplary framework for police officers with appropriate guidance and training for them’ (2019: 17).

Manning produces evidence from his own interviews with police officers and other sources indicating that the Code has had scarcely any effect on actual decision-making within the police; that it is badly taught, if at all, to budding police constables; and that they barely understand it or just refer to it flippantly (2019: 208–209); a finding that is largely mirrored by Williams (2021) . Yet on a more positive note, Hough writes: ‘The College of Policing has struck a sensible balance between issues of ethics and those of morality in its code’ (2021: 89)—where the word ‘ethics’ clearly refers to a deontological understanding but ‘morality’ to a characterological or virtue ethical one. Hough does point out, however, that more guidance could be provided in terms of explicating what is expected in terms of adherence to the principles/virtues. His critique also seems to suggest that the ethical theory underpinning the principles ought to be made more explicit, which again finds corroboration in the recent work by Williams (2021) .

It is no novelty to see virtue ethically minded theorists speak of ethical codes with opprobrium. Such codes, after all, tend to have a deontological provenance; and even when understood as a broad set of simple techniques used to solve challenging problems, they rarely seem to have much direct practical value in terms of guiding conduct. Constructing an ethical code is often little more than a tick-box exercise to satisfy quasi-legal requirements. Importantly, though the police Code is not defined in terms of rules and protocols, and thus it is not at risk of promoting unreflective or conformist mindsets, nor does it discourage professional autonomy and independent moral reasoning. When viewed through the lens of virtue ethics, the Code sets out the moral mission of policing, and while failing as a vehicle of codifiable decision-making (as all ethical codes will do), the message conveyed in it appears conducive, rather than inimical, to the cultivation of reflective police professionals.

Hough (2021) raises two important questions in this regard: (i) whether police officers should be promoted and recruited based on their character or on their competencies, and (ii) whether the Code simply requires police officers to perform the virtues (or principles) listed, such as respect and fairness, or whether they be required to internalize them. In Hough’s view, ethical deontologically inspired behaviour and moral/virtuous action may be distinct from one another. What we advocate for in this paper, however, is that this distinction is not a clear-cut one, and that ethical behaviour (i.e. doing the right/prescribed thing) can ideally be morally informed (i.e. doing it for the right/virtuous reason). While not explicitly indicating a virtue ethical foundation, the classic virtue ethical formulation of ‘doing the right thing for the right reason’ does feature in the Code , as well as references to making ‘wise decisions’, and ensuring that moral virtues such as empathy and compassion guide police officers in their interactions with the public ( College of Policing, 2014 ). The ‘wisdom’ employed in decision-making comes, according to a virtue ethical outlook, through the cultivation of phronesis (practical wisdom)—excellence in moral decision-making—which the present study explores below (cf. Darnell et al. , 2019 ). What will be revealed in the discussion about the emergent themes is that these two aspects (i.e. the ethical and the moral) need not be viewed separately, but rather come together complementarily under the overarching theme of ‘autonomy and discretion’.

The present qualitative study stems from a larger mixed-method project, Virtues in Policing ( Kristjánsson et al ., 2021 ). Semi-structured interviews were conducted with 30 UK police officers with varied operational and managerial responsibilities (two of whom had retired) between 19 February 2021 and 29 April 2021.

Recruitment of participants was facilitated through the Ethics Leads at each of the 43 police forces in all four regions of the UK, via the distribution of an optional survey. At the end of the survey, respondents were asked if they would be interested in participating in a follow-up interview. Three hundred and thirty-five police officers completed the survey, of which 137 indicated a willingness to be interviewed, from which the final cohort of 30 was selected based on purposive sampling. Interviews lasted approximately 1 hour, were conducted online via videoconferencing, and then were recorded and transcribed. Biographic data of interview participants and aspects related to their work as police officers is captured in Table 1 .

Biographic and work-related data of interview participants

a The urban/rural/mixed location identifier applies to the location in which the police officer was working at the time of their participation. It is important to bear in mind that many of the officers included in the study had previously worked in other locations, including more rural locations as well as more urban locations, such as the Metropolitan Police Service.

b Years of Service.

Interview questions related to the following overall research themes:

1) What the officers’ journeys were to becoming police officers, their inspiration to pursue the profession, and the kind of qualities they aspire(d) to show in their work (both at the beginning of their career and at present).

2) What they consider to be their most important character strengths and qualities, and how these qualities influence their work.

3) How useful they find the Code and to what extent it informs their actions.

4) Any challenges or obstacles they have had (or currently experience) in being able to display the qualities and character strengths they consider to be important in the role.

5) If they had any recommendations to make regarding the education and preparation of pre-service police officers, particularly in relation to the professionalization of policing.

Data were analysed using inductive thematic analysis, which enabled the identification and analysis of the key emergent themes ( Braun and Clarke, 2006 ). The analysis was undertaken by two researchers with the use of Nvivo, and began with a purely inductive process prior to engaging with the literature related to the identified themes. The researchers conducting the recursive analysis met frequently to discuss the prominent identified themes, before agreeing on a set of ‘priority’ themes, including main and subcategories to look out for, which fitted coherently with the overarching construct of autonomy and discretion. This was undertaken following a similar process to that proposed by Braun and Clarke (2006) . All findings discussed below that might link to the identities of individuals or forces have been anonymized.

The study was given full ethical approval by the University of Birmingham Ethics Committee.

Limitations

The sampling of participants was reliant on the use of gatekeepers at each force and the voluntary nature of participants who indicated their willingness to be interviewed. Self-selection bias cannot be avoided in a study like this; however, the researchers note that the explicit purpose of the research relating to policing as a moral practice, may have meant that police officers who identify positively with the virtuous aspect of policing were more likely to register interest in being interviewed. Additionally, due to the self-report nature of interviewing, it is possible that responses may be subject to the problems of (i) social-desirability bias, whereby participants respond so as to be viewed favourably; (ii) self-confirmation bias, whereby participants uphold their prior beliefs and discard anything that contradicts those beliefs; and (iii) self-deception bias, whereby self-representations do not cohere with reality.

The overarching theme that emerged from the analysed data related to the value, need, and purpose of police autonomy and discretion. Within this overarching theme, four subthemes emerged: personal judgement, doing the right thing, emotional regulation, and effective and adaptive communication. There was overlap between the themes, and an indication that themes reinforce one another; however, for the sake of clarity, quotes that reveal this overlap have not been included.

Personal judgement

All professions require an element of personal judgement, and findings in our interviews revealed that the capacity of police officers to exercise personal judgement—based on their expert knowledge, evaluation of laws and policies, and with sensitivity to the specific context—was of crucial importance to them, in order to do their job well. Many emphasized that the ability to adjudicate between competing priorities was central to the execution of their duty, and that in some circumstances, going against the accepted norm, or even employing an unconventional approach, had been key to diffusing a challenging situation. This indicates a virtue ethical ‘ phronetic ’ orientation over a deontological one.

As one interviewee described the role:

… the fact is, as a police officer you are your own boss. Yes, there’s a hierarchy of a rank... [but] when you step outside that station and step into the community, it’s up to you whether you stop that car, whether you stop that person, how you deal with that person. (A25)

And another:

…the overriding [quality] which I think a police officer needs is judgement in knowing what is appropriate… about when and how the law should be applied and it’s a very fine line. (S3) … before the Code of Ethics …there was very rarely a hard and fast rule about what you should do here. I suppose a lot of it was using good judgement, being prudent when it called for it, being brave when it called for it and just having that sense of judgement about what was appropriate in a given set of circumstances. (S3)

The emphasis on being personally responsible and accountable for their own decisions was evident across the interview cohort. To illustrate: ‘could I stand in front of the person that’s going to be most adversely affected by it and explain why I’ve done it?’ (A9). Thus, it is not a matter of blindly making personal judgements, but doing so in a way that incorporates their vast experience, built upon self-reflection and imitation of role models. As one interviewee commented, ‘…can you justify your actions? The answer must always be ‘yes’’ (A1).

Through the experience of working with and for different individuals, police officers described how they were able to develop their own constructs of appropriate behaviour, both towards the public and in relation to other colleagues; being creative, and learning from others:

We had a big stand up at a nightclub one night, and the lad next to me got out of his pocket a white tatty bag full of mint imperials, thrust it forward to this lad that was going to kick off and said ‘you want a mint?’ … and the other one said ‘yeah, go on then’ and that was it, situation diffused, everybody happy… (A1)

For others, this translated to a nuanced understanding of more formal structures of decision-making, such as the Code :

…there’s nothing written in the Code of Ethics that doesn’t pass through my social filter already. … And it’s not there to measure how well we’ve done, it’s there to measure somebody who’s done poorly. (A27)

More experienced police officers referred not to the Code as their primary reference for decision-making, but rather their own moral code, shaped by their experiences in the job and the examples set by role models: ‘If you don’t believe in those things, you shouldn’t be a copper’ (A3). Once again, references to a personal moral compass and the salience of moral role models are standard virtue ethical indications.

Doing the right thing

Participants were consistent in their presentation of exercising personal judgement in the pursuit of ‘doing the right thing’ and with the goal of ‘making the right decision’. Most often, the desire to do the right thing stemmed from a deeper sense of professional purpose and meaning within the role:

I come back to who am I doing this job for? Am I doing this job for senior managers or the government? No, I’m doing this job for the victims. So, what is the right outcome for the victims? (A3)

The right outcome, however, is not something that participants found easy to determine:

… I think sometimes that’s really, really difficult to articulate what that is… What is the right thing to do there? (A9) Doing the right thing isn’t always easy. In fact, most of the time doing the right thing is hard. That’s where the bravery comes in… (A7)

Participants frequent reference to ensuring they do the right thing reflects the complexity of policing; when knowing the right thing to do is neither easy nor obvious, but is always underpinned by good and well-meaning intentions. The challenges surrounding being able to make ‘the right decision’ are linked to the recognition of context sensitivity, as well as balancing competing demands and available options based on knowledge and experience, and how these can be applied in different situations. This requires ‘problem solving and analytical decision making’ (A17), even with less complicated cases:

We deal with a lot of grey areas, and sometimes arrest and court isn’t suitable for the officer, the victim or the offender…It’s great being able to kind of deviate and think outside the box, given most officers will be able to do that and they really excel in doing it, choosing the right thing for the right people. Having the ability to do that is massive, absolutely massive for victims and offenders and the whole system… (A26)

Many participants also mentioned the role of reflection, both pro-active and retro-active, in their efforts to ensure that they were doing the right thing and ensuring that they can learn from lessons and mistakes of the past:

...You think afterwards ‘How did that go?’... ‘How did I come across? How did the profession come across? And how could we [have] done it better, explained it better, or just been better for that person in general?’ (A13)

The questions raised in the quote above indicate what informs much of the participants’ thinking around mistakes, which typically reveal concerns that the outcome of the interaction might not reflect the good intentions of the officer, or the general desire of providing a service that aims to help and support vulnerable individuals and communities. To this extent, participants spoke about the value of having colleagues with whom they could discuss ideas and invite friendly challenge to their ideas and solutions:

You’re trying to weigh all this up, it can be quite a difficult decision, and again having the conversation helps, bouncing those ideas off, those ethical decisions off. I don’t think you can undervalue the impact of or the benefit of seeking wider views around stuff and getting a different perspective on it. (A2)

Indeed, when collegial support structures and networks are not in place, this can serve as an obstacle to deliberating between different options and finding the right outcome: ‘I think the only real things that stop you are the lack of support from higher up...’ (A26)

Personal reflections from making and learning from mistakes were described as profound, and often impacted the interviewees throughout their careers:

…you just need to put your hand up and say, “I fucked up.” This is why, I acted with good intentions, but I did the wrong thing. And if you do that, then nobody can ever accuse you of being anything less than upright and honest. So they were such powerful lessons to learn. (A7)

Participants also emphasized the need to provide junior officers with ‘an environment where they can be honest enough to learn for the benefit of the organisation’ (A9) and mentors with whom ‘…you always felt secure to make mistakes, and to constantly seek advice and affirmation’ (A24). This ability to reflect on and learn from mistakes was mentioned repeatedly, and when mistakes are made there is a need to be vulnerable—with oneself, to colleagues, and with citizens directly implicated in a case. As one participant reflected, ‘I had to sort of open myself up and show, unfortunately, I had to show the complainant that there is a human, a fallible human being…’ (A7).

Although none of the interviewees mentioned the word phronesis , they described their decision-making in terms that are more in line with virtue ethical phronetic reasoning ( Darnell et al ., 2019 ), rather than a deontological or consequentialist one.

Effective and adaptive communication

Police officers must navigate vast and varied roles, meeting the disparate needs of those who call on them for assistance, and balancing the divergent interactions and encounters they have with members of the public. In order to do this, many participants identified the absolute centrality of effective communication. Indeed, for most of the police officers interviewed, the ability to talk to a wide range of people with relative ease was held to be among the most vital capabilities for a police officer to have:

It is the most important skill. Policing is about people isn’t it? … At the end of the day, you can have the best evidence in the world but unless you can sit across the table from someone and talk to them, you’ve got no place in the police service. (S6) …probably the biggest thing for an officer to have is the ability to communicate. I have talked my way out of more violent situations than I can count and purely by the way that I’ve interacted with people. That’s probably…the biggest and best tool that we have… (A17)

It is through these conversations that police officers are able to understand the public, the particular needs and concerns of the victim or (alleged) perpetrator, and adapt their responses as appropriate to achieve effective resolutions. As one interviewee reflected, ‘I had to rein in, I suppose, a little bit of natural exuberance… I had to think quite carefully about what I said to people and how it came across’ (S9). In this regard, police officers indicated their awareness of adapting their communication styles—both elevating and deescalating—in order to achieve their desired outcomes:

It’s about being able to talk to people, in a way that they understand. Be respectful and courteous when someone’s in your face trying to spit at you … (S9) There may be times you need to be very compassionate and caring, then other times you need to be very forthright. It’s that ability to be fluid in that approach and to be able to read people. (S19)

Moreover, interviewees also spoke of the importance of communicating well with colleagues and superiors in order to achieve good outcomes in their work. Where communication is poor, the job becomes more difficult:

So, being able to communicate with people… we police by consent, we don’t enforce the law… So, it is a lot about communication, calming people down by speaking to them and I think that’s probably the most important. (A13) So you might have someone that comes in and delivers a fabulous detection rate, but breaks every other copper, or every investigator that they’ve got. And people don’t want to work with them generally… (A8)

The weight given to the ability to communicate also brought out an interesting trend in the kind of virtues prioritized by police officers. Empathy and compassion (and other similarly emotion-related virtues) were often listed, and were regarded as important components of effective listening. As two interviewees remarked:

… if you can empathise with someone, you understand a bit of where they’re coming from; every officer has had their own struggles at some point. You have that little bit of empathy with people, you’re able to put yourself in their shoes and communicate with them well. (A17) There’s a time and a place for frivolity, there’s a time and a place for humour… there’s a time and a place for no reaction whatsoever, sympathy, empathy, it’s knowing when to give the right feeling at the right time. (A1)

Several interviewees felt that focusing on compassion and empathy is a new concept in policing, which may well have been born of necessity due to the changing role that the police hold in society. Rather than ‘locking up the bad guys, now it’s much more around identifying vulnerability, protecting the innocent, protecting people from harm’ (A2).

This theme of compassion and empathy (which speaks directly to virtue ethical insights) was reflected also in the relationship between leaders and staff within the ranks of the police, with many in positions of authority mentioning being aware of the struggles and vulnerabilities of staff members as well members of the public:

I want my team just to get on and do a really, really good job around the objectives that we’re given to do. But actually, to enable them to do that, for them to achieve that, I need to be more compassionate towards them. More understanding of them and their needs, in order to unlock them to achieve their potential. (A9)

Emotional regulation

In line with adapting modes of communication effectively, there was a clear indication of the need for participants to manage or regulate their emotions, so as to better anticipate how to ensure their actions, behaviours, and communication could aid in better decision-making. ‘You need be assertive, come across that you are in control, and be able to calm situations down’ (A13). Importantly, this illustrative quote may also reveal a slight nuance in how one’s ability to ‘control’ one’s emotions may also be used to dominate interactions.

One participant expressed the importance of ‘Strength of character, just to keep it together at the time, fall [apart] later, but don’t let the public know that you’re panicking or you’re sick by what’s going on, if anything, seem [as] laid back as possible…’ (A1). This ability to self-regulate one’s actions and underpinning emotions was seen as fundamental to modern policing, especially in contexts where ‘…that emotion, that adrenaline that’s, you know, pumping through you, and you’ve got your heart in your mouth…’ (S15).

Much of this ability to adapt and regulate one’s emotions was gained through direct observation of more skilled and senior colleagues. When reflecting on the role of the tutor constable and other role models, one participant expressed:

…you would see with some people, if they went into a situation very calmly, then someone’s always going to come down to your level. Whereas if you start off quite high, then they can’t go down further. So, I think that’s something that stuck with me, with colleagues that were really calm. I’m like, this seems to be a better way, explaining things to people before you even get hands on with them…it’s about your behaviour leads to your reaction, then that leads to their behaviour, and their reaction leads to your behaviour . (A13)

In this way, the ability to know how to manage one’s emotions and reactions can influence the nature of the interaction as well as the outcome and consequences of the experience. The importance of emotional regulation is further illustrated:

I really think maybe the emphasis should be on being really calm, keeping a level-headed approach to not letting whatever that person is saying affect you. (A13) … there’s even times that you can end up having a laugh and suspects you’ve arrested that you’ve maybe previously even been fighting with, then shaking your hand and saying, “You’re an alright copper.” It’s the way that you deal with them … (A17)

Indeed, there was also an indication of awareness by the participants that use of force could at times be used ethically, if the situation called for it, and that in return, police officers’ conduct—with the requisite ability to regulate their emotions—could influence the outcome of a situation and the extent to which they too would be recipients of force. This reveals the extremity of emotional regulation required by police officers, recognizing that at times it may be necessary to use force (deployed in an ethical way), as illustrated by the following quote:

I would always say honey always works better than vinegar and trying to get somebody on side, you’ve got to try that first, because if push comes to shove and you’re in a fight with them on a Friday night, if they knew that you’ve treated them with respect previously, they might not kick you in your head. They might kick you in your stomach, but they might avoid kicking you in your head . (A7)

Notably, emotional regulation is one of the components of the virtue ethical intellectual meta-virtue of phronesis ( Darnell et al. , 2019 ).

One of the defining criteria of a profession is professional autonomy, where its professionals are looked upon and expected to exercise professional judgement ( Anteby et al. , 2016 ). In highly bureaucratic institutions, expectations to conform and comply can prevent professionals from acting with judgement based on professional expertise and knowledge, and thus serve to threaten the professionalism of the role ( Walker, 2018 ). As the findings above illustrate, in policing, the fundamental necessity to be able to exercise autonomy when it comes to handling cases, adapting one’s style to ensure effective communication and outcomes, and using one’s personal judgement to ensure that the right decision is made, is of paramount importance to overall professional competence.

The four themes explicated above indicate that, although police officers may regularly be seen to fall short of the ideal moral character of a virtuous professional and expectations of the public (emphasized through often subjective media depictions focusing on their mistakes rather than their merits; MacVean and Spindler, 2015 : 110; Wood, 2020 : 1), autonomous decision-making is quintessential to the role. In the exercise of this autonomy, structures and frameworks provided to them—such as by the Code and the NDM—provide an important foundation and springboard from which ethical decisions regarding behaviour, communications, and differing levels of legal action can be discerned and dispensed, as deemed appropriate by the individual. However, these frameworks alone do not suffice for the exercise of autonomous decision-making. Rather, they are a springboard, and require officers to employ their own personal judgement—guided by their ability, foresight and expertise—to regulate their emotions, so as to ensure effective and adaptive communication. These are undertaken with the aim of making the ‘right judgements’ and in order to ‘do the right things, for the right reasons’ ( Neyroud and Beckley, 2001 : 27; cf. Kleinig, 1999 ) and ‘in the right way’ ( College of Policing, 2014 : 3).

With an explicit focus on displaying and enacting a wide array of virtues (which extend beyond the remit of the Code’s ‘principles’) throughout their work, participants in the present study clearly spoke of engaging with the public in a manner conducive to the context and associated professional requirements. Where possible, participants referred to engaging in professional conduct in ways that were representative of moral alignment with the public, and therefore, the corresponding justification of police legitimacy. While this theme of ‘moral alignment’ was not a prominent finding in Manning’s interview study ( 2019 ), the present research findings tally with the conceptualization proposed by Jackson and colleagues ( 2012 : 1051)—namely that the police act ‘according to a shared moral purpose with citizens’. While this may not seem explicit in the quotes included above, the emphasis on ‘knowing when to give the right feeling at the right time’ (A1) and ‘identifying vulnerability, protecting the innocent, protecting people from harm’ (A2), is indicative of a desire and appreciation of moral alignment. These findings partially reflect the considerations explicated in ‘self-legitimacy’, where links between normative orientations toward the public and ‘self-legitimacy’ are highlighted ( Tankebe, 2019 ). In this respect, police officers carefully reflect on how best to enforce the law while ensuring that their actions do not alienate the public, seeking to engage in ways that ‘establish attributions of legitimacy’, and which is linked to a range of positive attitudes and behavioural outcomes ( Bottoms and Tankebe, 2013 : 75; Tankebe, 2019 ). In this way, while police officers seek to legitimate policing by reflecting moral alignment with the public, it remains for the public (the audience) to accept this claim and confer the legitimacy to the police (the power holder, Bottoms and Tankebe, 2013 : 64).

Nevertheless, the challenges and obstacles identified by several participants highlighted the risks that the professionalization agenda of policing, insofar as it is deontologically motivated, poses to inhibit the exercise of autonomous decision-making. The perceived increase of ‘red tape’ serves (for some) as a barrier to exercising personal judgement, and ultimately to make the right decision. This implies a lack of trust in the police as professionals, despite a strong recognition of the need to be accountable and able to justify why certain decisions and judgements were made. Throughout the research, while some participants spoke of policing as ‘enforcing the rule of law’, all of them recognized that policing was neither instrumental nor narrowly defined in this way. Rather, it was grounded in a morally driven service aimed at helping and supporting the community by fighting crime, as well as fulfilling the role of a social service institution. Police officers’ reflections about ethical considerations and moral motivations in their work seem in this respect to contribute to the legitimation process, whereby the police officers self-proclaimed and authentically expressed ethicality in how they undertake their work are claims to self-legitimacy ( Tankebe, 2019 ). This aligns with findings suggesting that ethically motivated police officers, whose conduct is underpinned by impartial decision-making, treating the public with dignity, and boundary-awareness (bounded authority)—all practices which support the legitimation process—are essential to a healthy and well-functioning police service ( Trinkner et al. , 2018 ).

The four themes identified here underline the value of the Code , and that officers ensure their actions account for ‘doing the right thing in the right way’ (2014:3), which corresponds to a standard virtue ethical formulation for practical wisdom ( phronesis ), as elaborated upon, for instance, by Darnell et al. (2019) . Given that the police are often considered the most deontologically grounded of all professions (apart from the Army), the extent to which the interviews were steeped in a virtue of ethical language was surprising. Thus, as guided by the Code and in their application of its principles, police officers clearly saw the need to use autonomous phronetic discretion in interpreting and responding to situations, instead of the blind application of rules or maxims. From the interview data, there was a clear indication that the Code positively aligned with the interviewees’ personal moral codes, indicating that the Code had been internalized, or that individual moral codes were already reflected in the Code . Nevertheless, the findings from the survey, listed elsewhere ( Maile et al ., 2022 ), indicated that deontology is still the reasoning strategy that police officers and aspiring officers most often avail themselves of in solving police dilemmas—closely followed, however, by virtue of ethical reasoning.

Throughout the Code, the necessity for officers to demonstrate behavioural and emotion-related virtues, such as being able to show ‘compassion and empathy, as appropriate, to people you come into contact with’ (2014: 7), is foregrounded. This corresponds well with participants’ prioritization of effective and contextualized communication and emotional regulation. Alongside the Code sits the NDM, which offers five steps in guiding police officers in their ethical decision-making, emphasizing that police officers should use it ‘wisely’ (2014: 9) and with discretion. As such, the decision-making procedure is ‘inherently flexible’ (2014: 7).

Historically, police ethical codes have typically been offered as an antidote or replacement to covert yet prevalent cultural norms within the police known as the ‘blue code’ ( Westmarland and Rowe, 2018 ). Instead of becoming embroiled in a prevailing culture within the police, a written code thus enables officers to question and rethink the normative ‘police culture’ into which they are socialized. In addition, despite the Code not being explicitly grounded in a discrete ethical theory—or seeing the use of ‘principles’ as indicating a deontological heritage—its use of virtue terms and concepts, as opposed to a language of compliance, sets it apart from similar professional codes ( Earl and Moulin- Stožek, 2019 ). There is also no indication in the Code or in the NDM that either should replace the use of individual and shared discretion or reflective thinking by police officers. Rather, the NDM is offered as a guide for police officers through some of the necessary steps in order to reach a sound decision within the circumstances. When viewed through a virtue ethics lens, the NDM is similar to a neo-Aristotelian model and the components of phronesis , which encourages individuals to make well-grounded autonomous moral decisions ( Darnell et al. , 2019 ).

Importantly, while the overarching ‘duty’ of a police officer may be to uphold the public order, by applying the sections detailed in the Public Order Act (1986) in the same way as one would deductively apply formal rules, the concept of ‘public order’ remains a nebulous concept. While making an attempt to codify the concept of the ‘Queen’s Peace’ (referring to ‘public order’ in the UK), the formalistic royal road to determining what counts as a breach of public order is rocky, requiring professional knowledge and expertise. 4 It is a matter for the police to use their personal judgement and discernment, and the themes that have emerged from this research indicate that this is already occurring. This finding tallies with the emphasis in Wood’s book ( 2020 : 5) on ‘the ability of an officer to make sense of competing demands within different contexts, as opposed to following ethical rules of one kind or another’. This ‘ability’ is typically referred to in the virtue ethical literature as (professional) phronesis .

This understanding of phronesis , and the emergent themes from our qualitative analysis, highlight the necessity for police officers to cultivate and display the necessary personal and professional virtues required of them to do the right thing, at the right time, for the right reasons. Such phronetic decisions are critical, morally tethered, and properly deliberated and adjudicated upon.

Contributing to the professionalization of policing, the watershed review by Neyroud (2011) led to the establishment of the professional qualification framework for policing in England and Wales (PEQF; Brown et al. , 2018 ; Hunter and May, 2019 ). This educational focus of police training aligns with the need for policing to provide a more effective and knowledge-based response to contemporary issues ( Hallenberg and Cockcroft, 2017 ). However, as this paper reveals, more could be done to incorporate a stronger and more explicit ethical focus in police education and training, mirroring findings by Williams (2021) . Indeed, the rhetoric of professionalization and the need for autonomy and discretion in policing, highlighted in this paper, emphasize the requirement for policing environments and educational provision to support the development and exercise of professional phronesis . Such a shift is necessary in a profession that exacts significant levels of responsibility on its professionals, faces situations of high-level moral complexity, and demands professional competence at all times.

The themes identified in this study emphasize the role that phronesis (or phronesis -like discernment) plays in the moral practice of policing. As such, the authors of this paper and the associated project ( Kristjánsson et al ., 2021 ) recommend that any revisions to the Code ought to make explicit its ethical theoretical grounding, and recognize (and celebrate) the virtue ethical considerations it reflects. Such a move would reflect well the sentiments expressed by the 30 police officers interviewed, as well as developments in general professional ethics where virtue ethics is gradually replacing deontology as the theory of choice.

This work was supported by funding from the John Templeton Foundation, grant number 962685.

Note that while the Code of Ethics applies only to England and Wales, the participant sample includes police officers from Scotland.

The Code of Ethics is currently undergoing a review process by the College of Policing, which is likely to result in it being re-written. This paper focusses on the first iteration of the Code.

To clarify, this paper is specifically interested in the principles of the first iteration of the Code and the associated narrative, not the code of conduct utilized in misconduct proceedings.

While ss. 1-3 of the Public Order Act (1986) are tightly defined, there remains a role for professional judgement and discretion in interpreting whether s4, 4A, or 5, has been breached in any given circumstance.

Anteby M. , Chan C. K. , and Dibengigno J. ( 2016 ). ‘Three Lenses on Occupations and Professions in Organizations: Becoming, Doing, and Relating.’ The Academy of Management Annals 10 : 183 – 244 .

Google Scholar

Aristotle. ( 1985 ). Nicomachean Ethics , trans. T. Irwin . Indianapolis : Hackett Publishing .

Google Preview

Arthur , J. and Earl , S. R. ( 2020 ). ‘Character in the Professions: How Virtue Informs Practice. Research Report’. Jubilee Centre for Character and Virtues . Retrieved September 12, 2022, from https://www.jubileecentre.ac.uk/userfiles/jubileecentre/pdf/projects/TransformativeBritain/Professions_Report.pdf .

Bottoms , A. and Tankebe , J. ( 2012 ). ‘Beyond Procedural Justice: A Dialogic Approach to Legitimacy in Criminal Justice.’ Journal of Law and Criminology 102 : 119 – 170 .

Bottoms , A. and Tankebe , J. ( 2013 ). ‘“Voice Within”: Powerholders’ Perspectives on Authority and Legitimacy.’ In Tankebe , J. and Liebling , A. (eds), Legitimacy and Criminal Justice: An International Exploration . Oxford : Oxford University Press , pp. 60 – 82 .

Braun , V. and Clarke , V. ( 2006 ). ‘Using Thematic Analysis in Psychology.’ Qualitative Research in Psychology 3 ( 2 ): 77 – 101 .

Brown , J. , McDowall , A. , Gamblin , D. , and Fenn , L. ( 2018 ). ‘Assessing Transmission and Translation of Learning about Evidence-Based Policing by Graduate Trainee Police Officers.’ Policing 14 ( 1 ): 119 – 134 .

Carr , D. ( 1999 ). ‘Professional Education and Professional Ethics.’ Journal of Applied Philosophy 16 ( 1 ): 33 – 46 .

Carr , D. ( 2000 ). Professionalism and Ethics in Teaching . London : Routledge .

College of Policing. ( 2014 ). ‘Code of Ethics.’ https://www.college.police.uk/ethics/code-of-ethics .

Critchley , T. ( 1967 ). A History of the Police in England and Wales . London : Constable .

Darnell , C. , Gulliford , L. , Kristjánsson , K. , and Paris , P. ( 2019 ). ‘ Phronesis and the Knowledge-Action Gap in Moral Psychology and Moral Education: A New Synthesis?’ Human Development 62 : 101 – 129 .

Earl , S. R. and Moulin-Stožek , M. ( 2019 ). ‘Moving Towards Virtuous Professional Practice: A Summary of a Content Analysis of Professional Regulatory Documents and Codes of Conduct . ’ Insight Series Paper. https://www.jubileecentre.ac.uk/userfiles/jubileecentre/pdf/insight-series/MovingTowardsVirtuousProfessionalPractice.pdf .

Hallenberg , K. M. and Cockcroft , T. ( 2017 ). ‘From Indifference to Hostility: Police Officers, Organizational Responses and the Symbolic Value of ‘In-Service’ Higher Education in Policing.’ Policing 11 ( 3 ): 273 – 288 .

Her Majesty’s Inspectorate of Constabulary. ( 2022 ). ‘Inspection of the Metropolitan Police Service.’ https://www.justiceinspectorates.gov.uk/hmicfrs/publications/an-inspection-of-the-metropolitan-police-services-counter-corruption-arrangements-and-other-matters-related-to-the-daniel-morgan-independent-panel/ .

Holdaway , S. ( 2017 ). ‘The Re-professionalisation of the Police in England and Wales.’ Criminology and Criminal Justice 10 ( 1 ): 11 – 22 .

Home Office. ( 2014 ). ‘Home Office Guidance: Police Officer Misconduct, Unsatisfactory Performance and Attendance Management Procedures.’ https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/330235/MisconductPerformAttendanceJuly14.pdf .

Hough , M. ( 2021 ). Good Policing: Trust, Legitimacy and Authority . Bristol : Policy Press .

Hunter , G. and May , T. ( 2019 ). ‘Views from the Frontline: Graduate Police Recruits on the Status of Evidence-based Practice.’ Policing 14 ( 1 ): 91 – 103 .

Independent Office for Police Conduct. ( 2022 ). ‘Operation Hotton.’ https://www.policeconduct.gov.uk/sites/default/files/Operation%20Hotton%20Learning%20report%20-%20January%202022.pdf .

Jackson , J. , Bradford , B. , Hough , M. et al.  . ( 2012 ). ‘Why Do People Comply with the Law? Legitimacy and the Influence of Legal Institutions.’ The British Journal of Criminology 52 ( 6 ): 1051 – 1071 .

Jubilee Centre for Character and Virtues. ( 2016 ). ‘Statement on Character, Virtue and Practical Wisdom in Professional Practice’. Retrieved September 12, 2022, from https://www.jubileecentre.ac.uk/userfiles/jubileecentre/pdf/Statement_Character_Virtue_Practical_Wisdom_Professional_Practice.pdf .

Kant , I. ( 1964 ). Metaphysics of Morals , trans. M. J. Gregor . New York : Harper & Row .

Kleinig , J. ( 1999 ). The Ethics of Policing , Cambridge : Cambridge University Press .

Kristjánsson , K. ( 2015 ). Aristotelian Character Education. London : Routledge .

Kristjánsson , K. , Thompson , A. , and Maile , A. ( 2021 ). ‘Character Virtues in Policing. Research Report.’ Jubilee Centre for Character and Virtues. Retrieved September 12, 2022, from https://www.jubileecentre.ac.uk/userfiles/jubileecentre/pdf/Research%20Reports/CharacterVirtuesinPolicing_ResearchReport.pdf .

Legislation.gov.uk. ( 1986 ). Public Order Act 1986. [online] Available at: http://www.legislation.gov.uk/ukpga/1986/64 (accessed 10 October 2022 ).

MacVean , A and Spindler , P. ( 2015 ). ‘Principled and Ethical Policing: Some Considerations for Police Leaders.’ In Fleming , J. (ed.), Police Leadership: Rising to the Top . Oxford : Oxford University Press , pp. 110 – 128 .

Maile , A. , Thompson , A. , McLoughlin , S. , and Kristjánsson , K. ( 2022 ). ‘The choice of Deontological, Virtue Ethical, and Consequentialist Moral Reasoning Strategies by Pre- and In-Service Police Officers in the U.K.: An Empirical Study’. Ethics and Behaviour.

Manning , M.L. ( 2019 ). Policing as a Virtue: Moral Alignment and Legitimacy . Unpublished PhD Thesis , University of Essex .

Mark , R. ( 1977 ). Policing a Perplexed Society . London : George, Allen and Unwin .

Mill , J.S. ( 1863 ). Utilitarianism . London : Parker, Son & Bourn .

Morrell , K. and Bradford , B. ( 2019 ). Policing and Public Management: Governance, Vices and Virtues . London : Routledge .

Neyroud , P. ( 2006 ). ‘Ethics in Policing: Performance and the Personalisation of Accountability in British Policing and Criminal Justice.’ Legal Ethics 9 ( 1 ): 16 – 34 .

Neyroud , P. ( 2011 ). Review of Police Leadership and Training . London : Home Office .

Neyroud , P. and Beckley , A. ( 2001 ). Policing, Ethics and Human Rights . Cullompton : Willan .

Rowe , M. , Turner , E. , and Pearson , G. ( 2016 ). ‘Learning and Practicing Police Craft’. Journal of Organizational Ethnography 5 ( 3 ): 276 – 286 .

Schwartz , B. and Sharpe , K. E. ( 2010 ). Practical Wisdom: The Right Way to do the Right Thing. Riverhead Books.

Tankebe , J. ( 2019 ). ‘In Their Own Eyes: An Empirical Examination of Police Self-legitimacy.’ International Journal of Comparative and Applied Criminal Justice 43 ( 2 ): 99 – 116 .

Trinkner , R. , Jackson , J. , and Tyler , T. ( 2018 ). ‘Bounded Authority: Expanding “Appropriate” Police Behaviour Beyond Procedural Justice.’ Law and Human Behaviour 42 ( 3 ): 280 – 293 .

Tyler , T.R. ( 2006 ). Why People Obey the Law . Princeton : Princeton University Press .

Walker , D.I. ( 2018 ). ‘Character in the British Army: A Precarious Professional Practice.’ In Carr , D. (ed.), Cultivating Moral Character and Virtue in Professional Practice . London : Routledge , pp. 135 – 148 .

Westmarland , L. and Rowe , M. ( 2018 ). ‘Police Ethics and Integrity: Can a New Code Overturn the Blue Code?’ Policing and Society 28 : 854 – 870 .

Williams , C.A. ( 2021 ). How Does Internal, Initial Police Officer Training Affect Police Officer Ethical Decision-making? Unpublished PhD Thesis , University of Central Lancashire .

Williams , E. , Norman , J. , and Rowe , M. ( 2019 ). ‘The Police Education Qualification Framework: A Professional Agenda or Building Professionals?’ Police Practice and Research 20 ( 3 ): 259 – 272 .

Wood , D.A. ( 2020 ). Towards Ethical Policing . Bristol : Policy Press .

Email alerts

Citing articles via.

• Recommend to your Library

Affiliations

• Online ISSN 1752-4520
• Copyright © 2024 Oxford University Press
• About Oxford Academic
• Publish journals with us
• University press partners
• What we publish
• New features  
• Open access
• Institutional account management
• Rights and permissions
• Get help with access
• Accessibility
• Advertising
• Media enquiries
• Oxford University Press
• Oxford Languages
• University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

• Copyright © 2024 Oxford University Press
• Cookie settings
• Cookie policy
• Privacy policy
• Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

Decision-Making for UK Police on the Transportation of Casualties with Life-Threatening Injuries

• December 2020
• Journal of High Threat & Austere Medicine 2(1)
• CC BY-NC 4.0
• This person is not on ResearchGate, or hasn't claimed this research yet.

Abstract and Figures

Discover the world's research

• 25+ million members
• 160+ million publication pages
• 2.3+ billion citations
• Recruit researchers
• Join for free
• Login Email Tip: Most researchers use their institutional email address as their ResearchGate login Password Forgot password? Keep me logged in Log in or Continue with Google Welcome back! Please log in. Email · Hint Tip: Most researchers use their institutional email address as their ResearchGate login Password Forgot password? Keep me logged in Log in or Continue with Google No account? Sign up

Make it happen

Data Breaches & the National Decision Model (NDM)

Data breaches and losses should be a thing of the past, although recent high profile incidents highlight that this is still not the case. The mismanagement of data and information comes with potentially serious and very costly implications, particularly where personal information about vulnerable members of the public is concerned.

I recently read an article about a local authority who lost a memory stick holding sensitive information. Whilst I gather the memory stick was encrypted, it does raise questions around data sharing and security particularly when secure and trusted collaboration platforms pose such a compelling alternative to old ways of working.

Typically such incidents occur as a result of procedure not being followed and / or poor decision making.  Increasingly public services are looking to bring their processes in line with recognised International Standards (i.e ISO 27001 as Alliantist have) and decision models. One such model that is being implemented on a national basis is the Association of Chief Police Officers (ACPO) National Decision Model (NDM).

The NDM was designed by ACPO to help police officers and staff develop the professional judgement necessary to make effective policing decisions and to learn from the outcomes of those decisions, be they successful or unsuccessful. The NDM provides a “simple, logical and evidence-based approach to making policing decisions.”

The implementation of the NDM highlights a focus from the police on supporting the increasing demands on limited policing resources & mitigating risks through evidence-based decision making. “Decision makers will receive the support of their organisation in all instances where they can demonstrate that their decisions were assessed and managed reasonably in the circumstances existing at the time. This applies even where harm results from their decisions and actions.”

The inherent flexibility of the model makes it suitable for all types of police decisions, from spontaneous incidents to planned operations, and can be utilised by individuals or teams. “In every case, the model stays the same, but users decide for themselves what questions and considerations they apply at each stage.”

After looking at the model, we quickly recognised an opportunity to digitise the process. Through an easily repeatable framework for the NDM, users can quickly integrate this national model into the work they are currently doing on pam, either alone or with their partners, to start making better decisions, together.

To find out how pam could benefit your organisation simply get in touch .

If you liked this, stay informed via email...

We won't share your details with anyone else.

Recent blog posts

• Make Time to Listen by playing as a multi-agency safeguarding team
• Multi Agency Safeguarding in Warwickshire – A case study in successful collaboration
• It’s not just safeguarding of people required, it’s their data too
• Alliantist earns Cyber Essentials certification
• Multi Agency Safeguarding Hub launches in Warwickshire

Blog Archives

Warning: The NCBI web site requires JavaScript to function. more...

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.

Kuntz K, Sainfort F, Butler M, et al. Decision and Simulation Modeling in Systematic Reviews [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2013 Feb.

Decision and Simulation Modeling in Systematic Reviews [Internet].

Overview of decision models used in research.

• Introduction

Decision analysis is a systematic, quantitative, and transparent approach to making decisions under uncertainty. The fundamental tool of decision analysis is a decision-analytic model, most often a decision tree or a Markov model. A decision model provides a way to visualize the sequences of events that can occur following alternative decisions (or actions) in a logical framework, as well as the health outcomes associated with each possible pathway. Decision models can incorporate the probabilities of the underlying (true) states of nature in determining the distribution of possible outcomes associated with a particular decision. These probabilities are not known to the decisionmaker but are critically important. For example, the value associated with the decision to do a diagnostic test depends on the probability that the patient has the disease of interest (i.e., the prior probability of disease), which in turn determines the probabilities of true and false positive findings, as well as true and false negative findings. Ultimately, the relative values assigned to these different outcomes determine the clinical value of a test, for example, how bad is it to fail to treat a person with disease relative to treating a person without disease? While the evidence requirement for a diagnostic test focuses on its diagnostic performance (e.g., sensitivity, specificity), the clinical value of the test, which should underlie the decision to use or recommend the test, depends on clinical evidence well beyond measurements of test performance. For example, does the information obtained from the diagnostic test change the decision that would be made, and does this change in decision lead to improved health outcomes? The methods of decision analysis have been found to be particularly useful in settings where multiple data inputs from a variety of studies are relevant in a particular decisionmaking context.

Since the first application in 1967, decision-analytic models have been increasingly used to evaluate and compare competing public health and medical interventions. Decision models can vary from a very simple “back of the envelope” type of calculation to extremely complex computer-based microsimulation models. While the term “model” has different meanings in different clinical settings, 27 such as a statistical model, the fundamental feature of a decision model is that the goal of a decision model is to assist with decisionmaking and not to make statements about truth. Statistical study designs such as randomized clinical trials and case control studies are focused on gathering evidence; however, decision analysis studies are aimed at processing evidence. In the absence of a systematic and formal approach that assists the decisionmaker in the processing of the (often disparate and complex) evidence, the processing occurs in a more informal way.

There has been little documentation of the use of decision models in the literature. Hence, the goal of our analysis was to profile decision-analytic models published in the literature over the past 5 years, and to provide a summary of who is publishing models, in what disease areas, in what journals, and for what types of interventions.

We sought to obtain a “bird's eye” view of the number of articles in the medical literature that used a model for medical decisionmaking purposes. Our goal was to capture studies that used the methods of decision analysis to project health outcomes associated with two or more competing strategies for purposes of decisionmaking. We were not interested in studies with a goal of estimating efficacy or effectiveness parameters, nor were we interested in studies that evaluation environmental interventions or regulatory policies. Our goal was to document the current use of decision models in the medical literature (within the past 5 years) and document basic information about the authors and analyses.

Literature Search

We conducted a systematic review of the medical literature and the grey literature to document and synthesize all analyses that used a decision-analytic model, including relevant cost-effectiveness analyses, conducted within the past 5 years.

We searched Medline for articles that used some form of decision model. We relied on key word searches to locate decision models since medical subject headings (MeSH) terms are not well indexed for decision modeling topics. Searches employing a model term other than “decision analysis” or “decision analytic model” required filters to eliminate nonrelated articles. For these filters, we used some form of quality adjusted life OR some form of incremental cost effectiveness ratio OR some form of disability adjusted life OR some form of modeling technique to further specify related articles. In addition, since cost-effectiveness models are the most commonly noted form of economic modeling references in the literature, we created a filter to further limit cost-utility and cost-effectiveness models to those that were not reported as simple extensions of a single clinical trial, as these were viewed as not incorporating multiple sources of data. Articles were limited to English language and human subjects. The full search string is provided in Appendix A . The search was performed on October 15, 2009, with an update search performed on February 18, 2010, for articles published from 2005 through 2009.

Inclusion/Exclusion Criteria

Articles were not excluded by country of origin, type of disease condition, or treatment provided. We also did not include decision aids or statistical analyses for estimating effect size, inference, or prediction. We included cost-effectiveness analysis as long as the model was used to project the costs and the effectiveness of two or more strategies, but excluded papers that evaluated costs only. Since our focus was on the effectiveness/comparative effectiveness of health care interventions at the individual level, we did not include dynamic models, or infectious disease models, which focus on describing epidemics rather than on evaluating strategies for decisionmaking purposes. We also excluded papers that reviewed current evidence around a topic, which may include pertinent decision analysis models, but did not use a model explicitly.

Screening and Data Abstraction

Model definition.

To ensure a broad look at the use of models in the literature, we defined a decision model as a mathematical structure developed to synthesize two or more sources of evidence, used to project out the health outcomes associated with alternative policies. This model definition is broader than what was applied to determine eligibility for the database discussed in the Potential Modeling Resources section of this report. Again, the purpose here was to generate a broad sense of the state of the literature.

Abstract Review

References generated from the search string were imported to Refworks for screening. Two reviewers screened all reference titles and abstracts for inclusion or exclusion. Conflicts were reconciled by a third reviewer through consensus. Included references were retained for abstracting. For excluded abstracts, the reason for exclusion was noted. Each abstract retained was reviewed independently by two readers. The reviewers abstracted the following information: title, journal, country of primary author, type of intervention, class of intervention, target disease, country of the target population, age group of the target population (children, adult, elderly), and the type of health outcome reported. The reviewers met to compare abstraction results; conflicts were reconciled through consensus, using a third reviewer if necessary.

Intervention Classification

The modeling analyses were classified according to the type of interventions that were evaluated: prevention, screening, diagnosis, and treatment.

Models were classified as “prevention” if the model studied an intervention aimed at a population with no symptoms with the express intent to prevent a certain disease. This most often includes models to evaluate the effectiveness of vaccinations, but also included pharmacological and other types of interventions, e.g., bed nets for malaria prevention.

Models were classified as “screening” if the intervention pertained to patients with no specific symptoms or diseases but was a method to determine a certain affliction from a general population. There were some caveats allowed in this definition in cases where a higher-risk population could be identified by factors other than signs and symptoms of a disease or condition. For example, we would categorize colonoscopy screening every 5 years among individuals who have had a polypectomy in the past as screening. Although this group is at higher risk of having colorectal cancer or an adenomatous polyp than the general population, members do not have signs and symptoms of disease at the time of the colonoscopy. (Note that in practice, this is referred to as surveillance and not screening.) Screening also includes those models that examine strategies of screening versus treatment of certain disease.

The “diagnosis” categorization includes models that target a population with a set of signs and symptoms related to a disease. Diagnostic interventions include the use of a diagnostic test to gather more information about the disease status of the patient and to target subsequent treatment or further testing based on the test results. This category includes lab testing, radiological test, and genetic tests.

Models were given the categorization of “treatment” if the model examined a certain intervention to be used on patients who had a specific disease. This can include pharmacological interventions, procedural interventions (surgery, noninvasive device), or organizational interventions, such as stroke unit care. Treatment can also include strategies aimed at preventing complications in the normal course of treatment for disease, such as using a certain kind of surgical instrument to avoid infections when treating a patient.

Disease Classification

The articles were also classified according to the diseases targeted with the interventions. These diseases represent either the disease that the target population already had (in the case of treatment), the disease that the target population might have (in the case of screening or diagnostic interventions), or the disease that the intervention is aimed at preventing, (in the case of prevention.) Table 4 shows the disease categories that we used.

Disease categories.

Summary Statistics

For the 5-year time period, we calculated the distributions of the home country of primary author, the journal, type of intervention, disease categories, the target population, and outcomes measures.

We present an overview of the use of decision-analytic models in the medical literature over the past 5 years. We identified 1,773 articles published in 2005–2009 that included the use of a decision model to compare the clinical outcomes associated with two or more strategies and excluded an additional 1,075 articles. Of the 1,075 articles excluded, the majority (56 percent) were deemed to be outside the scope of the review because they did not involve the comparison of two or more health care interventions. Other reasons for not including articles are shown in Table 5 .

Reason for rejection (2005–2009).

The home country of the primary author was the United States for 42 percent articles, and this was fairly consistent across years. The second most common home country of the primary author was the United Kingdom (16 percent), which was also consistent across this timeframe. Table 6 shows the distribution of the home country of the primary author of the 2005–2009 modeling articles.

Home country of primary author (2005–2009).

Health Technology Assessment published the most articles (5 percent) in 2005–2009, followed by Pharmacoeconomics (4 percent). Table 7 shows the distribution of articles by journal name. Table 8 shows the type of interventions modeled in the 2005–2009 modeling papers. The majority of the intervention types were treatment, representing 70 percent of the total articles. Among the articles that targeted a diseased population, 46 percent of the interventions were pharmaceutical interventions, and 14 percent of them evaluated procedural interventions. The second most common intervention type (12 percent) was prevention, most of which pertained to the evaluation of vaccinations. Screening strategies were evaluated in 12 percent of the articles and diagnostic interventions were evaluated in 6 percent of the 2005–2009 articles. These findings were consistent across years. (Data not shown.)

Journals represented with more than seven articles (2005–2009).

Type of intervention modeled (2005–2009).

Table 9 shows the classes of disease types that were the focus of the 2005–2009 articles. The most common disease type evaluated was cancer (20 percent of the articles), followed by cardiovascular diseases (14 percent of the articles), and this was consistent across the 5 years (data not shown).

Diseases addressed by models (2005–2009).

Table 10 shows the age (in broad categories) and sex of the target populations for those modeling articles that specified age and sex in the abstracts for 2005–2009. Approximately 10 percent of the 1,773 articles specifically stated that the focus of the analysis was on a pediatric population, while 6 percent of the stated a focus on an elderly only population and 15 percent with a stated focus on women only.

Target populations (2005–2009).

Table 11 shows the primary outcome measures reported in the 2005–2009 modeling articles. The majority of the articles reported adjusted life years as an outcome (62 percent). Other outcomes reported included life years, survival, and cases detected.

Primary outcome measures (2005–2009).

Table 12 shows the intervention types by disease category. While 22 percent of the overall articles focus on cancer, 36 percent of the screening articles focus on cancer. Some diseases are associated almost exclusively with treatment-related interventions (>90 percent of disease-specific articles are related to treatment). These include mental disease, kidney diseases, and multiple sclerosis. The diseases that were more associated with prevention interventions (>25 percent of the disease-specific articles) were lung diseases, influenza, and rotavirus. Most of the interventions pertaining to gastrointestinal disease were related to diagnostics.

Intervention types modeled by associated disease (2005–2009).

We provide a bird's eye view of the articles that used a decision model to incorporate data from multiple sources to evaluate two or more strategies for decisionmaking purposes. Over a 5-year period, we identified more than 1,700 articles in several disease areas and intervention types published in journals listed in Medline. It is possible that we missed some studies as we did not search other databases (e.g., EconLit, Heath Economic Evaluations Database). The diseases areas and interventions types are fairly representative of those evaluated by systematic reviews. The majority of the decision analysis papers evaluated an intervention aimed at individuals with existing disease, and surprisingly few were focused on diagnostic-related strategies.

Overall, 42 percent of the articles came from the United States, though over 45 countries were represented (home country of primary author). A wide variety of journals publish decision analysis papers. The journal that published the greatest number of articles, Health Technology Assessment, only represents 5 percent of the total papers. However, it is not surprising that Health Technology Assessment is a leading journal for publishing models because they publish peer-reviewed reports from the Health Technology Assessment program in the United Kingdom.

The most common disease type evaluated was cancer (20 percent of the articles), followed by cardiovascular diseases (14 percent of the articles). The majority of the target populations were adults, with approximately 10 percent of the 1,773 articles with a focus on a pediatric population, and 6 percent with a focus on an elderly population. In addition, 15 percent of the articles were targeted towards women only. The majority of the articles reported adjusted life years as an outcome (62 percent), the majority of which used quality-adjusted life years as the outcome.

The search algorithm we developed for this analysis could be useful in the future to identify studies that use a decision model within a disease area. (See Potential Modeling Resources .) We found that the search algorithm has a reasonable yield (61 percent) in terms of the percent of papers subsequently found to include a decision model. However, we are not certain how many decision models we have missed.

An overview of the decision analysis field, as detailed by this analysis of the decision modeling literature, provides a first step in the development of a framework for using models with systematic reviews. While there are a relatively large number of papers that have used the methods of decision analysis in the past three years, they tend to be published in journals whose audience is modelers, or in somewhat lower-tiered clinical journals. There are exceptions to this, of course, but one goal for the field of decision analysis is to reach a broader audience through publication in high-impact clinical journals. Development of a model alongside a systematic review may help move the field in this direction, as there would be opportunities to publish in the Annals of Internal Medicine. Also, the connection to a well-done systematic review and formal decisionmaking process will likely add to the credibility and understanding of the use of decision models for policymaking.

• Cite this Page Kuntz K, Sainfort F, Butler M, et al. Decision and Simulation Modeling in Systematic Reviews [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2013 Feb. Overview of Decision Models Used in Research.
• PDF version of this title (1.2M)

In this Page

Other titles in these collections.

• AHRQ Methods for Effective Health Care
• Health Services/Technology Assessment Texts (HSTAT)

Recent Activity

• Overview of Decision Models Used in Research - Decision and Simulation Modeling ... Overview of Decision Models Used in Research - Decision and Simulation Modeling in Systematic Reviews

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

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 19 August 2024

Development and validation of a nomogram of all-cause mortality in adult Americans with diabetes

• Xia Shen 1   na1 ,
• Xiao Hua Zhang 2   na1 ,
• Long Yang 3   na1 ,
• Peng Fei Wang 4 ,
• Jian Feng Zhang 5 ,
• Shao Zheng Song 7 &
• Lei Jiang 6  

Scientific Reports volume  14 , Article number:  19148 ( 2024 ) Cite this article

50 Accesses

Metrics details

• Endocrinology
• Risk factors

This study aimed to develop and validate a predictive model of all-cause mortality risk in American adults aged ≥ 18 years with diabetes. 7918 participants with diabetes were enrolled from the National Health and Nutrition Examination Survey (NHANES) 1999–2016 and followed for a median of 96 months. The primary study endpoint was the all-cause mortality. Predictors of all-cause mortality included age, Monocytes, Erythrocyte, creatinine, Nutrition Risk Index (NRI), neutrophils/lymphocytes (NLR), smoking habits, alcohol consumption, cardiovascular disease (CVD), urinary albumin excretion rate (UAE), and insulin use. The c-index was 0.790 (95% CI 0.779–0.801, P  < 0.001) and 0.792 (95% CI: 0.776–0.808, P  < 0.001) for the training and validation sets, respectively. The area under the ROC curve was 0.815, 0.814, 0.827 and 0.812, 0.818 and 0.829 for the training and validation sets at 3, 5, and 10 years of follow-up, respectively. Both calibration plots and DCA curves performed well. The model provides accurate predictions of the risk of death for American persons with diabetes and its scores can effectively determine the risk of death in outpatients, providing guidance for clinical decision-making and predicting prognosis for patients.

Similar content being viewed by others

Prediction model for cardiovascular disease in patients with diabetes using machine learning derived and validated in two independent Korean cohorts

Development and validation of a risk prediction model for chronic kidney disease among individuals with type 2 diabetes

Identifying myoglobin as a mediator of diabetic kidney disease: a machine learning-based cross-sectional study

Introduction.

Diabetes mellitus is a metabolic disease marked by hyperglycemia 1 . In recent years, the prevalence of diabetes has shown an alarming increase with changes in diet and lifestyle and is developing into a major public health problem in the world 2 . Studies have reported that the prevalence of diabetes among adults (20–79 years) worldwide will increase to 7.7% by 2030, affecting 439 million adults 3 .

Diabetes is known to cause long-term and severe damage to multiple organs and body systems, such as the eyes, heart, kidneys, nerves, and blood vessels, and is a leading cause of blindness, heart attack, kidney failure, stroke, and amputation of lower limbs 4 . Studies reported that 2.9 million deaths a year are directly attributable to diabetes, a figure comparable to that of human immunodeficiency virus infection and the overall risk of death in participants with diabetes is almost twice that of those without diabetes 5 , 6 . Moreover, a multicounty data analysis from 19 data sources in 16 countries or jurisdictions demonstrated that the mortality of participants with diabetes decreased more rapidly than those without diabetes 7 . Furthermore, diabetes and its complications can also cause great physical suffering and financial burden to patients 8 . Therefore, it is necessary to predict mortality in diabetes and establish corresponding predictive models to develop aggressive prevention strategies for very high-risk participants that can help reduce the risk of premature death and improve prognosis.

The nomogram is not only a visual statistical prognostic tool, which is widely used in the prognostic assessment of clinical outcomes by calculating the score of potential predictors but also a simple, effective, and reliable prediction model, which can quickly provide clinical risk stratification for clinical care to make prognostic decisions 9 . To date, several studies have reported predictive models for diabetes and its complications 10 , 11 , 12 , but prediction models for all-cause mortality in a nationally representative person with diabetes have rarely been reported 13 . As there is no population-based study to establish a mortality risk prediction model for Americans with diabetes from NHANES, this study aims to construct an effective nomogram prediction model for adult (years ≥ 18) outpatient diabetes in the United States to predict their 3-year, 5-year, and 10-year risk of all-cause mortality, helping them benefit from better clinical interventions.

Baseline characteristics of participants

In this study, 7918 participants were included, as shown in Fig. 1 . We weighed and adjusted all the results when comparing the two datasets. Among all participants in the study, the median age was 59.014 years (IQR, 58.74–59.29 years) and with 3887(49.62%) females and 4031(50.38%) males were recruited. At a median follow-up of 96 months and a total of 258 (28.61%) participants occurred deaths. Based on the weighted analyses, most of the patients were without a partner, higher than 9th Grate or 9th Grate, smoked less than 100 cigarettes, those who don't drink, had Hypertension, use of hypoglycemic drugs and use of antihypertensive drugs. Participants were divided into a training set and a validation set in a 7:3 ratio. The mean age of individuals in the training set cohort was 58.91 ± 0.27 years old, with 2681 males (48.76%) and 2866 females (51.24%) and the mean age of individuals in the training set cohort was 59.27 ± 0.41 years old, with 1206 males (51.61%) and 1165 females (48.39%). Participants in the two groups were statistically different in Marital status, RBC and fasting glucose. The training set and validation cohort showed similar characteristics in living habits, drug use, dietary intake and chronic diseases. And the specific detailed baseline characteristics of the two group participants are summarized in Table 1 .

Flow chart of the study design.

Feature selection and development of an individualized prediction model

Table S1 shows univariable COX regression, and the results show that age, race, marital status, education levels, smoking status, alcohol consumption, CVD, hypertension, use of insulin, use of statins, use of antihypertensive drugs, use of antihyperlipidemic drugs, BMI, white blood cells, lymphocytes, neutrophils, monocytes, hemoglobin, Erythrocyte, platelets, albumin, total cholesterol, uric acid, serum creatinine, LDL, dietary intake of energy, protein intake, carbohydrate intake, fiber intake, coffee intake, calcium intake, magnesium intake, iron intake, zinc intake, NRI, UAE, UA/HDL and FTC/HDL are potential risk factors for death in the diabetic population. Spearman correlation matrix analysis was also performed on statistically the significant variables in the univariate analysis to evaluate the cross-correlation between these variables as shown in Fig. S1 . The 45 relevant features in the LASSO regression analysis are reduced to 11 potential predictors in Fig.  2 A and B (using the non-zero coefficients of Lambda.1se). These features included age, monocytes, erythrocyte count, urinary albumin excretion rate (UAE), NLR, smoking status, alcohol consumption, CVD, serum creatinine, NRI, and use of insulin (Table 2 ). Multivariate Cox regression analysis results showed that the 11 variables were independent risk factors of all causes of death in participants with diabetes as shown in Table 2 . Using these independent risk factors, we developed a nomogram prediction model for all-cause mortality risk in adults with diabetes over 3, 5, and 10 years (Fig.  3 ).

LASSO regression models to select risk factors for all-cause mortality in a diabetic population over 10 years. ( A ) 45 features of the coefficient curves of LASSO, where the coefficient curve plots are generated based on log(lambda) sequences. ( B ) Cross-validation is used in the optimal parameter (lambda) selection model in LASSO. Partial likelihood deviation (binomial deviation) curves were compared with plotted for log(lambda). The final LASSO coefficient curves for the 11 features were obtained.

Nomogram for estimating the risk of all-cause mortality in diabetes at 3,5,10 years. UAE urinary albumin excretion rate, NLR neutrophil–lymphocyte ratio, CVD cardiovascular disease, NRI National Death Index. Creatinine was categorized into 4 groups. Group 1 (Cr < 100); Group 2 (100 ≤ Cr < 200); Group 3 (200 ≤ Cr < 300); Group 4 (Cr ≥ 300). NRI was categorized into 4 groups: Group 1 (NRI < 83.5); Group 2 (83.5 ≤ NRI < 97.5); Group 3 (97.5 ≤ NRI < 100); Group 4 (NRI ≥ 100). How to use the nomogram: the position of each variable on the corresponding axis was found, a line was drawn on the point axis to indicate the number of points, the points for all variables were summed, and then a line was drawn from the total point axis to determine the probability of all-cause mortality at 3,5,10 years.

Performance of the all-cause mortality risk nomogram in adults with diabetes

Generally speaking, a C-index greater than 0.75 is considered a good predictive performance. The C-index of the nomogram was 0.790 (95% CI 0.779–0.801, P  < 0.001) and 0.792 (95% CI 0.776–0.808, P  < 0.001) in the training and validation sets, respectively. It indicates a fine accuracy in predicting all-cause mortality in an adult with diabetes. We plotted time-dependent ROC curves, the AUC value of the training set ( Fig.  4 A ) was 0.815, 0.814 and 0.827 for 3, 5, 10 years and the AUC value of the validation set for 3, 5, 10 years was 0.812, 0.818 and 0.829 ( Fig.  4 B ) , respectively. And the AUC curves for the training test and validation set are showed in Fig. S2 . Figure  5 shows the calibration curves of the model for predicting all-cause mortality risk in adults with diabetes at 3,5 and 10 years from the training and validation set. The calibration curves for the predicted all-cause mortality risk at 3, 5, and 10 years differ slightly from the ideal calibration curves in both datasets, but there is no significant difference between the predicted risk and the observed risk, in general, indicating that the prediction model is well calibrated. The DCA curve is performed on the data from the training and validation sets in Fig.  6 . To verify the clinical effectiveness of the model, we conducted a decision curve analysis. On the Y-axis, the vertical distance to the X-axis showed the standard net benefit, while on the X-axis the threshold probability for diabetes. Each graph represented the clinical usefulness of the predictive model for all-cause deaths in the participants with diabetes in different corresponding years, as shown in Fig.  6 , reflecting a good predictive efficiency. In addition, we also carried out tenfold internal crossover, and the verification results are shown in Table S2 . Finally, we also found a negative interaction between CVD and insulin use, as shown in Tables S3 , S4 and Fig. S3 .

Time-dependent Receiver Operating Curve (ROC) analysis of the Nomogram model. ( A ) Training set. ( B ) Validation set.

Calibration curves for 3, 5, and 10-year mortality risk prediction for diabetes. ( A ) Training set for 3,5, and 10 years. ( B ) Validation set 3,5, and 10 years. X-axis represents the predicted risk of morbidity. Y-axis represents the actual diagnosis of diabetes. The solid line represents the performance; the closer to the diagonal dashed line, the better the prediction (the dashed line on the diagonal represents a perfect prediction of an ideal model).

Decision curve analysis (DCA) of the nomogram for the risk of mortality over 3, 5, and 10 years in a diabetic population. ( A ) Training set for 3 years. ( B ) Training set for 5 years. ( C ) Training set for 10 years. ( D ) Validation set for 3 years. ( E ) Validation set 5 years. ( F ) Validation set for 10 years. y-axis represents net benefit. The dashed line represents the all-cause mortality risk nomogram. The thin solid line represents the assumption that all patients die. The thin solid line represents the assumption that no patients die.

Method of using nomogram

The Nomogram model is used to obtain the corresponding scores for each variable and the sum of these scores were recorded as the total score, with the highest score corresponding to 100 points. The total predicted probability was the risk of all-cause death for the participants with diabetes.

In the study, the formula for the Nomogram was as follows:

Score = (1.43*Age−21.43) + (Monocyte*10.08) + (− 2.38*Erythrocyte + 20.22) + (UAE*0.020) + (NLR*1.29) + (Score of Smoking status) + (Score of Alcohol consumption) + (Score of CVD) + (Score of Cr) + (Score of NRI) + (Score of Insulin using).

And the smoking status are classified as never (0), former (4.29), and now (12.91); Alcohol consumption are classified as never (3.73), mild (0), and moderate (6.36); CVD are classified as no (0) and yes (10.04); Cr are classified as Cr < 100 (0), and 100 ≤ Cr < 200 (5.68), 200 ≤ Cr < 300(10.54), Cr ≥ 300(14.11); NRI are classified as NRI < 83.5(17.96),83.5 ≤ NRI < 97.5(8.20), 97.5 ≤ NRI < 100(5.59), NRI ≥ 100(0); Insulin using are classified as no (0) and yes (4.64).

To the best of our knowledge, this is the first predictive model of all-cause mortality in an outpatient population with diabetes using the NHANES database. The nomograms not only have good recognition feature and higher prediction accuracy than the existing prediction tools but is easy to use. In addition, the scoring system can clearly and intuitively present the probability of clinical outcomes, which can facilitate better clinical decision-making and physician–patient communication 14 . However, only a few studies have attempted to use large national sample data to predict the risk of all-cause death in people with diabetes. In our nationally representative sample of American adults with diabetes, we developed and validated an 11-item mortality index to predict the 3-year, 5-year, and 10-year all-cause mortality risk, incorporating relevant covariates such as individual demographic characteristics, dietary condition, lifestyle, chronic disease, use of the drug, and laboratory examination indicators available in the NHANES database. We found that higher age, higher levels of Monocyte, UAE, Cr, and NLR, smoking, with a history of CVD, and use of insulin were significantly associated with increased risk of all-cause mortality, while higher levels of NRI, erythrocyte and mild alcohol consumption were associated with reduced risk of all-cause mortality. The proposed nomogram model includes only common and readily available clinical demographic data, blood analysis indicators, medical history, and medication history variables and shows good predictive performance. For example, the C index value of the training group was 0.790 (95% CI 0.779–0.801, P  < 0.001) and the verification cohort group was 0.792 (95% CI 0.776–0.808, P  < 0.001). Moreover, the calibration curve, ROC curve, and DCA curve also show good consistency between the actual observations and the predictions.

To date, many models 10 , 13 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 have been developed to predict the risk of all-cause mortality in patients with type 2 diabetes. However, predictive models developed based on one specific population are often not directly applicable to populations in different countries due to many differences in race, socioeconomic factors, and disease management. Several models 10 , 15 , 16 , 17 had been developed to predict the risk of all-cause mortality among people with type 2 diabetes in different parts of Asia, such as Hong Kong, Taiwan, Guangxi, and Shanghai. In addition, a national study reported an all-cause mortality risk prediction model for Chinese adults aged 20–80 years with diabetes 10 . Previous studies have pointed to some unique characteristics of diabetes populations in Western countries, such as a preponderant association with ethnicity, the older onset of diabetes, a higher risk of complications, and the preponderant use of medications 13 , 19 , 21 , 22 , 23 . However, predictors of all-cause mortality from diabetes developed based on the United States population covering very few ethnicities, only African Americans and European Americans 24 , 25 , 26 . Therefore, it is highly desirable to build a specific all-cause mortality risk prediction model for diabetes in the national American population to help achieve more accurate risk prediction.

In recent years, a growing number of studies 17 , 19 , 25 , 26 , 27 , 28 , 29 have reported that age, smoking habits, Cr, Monocyte, insulin use and prior history of CVD significantly affect the risk of death in persons with diabetes. In a prospective cohort study 30 , low B-containing lipoproteins to circulating erythrocytes was associated with increased mortality and cardiovascular risk, which supports our results. The association between UAE and all-cause mortality was reported in a 10-year follow-up study of residents from the Wan Shou Road community in northern China, which concluded that the urinary albumin-to-creatinine ratio was an important risk factor for all-cause mortality in this population 31 . And our results were also proved by a large study from the Netherlands 32 , which found that UAE predicted both cardiovascular and non-cardiovascular death in the general population. In addition, malnutrition is highly associated with mortality from various diseases 33 , 34 . Malnutrition is considered a global health problem that not only affects people's subjective quality of life, complication rates, and shortened life expectancy but also becomes more severe as the age pyramid changes. The NRI is a simple and accurate screening tool for nutritional status indicators 35 . Persons with diabetes are often at high risk of malnutrition due to increased nutritional requirements and severe acute inflammatory responses. Malnutrition, in turn, causes malnourished populations to be more susceptible to vascular endothelial cell injury and cardiovascular-related events because it significantly affects the body's regulation of blood glucose, cytokine parameters, and hormones 36 . Multiple studies have shown that NRI improved patients’ prediction of clinical outcomes 37 , 38 , 39 . In the study, NRI played a critical role in predicting the risk of all-cause mortality in persons with diabetes.

Diabetes is characterized by systemic chronic inflammation accompanied by increased oxidative stress. The NLR, in turn, is a good proxy for the balance between pro- and anti-inflammatory pathway activation and is often used as a proxy to quantify the degree of inflammation. Therefore, the importance of NLR level on the prognosis of diseases has been widely concerned 40 , 41 . In our statistical analysis, increased levels of NLR in the diabetic population were positively associated with all-cause mortality, which is consistent with the previous study 42 . Interestingly, we also found that it was mild and moderate alcohol consumption that significantly reduced the risk of death from diabetes. A review also noted that alcohol consumption may be associated with a reduced risk of death at light to moderate drinking levels of up to 14 g per day for women and 28 g per day for men 43 . Meanwhile, a study by Polsky and colleagues found that light to moderate alcohol consumption was associated with a lower risk of all-cause mortality in diabetes 44 .

The strengths of this study are as follows. Firstly, we use a sufficiently large nationally representative sample. Secondly, besides demographic information, we include not only lifestyle and laboratory-related indicators of chronic disease, but also dietary data on risk factors, which means that we are able to construct a comprehensive risk model that accurately predicts the risk of all-cause mortality in diabetes. Thirdly, this model obtained good prediction accuracy by cross-validation of the model, indicating strong internal validity of the results. Fourth, we also found a negative interaction between CVD and insulin use. Finally, we provide a nomogram tool that allows simple assessment of the 3, 5, and 10-year all-cause mortality risk in diabetes, which can be easily and time-effectively applied in clinical settings.

There are some limitations to this study. First, this is a retrospective study, and further prospective studies are needed at a later stage to validate our results. Second, although we included as many risk factors as possible, we cannot completely rule out the possibility of missing others. Finally, the population in this study was adults with diabetes in the United States, so the prediction model may not be applicable to populations in other countries.

The model provides a powerful tool for clinical care to identify the risk of death in persons with diabetes and facilitate timely interventions that ultimately improve the quality of survival and extend patient life.

Study design and population

This cohort study design strictly adheres to the guidelines of TRIPOD (Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis) 45 and met the framework of the prognostic Research Strategy (PROGRESS). Data from nine contiguous cycles of the National Health and Nutrition Examination Survey (NHANES) 1999–2016, a series of stratified, multistage, complex, and probability cluster samples designed by the National Center of Health Statistics of the Centers for Disease Control and Prevention aimed to assess the health status of citizens in the US, which is publicly released on a two-year cycle and all of it analyzed in the manuscript can available on the website: https://wwwn.cdc.gov/nchs/nhanes/search/default.aspx . Participants were invited to stay at home or to a mobile screening center (MEC) to answer relevant demographic information, lifestyle, and dietary questions posed by professionally trained staff, and to have their blood drawn as directed. From 1999 to 2016, the total sample size was 87,991 participants, including 50,787 (58%) who were ≥ 18 years. We selected the participants with legitimate follow-up data (50,695) and excluded people with non-diabetic (42,777), leaving a final sample of 7918. In this study, 70% of the participants in the NHANES 1999–2016 cycle were selected for the training set, while the validation set consisted of 30% of the participants from the nine cycles. Specific details about the study’s sampling and exclusion criteria are shown in Fig.  1 . The data were analyzed from January 4, 2023, to March 10, 2023. All participants signed informed consent forms and the NHANES survey protocol was approved by the National Center for Health Statistics Research Ethics Review Committee. The ethics approval numbers are Protocol #98-12, Protocol #2005-06, Continuation of Protocol#2005-06, and Protocol #2011-17. You can find it at this website: NCHS Ethics Review Board Approval (cdc.gov).

Diagnosis of diabetes

Diabetes is diagnosed by including these conditions and must meet one of the following criteria: the doctor told you have diabetes, HemoglobinA1c (HbA1c) (%) > 6.5, fasting glucose (mmol/L) ≥ 7.0, random blood glucose (mmol/L) ≥ 11.1, two-hour OGTT blood glucose (mmol/L) ≥ 11.1 and use of c or insulin. Hypoglycemic drugs include the following drugs: biguanides, sulfonylureas, thiazolidinediones, dipeptidyl peptidase 4 inhibitors, Glp-1 receptor agonists, sodium-glucose cotransporter-2, alpha-glycosidase inhibitors, and glycerides.

The follow-up all-cause mortality as of December 31, 2019, was determined from the NHANES Public-Use Association Mortality File, obtained by the National Center for Health Statistics (NCHS) in conjunction with the National Death Index (NDI) using a probabilistic matching algorithm 46 . During the entire follow-up period, any subjects who did not match any of the death records were considered alive.

Potential predictors

Potential predictive variables were evaluated based on clinical experience, scientific knowledge, and predictors identified in previously published articles 37 , 45 , 46 , 47 , 48 , 49 , 50 , 51 . The standard questionnaire collected information on demographic characteristics, habits of life, and medical history. Demographic characteristics (age, gender, ethnicity, Marital status, education levels), living habit (smoking status and drinking habits), dietary intake (energy, protein, carbohydrate, fiber, Caffeine, calcium, magnesium, iron and zinc), body measure characteristics (body mass index (BMI), Laboratory parameter (white blood cells, lymphocytes, neutrophils, monocytes, hemoglobin, Erythrocyte, platelets (PLT), uric acid (UA), urinary albumin excretion rate (UAE), serum creatinine(Cr), low-density lipoprotein (LDL), triglycerides (TG), fasting glucose (Glu), glycosylated hemoglobin, fast total cholesterol (FTC), UA/HDL, TG/HDL, FTC/HDL, neutrophils/lymphocytes (NLR), and albumin)), disease history(Hypertension, cardiovascular disease (CVD)), Medication history (hypoglycemic drugs, antihypertensive medication, statins, insulin drugs, Antihyperlipidemic drugs)were evaluated by staff. Ethnicities were categorized as non-Hispanic white, non-Hispanic black, Mexican American, or other. Marital status was classified as never married, having a partner (married or living with a partner), or without a partner (widowed, divorced, or separated). Education levels were classified as lower than 9th grade, and higher than 9th grade, or 9th grade. Smoking status was defined as the timeline of cigarettes and numbers in life: never (smoked less than 100 cigarettes), former (smoked more than 100 cigarettes in life and smoke not at all now), now (smoked more than 100 cigarettes in life and smoke some days or every day).

Participants were categorized as “never”, “mild”, and “moderate” based on the number of drinks per day he/she had drunk: Participants who drank before but don't drink now and those who never drank before were defined as “never”; “mild” were considered to be drinking alcohol ≤ 1 drink for women in 1 year and ≤ 2 drinks for men in 1 year; “moderate” were considered to be drinking alcohol ≥ 2 drinks for women in 1 year and ≥ 3 drinks for men in 1 year 52 . Data for the dietary interview portion were collected from the 24-h Recall Questionnaire, which was collected by the National Center for Health Statistics (NCHS) of the United States Department of Agriculture. BMI was calculated as weight (kg)/height (m 2 ). Blood pressure was measured by a trained technician according to a standard means, using a manual mercury sphygmomanometer to measure participants’ blood pressure, with every individual being measured three times. The laboratory parameters were obtained from the Hemal Biochemistry File. Fasting glucose 53 was monitored in the morning after a 9-h fast and after the initial venipuncture. CVD was defined as congestive heart failure/coronary heart disease/angina/heart attack or stroke. The hypertensive population was defined as the participants with a history of hypertension diagnosis or use of antihypertensive medication, or with SBP ≥ 140 mmHg or DPB ≥ 90 mmHg 54 . Antihypertensive drugs were defined as beta receptor blockers, calcium channel blockers, angiotensin receptor blockers, and diuretics. The Nutrition Risk Index (NRI) = (1.519*serum albumin (g/L)) + (41.7*body weight (kg)/ideal weight (kg)) 55 . Fasting triglyceride and glucose simple index (TyG) = Triglyceride(mg/dl) *fasting glucose(mg/dl)/2 56 .

Statistical analysis

We considered complex sampling designs and sample sizes during the data analysis in accordance with NHANES analysis guidelines 57 . And the present data can represent a sample population of 26,572,054. The missing data in selected variables were multiplied using the predictive average matching method in the chain equation algorithm, and the five estimated data sets were analyzed separately. Finally, the results were combined using Rubin's rule 58 . To compare the differences between the validation and training sets, continuous variables were expressed as weighted means with standard deviations and compared using Student's t-test; categorical variables were expressed as cases (n) and percentages (%) and compared using Rao Scott's χ2 test. Correlation analysis was performed using Person correlation for continuous variables.

In variable selection, univariate Cox-proportional hazard regression was used to select variables associated with outcome events, and uncorrelated variables are eliminated. After that, the minimum absolute contraction and selection operator (LASSO) regression (“Glmnet” package 4.2–6 edition) was used to screen the variables in the training set and non-zero coefficient variables were regarded as risk factors in the LASSO regression model. Finally, multivariate Cox-proportional hazard regression was used to screen out variables that are highly independent and correlated with outcome events. In univariate and multivariate bivariate Cox regression analysis, a two-sided P  < 0.05 was regarded as a significant difference. The penalty function in LASSO regression was beneficial to avoid overfitting and build a robust model. Based on the results of the multivariate logistic regression analysis, a nomogram was formulated. To test the accuracy of the prediction model of the nomogram, time-dependent receiver operating characteristic (ROC) curves, an area under the curve (AUC), and consistency index (C-index) were calculated using data from the training and validation sets, respectively. LASSO regression through selecting the minimum λ criterion to determine potential predictors as the plots of versus log (λ) and area under the receiver operating characteristic (AUC) value. In addition, both calibration plots and decision curve analysis (DCA) used a bootstrap method with 1000 resamples. To assign a value to each variable when creating the risk scores, we create categories for some of the continuous variables in the model. These categories were created based on the distribution of each variable. The β coefficients of the covariates of the final selected variables were used to assign points to each risk factor, which were then added together to obtain an overall score. All analyses were performed using the statistical software package R ( http://www.r-project.org ; version 4.2.2, The R Foundation) and Free Statists (version 1.9).

Ethics approval and consent to participant

This study is an analysis of the publicly available NHANES data. Informed consent was obtained from NHANES participants by the National Center for Health Statistics Research Ethics Review Board and the ethics approval number are Continuation of Protocol #98–12, Protocol #2005–06, Continuation of Protocol#2005–06, and Protocol #2011–17. You can find it at this website: NCHS Ethics Review Board Approval (cdc.gov). The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.

Data availability

All the data are available to the public and were used in the manuscript and can be available on the website: https://wwwn.cdc.gov/nchs/nhanes/search/default.aspx .

Abbreviations

National health and nutrition examination survey

Least absolute selection and shrinkage operator

Receiver operating characteristic

Decision curve analysis

Nutrition risk index

Neutrophils/lymphocytes

Cardiovascular disease

Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis

Prognostic research strategy

Mobile screening center

HemoglobinA1c

National center for health statistics

National death index

Body mass index

White blood cell

Red blood cell

Triglycerides

Urinary albumin excretion rate

Fast total cholesterol

Low-density lipoprotein cholesterol

High-density lipoprotein cholesterol

Triglyceride-glucose index

Diabetes mellitus

Dong, P.-T., Lin, H., Huang, K.-C. & Cheng, J.-X. Label-free quantitation of glycated hemoglobin in single red blood cells by transient absorption microscopy and phasor analysis. Sci. Adv. 5 (5), eaav0561 (2019).

Article   ADS   PubMed   PubMed Central   Google Scholar  

Gautheron, J. et al. The necroptosis-inducing kinase RIPK3 dampens adipose tissue inflammation and glucose intolerance. Nat. Commun. 7 , 11869 (2016).

Shaw, J. E., Sicree, R. A. & Zimmet, P. Z. Global estimates of the prevalence of diabetes for 2010 and 2030. Diabetes Res. Clin. Pract. 87 (1), 4–14 (2010).

Article   PubMed   Google Scholar  

Gregg, E. W. et al. Trends in cause-specific mortality among adults with and without diagnosed diabetes in the USA: An epidemiological analysis of linked national survey and vital statistics data. Lancet 391 (10138), 2430–2440 (2018).

Roglic, G. et al. The burden of mortality attributable to diabetes: realistic estimates for the year 2000. Diabetes Care 28 (9), 2130–2135 (2005).

Li, S., Wang, J., Zhang, B., Li, X. & Liu, Y. Diabetes mellitus and cause-specific mortality: A population-based study. Diabetes Metab. J. 43 (3), 319–341 (2019).

Article   PubMed   PubMed Central   Google Scholar  

Magliano, D. J. et al. Trends in all-cause mortality among people with diagnosed diabetes in high-income settings: A multicountry analysis of aggregate data. Lancet Diabetes Endocrinol. 10 (2), 112–119 (2022).

Selvin, E., Coresh, J. & Brancati, F. L. The burden and treatment of diabetes in elderly individuals in the U.S. Diabetes Care 29 (11), 2415–9 (2006).

Balachandran, V. P., Gonen, M., Smith, J. J. & DeMatteo, R. P. Nomograms in oncology: More than meets the eye. Lancet Oncol. 16 (4), e173–e180 (2015).

Shao, X. et al. Development and validation of a prediction model estimating the 10-year risk for type 2 diabetes in China. PLoS One 15 (9), e0237936 (2020).

Mo, M. et al. Development and validation of a prediction model for survival in diabetic patients with acute kidney injury. Front. Endocrinol. (Lausanne) 12 , 737996 (2021).

Wu, B., Niu, Z. & Hu, F. Study on risk factors of peripheral neuropathy in type 2 diabetes mellitus and establishment of prediction model. Diabetes Metab. J. 45 (4), 526–538 (2021).

McEwen, L. N. et al. Predictors of mortality over 8 years in type 2 diabetic patients: Translating Research Into Action for Diabetes (TRIAD). Diabetes Care 35 (6), 1301–1309 (2012).

Wang, H. et al. Predicting medication nonadherence risk in a Chinese inflammatory rheumatic disease population: Development and assessment of a new predictive nomogram. Patient Prefer Adher. 12 , 1757–1765 (2018).

Article   Google Scholar  

Wan, E. Y. F. et al. Prediction of five-year all-cause mortality in Chinese patients with type 2 diabetes mellitus—A population-based retrospective cohort study. J. Diabetes Complicat. 31 (6), 939–944 (2017).

Li, T.-C. et al. Development and validation of prediction models for the risks of diabetes-related hospitalization and in-hospital mortality in patients with type 2 diabetes. Metabolism 85 , 38–47 (2018).

Qi, J. et al. Developing a prediction model for all-cause mortality risk among patients with type 2 diabetes mellitus in Shanghai, China. J. Diabetes 15 (1), 27–35 (2023).

Yang, X. et al. Development and validation of an all-cause mortality risk score in type 2 diabetes. Arch. Intern. Med. 168 (5), 451–457 (2008).

Wells, B. J. et al. Predicting 6-year mortality risk in patients with type 2 diabetes. Diabetes Care 31 (12), 2301–2306 (2008).

Islam, Z. et al. Prediabetes, diabetes, and the risk of all-cause and cause-specific mortality in a Japanese working population: Japan Epidemiology Collaboration on Occupational Health Study. Diabetes Care 44 (3), 757–764 (2021).

Fernandez, C. et al. Plasma levels of the proprotein convertase furin and incidence of diabetes and mortality. J. Intern. Med. 284 (4), 377–387 (2018).

Ramírez-Prado, D. et al. Predictive models for all-cause and cardiovascular mortality in type 2 diabetic inpatients. A cohort study. Int. J. Clin. Pract. 69 (4), 474–84 (2015).

De Cosmo, S. et al. Development and validation of a predicting model of all-cause mortality in patients with type 2 diabetes. Diabetes Care 36 (9), 2830–2835 (2013).

Hayek, S. S. et al. Predicting mortality in African Americans with type 2 diabetes mellitus: Soluble urokinase plasminogen activator receptor, coronary artery calcium, and high-sensitivity C-reactive protein. J. Am. Heart Assoc. 7 (9), e008194 (2018).

Gillum, R. F., Mussolino, M. E. & Madans, J. H. Diabetes mellitus, coronary heart disease incidence, and death from all causes in African American and European American women: The NHANES I epidemiologic follow-up study. J. Clin. Epidemiol. 53 (5), 511–518 (2000).

Raffield, L. M. et al. Predictors of all-cause and cardiovascular disease mortality in type 2 diabetes: Diabetes heart study. Diabetol. Metab. Syndr. 7 , 58 (2015).

Mlekusch, W. et al. Serum creatinine predicts mortality in patients with peripheral artery disease: Influence of diabetes and hypertension. Atherosclerosis 175 (2), 361–367 (2004).

Yang, L. et al. Peripheral monocyte count is an independent predictor of all-cause mortality in type 2 diabetes with macro-vascular complications. Medicine (Baltimore) 99 (4), e18876 (2020).

Chiu, S.Y.-H., Chen, Y. I., Lu, J. R., Ng, S.-C. & Chen, C.-H. Developing a prediction model for 7-year and 10-year all-cause mortality risk in type 2 diabetes using a hospital-based prospective cohort study. J. Clin. Med. 10 (20), 4779 (2021).

de Vries, M. A. et al. Erythrocyte-bound apolipoprotein B in atherosclerosis and mortality. Eur. J. Clin. Investig. 47 (4), 289–296 (2017).

Zhang, A. et al. The relationship between urinary albumin to creatinine ratio and all-cause mortality in the elderly population in the Chinese community: A 10-year follow-up study. BMC Nephrol. 23 (1), 16 (2022).

Hillege, H. L. et al. Urinary albumin excretion predicts cardiovascular and noncardiovascular mortality in general population. Circulation 106 (14), 1777–1782 (2002).

Kaegi-Braun, N., Mueller, M., Schuetz, P., Mueller, B. & Kutz, A. Evaluation of nutritional support and in-hospital mortality in patients with malnutrition. JAMA Netw. Open 4 (1), e2033433 (2021).

Kobayashi, I. et al. Geriatric Nutritional Risk Index, a simplified nutritional screening index, is a significant predictor of mortality in chronic dialysis patients. Nephrol. Dial. Transplant. 25 (10), 3361–3365 (2010).

Perioperative total parenteral nutrition in surgical patients. N. Engl. J. Med., 325 (8) , 525-32 (1991).

Rajamanickam, A., Munisankar, S., Dolla, C. K., Thiruvengadam, K. & Babu, S. Impact of malnutrition on systemic immune and metabolic profiles in type 2 diabetes. BMC Endocr. Disord. 20 (1), 168 (2020).

Ma, X.-T. et al. Nutritional risk index improves the GRACE score prediction of clinical outcomes in patients with acute coronary syndrome undergoing percutaneous coronary intervention. Front. Cardiovasc. Med. 8 , 773200 (2021).

Correa-Rodríguez, M. et al. The prognostic nutritional index and nutritional risk index are associated with disease activity in patients with systemic lupus erythematosus. Nutrients 11 (3), 638 (2019).

Jiao, Z. et al. Prognostic utility of nutritional risk index in patients with head and neck soft tissue sarcoma. Nutrients 15 (3), 641 (2023).

Scilla, K. A. et al. Neutrophil-lymphocyte ratio is a prognostic marker in patients with locally advanced (Stage IIIA and IIIB) non-small cell lung cancer treated with combined modality therapy. Oncologist 22 (6), 737–742 (2017).

King, A. H. et al. Elevated neutrophil-lymphocyte ratio predicts mortality following elective endovascular aneurysm repair. J. Vasc. Surg. 72 (1), 129–137 (2020).

Lee, S. et al. Development of a predictive risk model for all-cause mortality in patients with diabetes in Hong Kong. BMJ Open Diabetes Res. Care 9 (1), e001950 (2021).

Hendriks, H. F. J. Alcohol and human health: What is the evidence?. Annu. Rev. Food Sci. Technol. 11 , 1–21 (2020).

Polsky, S. & Akturk, H. K. Alcohol consumption, diabetes risk, and cardiovascular disease within diabetes. Curr. Diab. Rep. 17 (12), 136 (2017).

Collins, G. S., Reitsma, J. B., Altman, D. G. & Moons, K. G. M. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): The TRIPOD statement. BMJ 350 , g7594 (2015).

Wolffenbuttel, B. H. R., Heiner-Fokkema, M. R., Green, R. & Gans, R. O. B. Relationship between serum B12 concentrations and mortality: Experience in NHANES. BMC Med. 18 (1), 307 (2020).

Katsimpris, A. et al. Prediction of type 2 diabetes mellitus based on nutrition data. J. Nutr. Sci. 10 , e46 (2021).

Alkhatatbeh, M. J., Hmoud, Z. L., Abdul-Razzak, K. K. & Alem, E. M. Self-reported sleep bruxism is associated with vitamin D deficiency and low dietary calcium intake: A case-control study. BMC Oral Health 21 (1), 21 (2021).

Quan, J. et al. Development and validation of the CHIME simulation model to assess lifetime health outcomes of prediabetes and type 2 diabetes in Chinese populations: A modeling study. PLoS Med. 18 (6), e1003692 (2021).

Deberneh, H. M. & Kim, I. Prediction of type 2 diabetes based on machine learning algorithm. Int. J. Environ. Res. Public Health 18 (6), 3317 (2021).

Xiao, Y. et al. Elevated serum creatinine levels and risk of cognitive impairment in older adults with diabetes: A NHANES study from 2011–2014. Front. Endocrinol. (Lausanne) 14 , 1149084 (2023).

Rattan, P. et al. Inverse association of telomere length with liver disease and mortality in the US population. Hepatol. Commun. 6 (2), 399–410 (2022).

Xie, X. et al. Association between triclocarban and triclosan exposures and the risks of type 2 diabetes mellitus and impaired glucose tolerance in the National Health and Nutrition Examination Survey (NHANES 2013–2014). Environ. Int. 136 , 105445 (2020).

Chobanian, A. V. et al. The seventh report of the joint national committee on prevention, detection, evaluation, and treatment of high blood pressure: The JNC 7 report. JAMA 289 (19), 2560–2572 (2003).

Przekop, Z. et al. Efficacy of the nutritional risk index, geriatric nutritional risk index, BMI, and GLIM-defined malnutrition in predicting survival of patients with head and neck cancer patients qualified for home enteral nutrition. Nutrients 14 (6), 1268 (2022).

Minh, H. V. et al. Assessment of preferred methods to measure insulin resistance in Asian patients with hypertension. J. Clin. Hypertens. (Greenwich) 23 (3), 529–537 (2021).

Ruan, Z. et al. Association between psoriasis and nonalcoholic fatty liver disease among outpatient US adults. JAMA Dermatol. 158 (7), 745–753 (2022).

Pabinger, I. et al. A clinical prediction model for cancer-associated venous thromboembolism: a development and validation study in two independent prospective cohorts. Lancet Haematol. 5 (7), e289–e298 (2018).

Download references

Acknowledgements

We thank the researchers, staff, and participants of the National Health and Nutrition Examination Survey for their contributions to making this analysis possible. We also thanks to Zhang Jing (Shanghai Tongren Hospital) for his work on the NHANES database. His outstanding work, nhanesR package and webpage, makes it easier for us to explore NHANES database.

Author information

These authors contributed equally: Xia Shen, Xiao Hua Zhang and Long Yang.

Authors and Affiliations

Department of Nursing, School of Health and Nursing, Wuxi Taihu University, 68 Qian Rong Rode, Bin Hu District, Wuxi, China

Cardiac Catheter Room, Wuxi People’s Hospital, Jiangsu, No.299 Qing Yang Road, Wuxi, 214000, China

Xiao Hua Zhang

Department of Pediatric Cardiothoracic Surgery, The First Affiliated Hospital of Xinjiang Medical University, 137 Li Yu Shan Road, Urumqi, 830054, China

Department of Traditional Chinese Medicine, Fuzhou University Affiliated Provincial Hospital, 134 East Street, Gu Lou District, Fuzhou, 350001, China

Peng Fei Wang

Research and Teaching Department, Taizhou Hospital of Integrative Medicine, Jiangsu Province, No. 111, Jiang Zhou South Road, Taizhou City, Jiangsu, China

Jian Feng Zhang

Department of Radiology, The Convalescent Hospital of East China, No.67 Da Ji Shan, Wuxi, 214065, China

Department of Basci, School of Health and Nursing, Wuxi Taihu University, 68 Qian Rong Rode, Bin Hu District, Wuxi, China

Shao Zheng Song

You can also search for this author in PubMed   Google Scholar

Contributions

Conceptualization, X.S.; methodology, X.S.; software, X.S., L.Y.; validation, L.Y; formal analysis, X.S., L.Y.; investigation, X.S., X-H-Z, L.Y.,; resources, X.S; data curation, X.S., L.Y and X-H-Z.; writing—original draft preparation, X.S.; writing—review and editing, X.S., X-H-Z., L.Y., J-F-Z and P-F-W; visualization, X.S., X-H-Z., J-F-Z and P-F-W.; supervision, L.J and S-Z-S. All authors have read and agreed to the published version of the manuscript.

Corresponding authors

Correspondence to Shao Zheng Song or Lei Jiang .

Ethics declarations

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.

Supplementary Information

Supplementary legends., supplementary figure s1., supplementary figure s2., supplementary figure s3., supplementary table s1., supplementary table s2., supplementary table s3., supplementary table s4., 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.

Shen, X., Zhang, X.H., Yang, L. et al. Development and validation of a nomogram of all-cause mortality in adult Americans with diabetes. Sci Rep 14 , 19148 (2024). https://doi.org/10.1038/s41598-024-69581-3

Download citation

Received : 28 September 2023

Accepted : 06 August 2024

Published : 19 August 2024

DOI : https://doi.org/10.1038/s41598-024-69581-3

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

• LASSO regression

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.