We should nudge clinicians and trainees to participate in health professions education programmes

Abstract

Health Professional Education (HPE) programmes, such as mentorship, are widely regarded as being advantageous to the personal and professional development of clinicians and trainees. Involvement in a mentoring relationship is associated with positive outcomes for both mentees and mentors, including improved career preparation, increased career success, higher job satisfaction and reduced risk of burnout. Despite these data, a minority of trainees report having a mentor.

In this Cross‐Cutting Edge article, the authors focus on an impediment to participation in HPE programmes that they feel are both highly prevalent and modifiable: habit. Taking the example of mentorship, they use dual processing as their theoretical framework and describe how we use both System 1 and System 2 processing to make decisions that, in turn, promote habitual and goal‐directed actions, respectively.

The authors discuss the relationship between habitual and goal‐directed actions and suggest that habits can both facilitate and hinder our goals. Drawing on the clinical literature on adherence to clinical practice guidelines, they describe how habits and contextual factors can interfere with clinical goals and how manipulating the clinical environment can move behaviour in the desired direction. They then branch into behavioural economics to describe the features of a nudge (and a sludge) and review the literature on the effectiveness of this type of intervention – including potential ethical concerns around the use of nudges as behavioural interventions. Using the MINDSPACE mnemonic/framework they suggest different types of transparent and non‐transparent nudges that could be used to increase participation in mentorship.

Recognizing that mentorship is complex and the impact of a single nudge on behaviour may be ineffective or wane over time, the authors propose a process of ongoing programme evaluation and quality improvement that could help create and maintain a culture of mentorship and that can also be applied to other HPE programmes.

Short abstract

Drawing on behaviour al economics, Harper et al. suggest ways in which we might turn mentoring into a habit to increase the rate at which trainees receive the suport they need.

1. INTRODUCTION

“You cannot force commitment, what you can do … You nudge a little here, inspire a little there, and provide a role model. Your primary influence is the environment you create.”

Peter Senge

The literature on healthcare professions education (HPE) is replete with evidence‐based recommendations on behaviours that clinicians could adopt to improve their personal well‐being and clinical performance, in addition to strategies that should enhance the performance of their trainees and their multidisciplinary care team. ¹ , ² , ³ , ⁴ , ⁵ , ⁶ And, in most academic centres there is the opportunity for clinicians to participate in mentorship programmes, wellness initiatives, faculty development, continuing professional education, and training in interprofessional collaboration designed to facilitate the desired behaviour changes. Yet, despite their availability, uptake of HPE programmes is disappointingly low. ³ , ⁴ , ⁷ , ⁸ , ⁹ If we consider the example of mentorship, this is now widely acknowledged as being advantageous to the personal and professional development of trainees and clinicians, and involvement in a mentoring relationship associates with positive outcomes for both mentees and mentors, including improved career preparation, increased career success, higher job satisfaction and reduced risk of burnout. ³ , ⁴ , ⁷ , ¹⁰ , ¹¹ But, regardless of these data, a minority of trainees report having a mentor. ³ , ⁴ , ⁷ So, why are so few clinicians and trainees involved in programmes such as mentorship, and how can we narrow this value‐action gap? ¹²

Rather than addressing all potential programmes, in this Cross‐Cutting Edge article, we will focus on mentorship as the prototypical underutilized HPE intervention that is available to most clinicians and trainees. And, instead of compiling a comprehensive list of potential mentorship barriers, we will address one impediment to participation in mentorship (and other HPE programmes) that we believe is highly prevalent and modifiable: habit. We use dual‐processing theory as the framework for our discussion and begin by describing how we use both System 1 and System 2 processing to make decisions that, in turn, promote habitual and goal‐directed actions, respectively. We then discuss the relationship between habitual and goal‐directed actions and suggest that habits can both facilitate and hinder our goals. We review the clinical literature on adherence to clinical practice guidelines to illustrate how habits and contextual factors can interfere with clinical goals and how manipulating the clinical environment can move behaviour in the desired direction. We then venture into the field of behavioural economics as we discuss how nudges can consciously or subconsciously change behaviour and how these could be used to narrow the value‐action gap in mentorship. We describe different categories of nudges, organized around the MINDSPACE framework, ¹³ and suggest ways in which these could be applied in the context of mentorship with the goal of increasing participation by both mentors and mentees – while at the same time acknowledging some of the concerns with using nudges as behavioural interventions. Finally, recognizing that mentorship is complex and the impact of a single nudge on behaviour may be ineffective or wane over time, we propose a process of ongoing programme evaluation and quality improvement that could help create and maintain a culture of mentorship among clinicians and trainees. ¹⁴

2. DUAL‐PROCESSING THEORY, HABITS AND GOAL‐DIRECTED ACTIONS

Based upon observational and experimental data spanning a period of more than 50 years, it is now commonly accepted that when we make decisions, we can access two cognitive processes. ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ The first is an automatic, subconscious and rapid process that is triggered by context and emotional reactions and is referred to as System 1 processing (or non‐analytic information processing). ¹⁸ By contrast, System 2 processing (or analytic information processing) is optional and intentional and is typically slower due to the fact that this involves reflection and logic with or without the integration of personal values and intentions. ¹⁸ Rather than being restricted to a specific decision‐making domain, dual‐processing appears to be ubiquitous with data supporting some version of dual‐processing theory found in all areas of psychology, sociology and related disciplines, including economics and clinical medicine. ¹⁶ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² While System 1 processing is shared with other animals, System 2 processing is thought to be unique to humans and some have argued that the cognitive advantage that humans have over other animals is largely attributed to System 2 processing being able to suppress the System 1 default and allow for ‘hypothetical thinking’. ²³ Yet, this does not imply the inherent superiority of System 2 processing. ²⁴ , ²⁵ Instead, our cognitive ascendancy more likely relates to having two forms of processing that complement each other and offset the limitations of the other. ²⁶

System 1 and 2 processing results in corresponding behaviours that are typically referred to as habitual versus goal‐directed actions, respectively. ²⁷ Habits are learned automatic responses that are heavily context‐dependent. They are activated by contextual cues, reinforced by repetition in association with the same cues and, by virtue of diminishing cognitive input, habits become increasingly inflexible and insensitive to changing goals. ²⁷ , ²⁸ By contrast, goal‐directed behaviour is shaped by the predicted outcome of a certain behaviour in a specific context. Rather than being determined by the context, goal‐directed behaviour adapts to the context in a way that is designed to achieve the desired outcome. ²⁷ In the real world, we continually switch between habits and goal‐directed behaviour, and a prior study of self‐reported behaviours estimated that close to 50% of the actions that we perform on a daily basis are habitual. ²⁸ , ²⁹

3. THE RELATIONSHIP BETWEEN HABITS AND GOALS

Studies using functional neuroimaging show that habits and goal‐directed actions use distinct neural pathways: goal‐directed actions involve neural connections to the prefrontal and orbitofrontal cortex where working memory is located, whereas habits involve activation of a sensorimotor loop without input from working memory. ²⁷ , ³⁰ , ³¹ The coexistence of two pathways leading to different actions is presumably related to the fact that each pathway offers potential advantages and disadvantages. The pros of goal‐directed actions are self‐explanatory, but pursuing goals comes with a higher cognitive load and the risk of overloading working memory capacity. ³² By contrast, automatically repeating behaviours in recurring contexts allows us to perform actions with minimal cognitive load – so habits can help us achieve our goals by freeing up working memory for this purpose. Many of our habits actually evolve from goal‐directed actions and through experience we discover that performing certain actions in distinct contexts produces specific rewards – so we continue to perform these actions in these contexts and gradually these migrate from being actions regulated by working memory to ones that are automated within the sensorimotor loop. ³³ A potential downside to this transition is that these actions may become entrenched through reinforcement and persist even if the rewards diminish or become inconsistent – in which case habits are maladaptive and can prevent us from achieving our goals. ³⁴

Currently, there is a paucity of data explaining why clinicians and trainees do not pursue involvement in mentorship and other HPE interventions. However, there are data to suggest that maladaptive habits can prevent these groups from achieving other career goals.

4. HOW HABITS CAN PREVENT US FROM ACHIEVING CLINICAL GOALS

Clinical practice guidelines (CPGs) are evidence‐based, systematically developed statements designed to help clinicians make the best healthcare choices for their patients. Thus, if our intention is to improve patient outcomes, then adherence to CPGs could be viewed as a measure of our goal‐directed actions. Yet, adherence to CPGs has always been rated as suboptimal, implying that clinicians' actions are not consistently goal‐directed. ³⁵ (This statement is an oversimplification since clinicians may disagree with guidelines, in which case their goal is to not follow the guideline, but adherence to CPGs is generally accepted as an indicator of good clinical care.) A study by Cabana and colleagues published more than 25 years ago identified “inertia from previous practice” as a contributory factor to non‐adherence in more than 40% of studies and linked these habits to “environmental factors” that included limited time and resources, lack of incentive to change and organizational constraints. ³⁶ A more recent systematic review identifies similar environmental factors but added emotional factors, including feeling “overloaded” by the amount of information in CPGs and experiencing “guideline fatigue.” ³⁷ These data not only suggest that habits frequently interfere with goal‐directed actions, but they also suggest that contextual factors reinforce habits since time limitations, feeling overloaded and lack of an incentive to change are not conducive to system 2 processing and goal‐directed actions. ¹⁸ The importance of contextual factors on actions is also supported by the observation that continuing medical education that targets clinicians' knowledge is unlikely to change behaviour unless this is combined with context‐specific interventions designed to enable and reinforce behaviour change. ³⁸ , ³⁹

5. FOLLOWING THE MONEY: HOW WE CAN LEARN FROM BEHAVIOURAL ECONOMICS

Since maladaptive habits are pervasive, we can learn from research in other areas of ways to move behaviour in a desired direction. Behavioural economics is a field positioned at the intersection between psychology and economics that explores how we make decisions in real life. ⁴⁰ , ⁴¹ Whereas traditional economic models assume that decisions are rational and based upon self‐interest (System 2 processing), behavioural economics recognizes that choices are impacted by other variables, such as context (including time constraints), emotions, cognitive limitations, complexity/cognitive load and biases (System 1 processing). ²⁰ , ⁴² Consequently, rather than being entirely rational, real‐life decisions are typically made under conditions of “bounded rationality” where we integrate rational thought with cognitive shortcuts that allow us to adapt to other variables that impact decision‐making. ⁴³ , ⁴⁴ , ⁴⁵ And, by offering a more comprehensive understanding of variables that impact decision‐making, behavioural economics can suggest strategies to influence decision‐making and behaviour beyond the traditional approach of providing knowledge and anticipating an appropriate [rational] response. One intervention rooted in behavioural economics that has garnered a lot of attention over the past two decades is The Nudge.

6. THE NUDGE

The concept of nudging was popularized in the early 2000s by Richard Thaler and Cass Sunstein. In their book, ‘Nudge: Improving Decisions about Health, Wealth, and Happiness’, they define a nudge as “any aspect of the choice architecture that alters people's behavior in a predictable way without forbidding any options or significantly changing their economic incentives.” ⁴⁶ Since the publication of this influential text, nudging as a behavioural intervention has become mainstream in many areas of society, ⁴⁷ and a systematic review of the literature on the effectiveness of nudging in a variety of contexts found that this type of intervention typically has a small‐to‐medium effect size (Cohen's d = 0.45). ⁴⁸ Multiple studies of nudging in a healthcare setting have also demonstrated that goal‐directed actions can be increased, and examples of successful nudges include placing prominent inserts in a patient chart (increased rate of mammography screening and CPG‐adherent prescribing) ⁴⁹ , ⁵⁰ or hand sanitizer in each consultation office. ⁵¹ A systematic review of 42 trials of healthcare nudges found that almost 90% moved the behaviour of healthcare providers in the desired direction with a median effect size of 0.39 for continuous outcomes and an odds ratio of 1.62 for dichotomous outcomes. ⁵²

7. NUDGING THOSE WHO SHOULD BE INVOLVED IN MENTORSHIP

If habits, such as non‐participation in mentorship, are initiated by features of the choice architecture and automated within a sensorimotor loop, then we can categorize nudges into those that assimilate working memory/System 2 processing into the choice architecture versus those that interrupt the initiation of habits (Figure 1). Nudges that incorporate System 2 processing are referred to as transparent nudges, and these are designed to encourage goal‐directed actions. For example, we may be more likely to set a goal and then commit to involvement in mentorship when we discover that a highly esteemed colleague attributes most of their early career success to the mentoring they received during their training and describes how their current role as a mentor is both motivating and rewarding (messenger nudge). ¹³ Similarly, when we discover that most of the other busy clinicians in our department have one or more mentees that they meet with on a regular basis, we might decide that we actually do have something to offer trainees and can find time in our schedule to meet with at least one trainee (social norm nudge). Non‐transparent nudges target aspects of the choice architecture that initiate the sensorimotor loop, and an example of this type of nudge would be changing our process of mentor recruitment so that clinicians are automatically included in the mentorship programme unless they choose to opt out (default nudge). Non‐transparent nudges can also target the emotions that trigger the sensorimotor loop and by repeatedly exposing potential participants to positive testimonials from those involved in mentoring we might be able to change the choice architecture to one where the term mentorship evokes a positive emotional reaction and, ultimately, willingness to participate.

FIGURE 1.

Proposed mechanism of action of transparent and non‐transparent nudges.

The literature on mentoring and social influence suggests that we are typically drawn to individuals that we perceive as being similar to ourselves. ⁵³ , ⁵⁴ Thus, when designing both transparent (e.g., messenger, social norm) and non‐transparent (e.g., priming, affect, ego) nudges, we should include diverse role models/individuals in order to promote inclusion, particularly of individuals who are traditionally less likely to report being involved in mentoring relationships. ⁵⁵

In Table 1 we use the MINDSPACE mnemonic as a framework to describe nudges that we could consider if trying to increase participation in our mentorship programme. ¹³ We also highlight the specific behavioural tendency that each nudge aligns with. In terms of the ideal number of nudges and synergy between nudges, there are limited data on this, but the literature on CPG adherence suggests that multiple interventions are typically better than a single intervention, ³⁹ and it is likely that combinations of nudges are more effective when they target complementary mechanisms – so it would seem logical to add a transparent nudge alongside a nudge that is designed to interrupt the sensorimotor loop.

TABLE 1.

Nudge types, behavioural tendencies, and how these could be adapted to mentorship.

Nudge type*	Behavioural tendency	Mentor nudge	Mentee nudge
Messenger (T)	We are more likely to change behaviour if someone we know and admire suggests this	A clinician role model describes their experiences as both mentor and mentee and recommends participation	A trainee role model describes the contribution of mentorship to their success and recommends participation
Incentive (NT)	We typically have a strong desire to avoid loss	Mentorship is recognized on annual reports, decisions on faculty awards, and in promotion decisions	A trainee retreat is held for trainees who have participated in the mentorship programme
Norms (T)	We are naturally influenced by the behaviour of others	All clinicians receive a communication that names and thanks those who participate in mentorship	All trainees receive a communication that names and thanks those who participate in mentorship
Defaults (NT)	We generally go with the flow and choose to preselected option	All clinicians are included in the mentorship programme unless they choose to opt out	All trainees are included in the mentorship programme unless they choose to opt out
Salience (T)	We are drawn to choices that grab our attention, are novel, and seem relevant	Mentorship programme updates and opportunities are included at the start of the departmental bulletin	Mentorship programme updates and opportunities are included at the start of the academic half‐day
Priming (NT)	Our choices are typically influenced subconsciously by contextual cues	“I am also a mentor” posters are advertised in clinics, elevators, and departmental offices	“I am also a mentee” posters are advertised in clinics, elevators, and resident lounges
Affect (NT)	Our choices are typically influenced subconsciously by emotional associations	Clinicians share their experiences of being a mentee and/or mentor at trainee/faculty events	Trainees share their experiences of being a mentee and/or mentor at trainee/faculty events
Commitments (T)	We generally strive to keep public promises and reciprocate positive actions	At events promoting mentorship, clinicians use a QR code to declare their intention to mentor	At events promoting mentorship, trainees use a QR code to declare their intention to seek mentorship
Ego (NT)	We mostly prefer to choose actions that make us feel better about ourselves	“Thanks for being a mentor” posters are advertised in clinics, elevators, and departmental offices	“Thanks for being a mentee” posters are advertised in clinics, elevators, and trainee lounges

^*T = Transparent nudge, NT = Non‐Transparent nudge.

8. CRITICISMS OF NUDGING

It is important to recognize that not all nudges are effective and that some may actually have a negative effect on behaviour. This near enemy of the nudge is commonly referred to as The Sludge. ⁵⁶ , ⁵⁷ The risk of sludging is increased when our intervention runs counter to behavioural tendencies; for example, if it adds complexity or friction to the choice architecture, if it appears to benefit an organization rather than the decision maker, or when an attempted nudge is perceived as being manipulative or coercive. ⁵⁷ Consider the following two‐step process for recruiting clinicians as volunteer mentors: 1) all potential mentors complete an application form and submit this to the mentorship selection committee along with their curriculum vitae, personal letter and reference letters from their department head and one other colleague; and 2) after considering all applications, the mentorship selection committee invites potential mentors for an interview, after which the committee decides on the final list of mentors to be included in their programme in addition to creating a “reserve list” of mentors. To mitigate the risk of sludging, the Behavioural Insight Team in the UK created another mnemonic (EAST) to describe desirable attributes of a nudge: make it Easy (simple messaging, reduce the hassle of making the desired choice), Attractive (to the individual choosing); Social (highlight that most individuals perform the desired behaviour) and Timely (introduce when individuals are receptive and consider short‐term rewards). ⁵⁸ When we apply the EAST mnemonic to the recruitment process described above (which is a process frequently used to recruit applicants to competitive positions), we can readily appreciate why this type of intervention is unlikely to nudge busy clinicians towards participation in mentorship.

The ethics of nudging have also come under scrutiny, with critics arguing that this employs manipulation and exploits the power and position of a select few: the choice architects. ⁵⁹ , ⁶⁰ , ⁶¹ , ⁶² Arguments to counter these concerns include 1) all choices occur within a specific context or environment such that all decisions are already nudged, even if this is not by design ⁴⁶ ; 2) rather than forcing a specific behaviour or removing a pre‐existing choice, nudging simply changes the likelihood that a “more desired” choice (such as participation in mentorship) will be made; 3) nudging is, by design, a compromise intervention that balances respect for an individual's choice with the potential advantages to this individual and/or society (including mentorship programmes) from influencing this choice; and 4) a more detailed analysis suggests that most health‐promoting nudges, whether these involve placing soda on the bottom shelf of a fridge or having an opt‐out system for organ donation, do not violate important ethical principles, such as autonomy. ⁶³ So, although there are undoubtedly concerns regarding the effectiveness and the ethics of nudging, we feel that, in principle, it is appropriate to introduce and then evaluate the effectiveness of nudges as interventions designed to increase participation in mentorship and other HPE interventions.

9. EVALUATING AND ADAPTING TO SLUDGES AND WANING NUDGES

As discussed above, nudges have the potential to be ineffective or even counterproductive. Also, since mentorship is complex and the decision to participate in mentorship can be impacted by multiple external variables, choices regarding mentorship will inevitably change over time. ¹⁴ When viewed through the lens of complexity theory, this apparent waning of a nudge's effectiveness is interpreted as a need for new nudges or other interventions rather than nudge failure. ⁶⁴ Based upon this perspective, we recommend that when we use nudges as a programme intervention to increase participation in mentorship then we also need to evaluate their effectiveness on an ongoing basis, for example by using Stufflebeam's Context Input Product Process programme evaluation model or realist evaluation, both of which are appropriate for evaluation of complex adaptive systems. ¹⁴ , ⁶⁵ , ⁶⁶ , ⁶⁷ , ⁶⁸ This way, we are more likely to prevent widening of the value‐action gap and, hopefully, create and maintain a culture of mentorship. ⁶⁹

10. CONCLUSION

In the movie, Fields of Dreams, the character played by Kevin Costner is moved to act by a whisper in a cornfield, “If you build it, he will come”, and so he builds a baseball diamond and indeed they do come. Unfortunately, this simple strategy has not worked for CPGs and is unlikely to achieve the desired participation in mentorship or other HPE interventions. In the excellent articles that provide tips on how to build a mentorship programme, the issue of how to increase participation is not addressed directly. ⁷⁰ , ⁷¹ , ⁷² It is obviously important to build the best possible mentorship programme, but in order for this to succeed we need to include interventions that are designed to increase the participation of overloaded clinicians and trainees who have a habit of non‐participation. Nudges are mostly effective, in most settings they are ethically acceptable, and they allow us to achieve a balance between influencing choice and respecting choice. Thus, we feel that nudges are interventions that can help bridge the value‐action gap and that if we build it – and then nudge them – we will have a more successful mentorship programme. And, since dual‐processing is ubiquitous and habits almost certainly contribute to low rates of participation in other HPE programmes, we would predict that by including nudges as a programme intervention we can increase participation in other HPE programmes that can then benefit clinicians, trainees and other members of the multidisciplinary care team.

AUTHOR CONTRIBUTIONS

Lea Harper: Conceptualization; writing—original draft; methodology; writing—review and editing. Sylvain Coderre: Conceptualization; methodology; writing—review and editing. Kirstie Lithgow: Conceptualization; methodology; writing—review and editing. Kenna Kelly‐Turner: Conceptualization; writing—review and editing; methodology. Melinda Davis: Conceptualization; writing—review and editing; methodology. Kevin McLaughlin: Conceptualization; writing—review and editing; writing—original draft; methodology.

CONFLICT OF INTEREST STATEMENT

None of the authors have a conflict of interest to report.

ETHICS STATEMENT

Ethics approval was not sought as there was no intervention or data reported in this manuscript.

Harper L, Coderre S, Lithgow K, Kelly‐Turner K, Davis M, McLaughlin K. We should nudge clinicians and trainees to participate in health professions education programmes. Med Educ. 2025;59(12):1333‐1340. doi: 10.1111/medu.15749

DATA AVAILABILITY STATEMENT

Data sharing not applicable to this article as no datasets were generated or analysed during the current study.