Integrating Insights from Implementation Science and Behavioral Economics to Strengthen Suicide Screening Strategies for Pediatric Populations

Abstract

Suicide is a leading cause of death for young people globally. Professional organizations such as the American Academy of Pediatrics (AAP) recommend suicide screening for every young person at every healthcare contact and these guidelines are now being implemented across the US. To optimize deployment of these guidelines, we draw on insights from two parallel fields, implementation science – or the study of how to support clinicians in providing guideline-concordant care within organizational constraints – and behavioral economics, which centers human decision-making. In this brief review, we offer potential barriers to suicide screening and strategies to address these barriers drawing on these two literatures, illustrating how best to integrate insights from these fields for maximal impact.

The know-do gap in suicide screening

Suicide is the fourth leading cause of death among young people globally [1]. Early identification of people who are at risk for suicide through screening can help save lives. Leading professional organizations such as the American Academy of Pediatrics (AAP) recommend screening young people for suicidal thoughts and intentions at every healthcare contact [2]. For example, one brief screening tool recommended by AAP is the 4-question Ask Suicide-Screening Questions (ASQ) tool [3]. Healthcare systems are now implementing these recommendations in a range of settings across the United States [4–7]*.

Despite the expectation of routine screening among young people, patient-, clinician-, organizational-, and system-level barriers can undermine the successful implementation of screening for suicide in healthcare settings, potentially missing opportunities to save lives. Mismatches between guidelines and practice are often described as ‘know-do gaps’ and suggest important areas for behavior change strategies to be used in healthcare. Implementing recommended suicide screening practices remains one such know-do gap, thus making it ripe for application of insights from behavioral economics and implementation science [8].

Promise of implementation science and behavioral economics for bridging this gap

Implementation science (IS) is “the study of methods to promote the adoption and integration of evidence-based practices” [9]. Insights from IS can inform the selection and design of strategies that can be used to encourage and support clinicians and organizations to incorporate universal suicide screening into routine practice [8]. Insights from behavioral economics (BE), or the study of how people make decisions, can also contribute to suicide prevention efforts by boosting our understanding of how to support clinicians and organizations in behavior change [10]. Behavioral economics posits that people make hundreds of decisions each day with incomplete information and limited time (i.e., bounded rationality). This has led humans to develop predictable patterns of decision-making that allow us to navigate the high cognitive load efficiently [11]. BE offers helpful insights on how to design interventions that account for this ‘bounded rationality’ [12]. Interventions from BE include “nudging,” which gently steers people towards certain decisions without limiting their ability to make a free choice; “boosting,” which makes it easier for people to understand relevant information; and “de-sludging,” which removes the “sludge” – excessive friction, confusion, and frustration – which makes it difficult for people to make optimal choices [13–16].

Emerging evidence suggests that the effectiveness of implementation strategies can be increased through the incorporation of insights from behavioral economics [8]. Both fields approach behavior change from different angles with the shared aim of improving health and well-being [17, 18]. However, insights from the two fields have rarely been combined in studies of implementation strategies for the promotion of suicide prevention efforts. A recent scoping review of suicide prevention interventions included 32 studies using a range of implementation strategies [19]. The authors identified 109 education-based strategies but none explicitly drew on insights from behavioral economics. These findings illustrate that insights from behavioral economics have not yet been integrated into implementation strategies in this context, providing an opportunity to strengthen approaches to implementing suicide screening.

In this brief review, we identify opportunities at the intersection of IS and BE to promote implementation of universal suicide screening in healthcare settings for young people. We begin by describing evidence showing that information-giving alone is insufficient to increase suicide screening. Then, we present barriers and potential implementation strategies to target these barriers. We structure this review using the EAST (Easy, Attractive, Social, Timely) framework, which is a parsimonious approach from the United Kingdom’s Behavioral Insights Team describing key principles from behavioral economics to inform strategy design, although other alternatives are also available such as MINDSPACE and the Taxonomy of Choice Architecture Techniques [12, 20, 21]. As universal screening for suicide among children is a newer recommendation, the majority of the research we draw on comes from targeted suicide screening aimed at people in primary care or psychiatric contexts. We conclude with next steps to actualize the promise of this combined approach.

Behavioral economics-informed implementation strategies to promote clinician action

Conventional approaches to behavior change center education and information-giving, but evidence from BE reveals that information alone is neither necessary nor sufficient for behavior change [22, 23]*. Although clinicians report high levels of training around discussing suicide, implementation remains low, suggesting that educational interventions alone do not change clinician screening behavior [24, 25]. Insights from BE can be used to inform the design of implementation strategies aimed at increasing healthcare provider delivery of suicide screening. For example, the BE literature suggests education-only interventions may be ineffective because they fail to account for factors that drive clinicians to turn their knowledge into action in the clinical encounter (i.e., complete screening during a clinical encounter).

Make the behavior easy.

Some clinicians find screening difficult and do not feel confident in their ability to ask about suicide, increasing the risk that they might avoid these potentially difficult conversations [26, 27]. Exposure-based roleplay may improve clinician confidence by allowing them to encounter the hardest scenarios under controlled conditions and revealing that they can handle them, reducing avoidance [28, 29]. Clinic workflow and processes can also make suicide screening difficult. “Sludge” can reduce clinician willingness to engage in evidence-based approaches, particularly when they face high cognitive burdens [16, 30, 31]. Clinicians have reported that barriers to completing suicide screening include becoming frustrated by completing paper forms which require multistep processing [32]. One approach to removing this ‘sludge’ is to simplify documentation either by adding a series of boxes to the electronic health record (EHR) where amending the EHR is feasible or by creating a short free text phrase for clinicians to use as an alternative. Specifically, clinicians at an organization could document that ASQ was conducted with no positive findings by simply writing “ASQ Negative”, thus quickly leaving an auditable record.

Although opportunities to simplify and streamline paperwork have not yet been subjected to empirical study, several studies have aimed to make screening easier to remember. Pop-ups such as best practice alerts in the EHR are used in a range of settings to alleviate the cognitive burden of recalling a practice which may not organically emerge in a given clinical encounter [26, 27]. For example, suicide screening prompts were tested in a pre-post study finding no significant effect [33]. In another study, wallet cards were disseminated to make it easier for clinicians to follow a new suicide screening protocol, but the behavior change impact was not measured [34]. These information-giving interventions do not reliably lead to behavior change, perhaps because they are not well-matched to the factors which make screening difficult in this case, or because EHR nudges such as best practice alerts risk frustrating clinicians and backfiring [35*, 36]. Therefore, we suggest careful matching of strategies to barriers and continued empirical testing before wider roll-out.

Make the behavior attractive.

Another barrier clinicians report is that suicide screening is not a reinforcing or attractive behavior. Specifically, there is a lack of feedback so clinicians are uncertain about the appropriateness of interventions, and this uncertainty contributes to low screening rates [37]. Some have argued that suicide screening is a low priority for clinicians because averted suicide is ‘silent’ and screening is not positively reinforced [38]. Providing positive feedback and support to improve practice could, therefore, make screening more attractive to clinicians.

Make the behavior social.

Most mental health care appointments take place one-on-one, creating uncertainty about exactly what questions other clinicians ask during assessments, including whether they habitually inquire about suicidal thoughts, plans and intentions. Nudges that entail giving clinicians information about their colleagues’ behaviors may therefore impact suicide screening routines by addressing this uncertainty. In one study, primary care physicians were offered support based on which barriers to suicide screening they reported facing. Those who reported low confidence were given a social norms intervention which included stories of similar primary care physicians who learned to provide screening [27]. Those who reported delaying suicide screening were offered a social comparison intervention to encourage prompt screening [27]. Compared with a control group, these two personalized social interventions (which also included educational interventions and feedback), increased suicide screening at 12 months. Another possible approach would involve inviting clinicians to make a public commitment to screen patients routinely, which could normalize screening among their colleagues. In a study of this approach with a different target behavior (reducing guideline non-concordant antibiotic prescribing), antibiotic prescriptions were reduced after 12 doctors shared a signed public commitment to prescribe antibiotics more carefully [39].

Make the behavior timely.

Managing a disclosed suicide risk is perceived as an immediate, difficult, and time-consuming task, so clinicians may avoid screening for this reason [40]*. A large implementation study investigated suicide screening at discharge from psychiatric units in New York, revealing that information avoidance made clinicians reluctant to conduct screening because they were concerned about the time it would take to manage the large number of people who may require further investigation after reporting ambiguous suicidal thoughts [32]. Assessment and treatment must, therefore, be readily available downstream to ensure screening does not threaten to derail a clinician’s appointment schedule [28, 29]. One practical approach would be to design a default approach to direct referral, including time until the patient can be seen, to refer directly to follow up after screening.

Table 1 shows how these and other exemplar barriers can be addressed by matched strategies.

TABLE 1:

Strategies drawing on insights from behavioral economics

	BARRIERS	Theory	Application		STRATEGIES	Theory	Application
EASY	Under-confidence	Lacking insight into own ability may contribute to avoidance of a challenging task [44].	Clinicians who believe suicide screening is difficult may avoid screening	EASY	Exposure	Roleplaying situations may help establish confidence [28, 29].	Clinicians with more confidence handling risk may find screening easier.
	Paperwork sludge	Multistep processes can frustrate and distract people from practices [31].	Unwieldy paper forms deter clinicians from screening.		Simplicity	When used consistently, mnemonics simplify team communication allowing prompt documentation [45, 46].	Provide a simple recognizable way of documenting suicide screening [47].
					Prompts	Reminders can overcome low salience to change behavior [48].	Clinicians may screen patients if a timely prompt reminds them.
ATTRACTIVE	Uncertainty About Outcomes	Uncertainty about outcomes impacts decision-making in diverse ways depending on outcome magnitude, probability, and framing [49]*.	Clinicians do not get positive feedback when referrals are appropriate.	ATTRACTIVE	Feedback	Informing physicians about appropriateness of past behavior could change future behavior [50].	Clinicians may screen patients if they get positive feedback.
SOCIAL	Uncertainty About Norms	Uncertainty about peers’ preferences and actions perpetuates harmful norms [51, 52].	Being uncertain about colleagues’ suicide screening practice can perpetuate non-screening.	SOCIAL	Social norms	People act in ways that reflect how they believe others act [53].	Clinicians are more likely to screen when told colleagues do.
					Peer comparison	Allowing people to compare and compete increases target desired behavior [54].	Clinicians are more likely to screen when colleagues can see whether they do.
TIMELY	Information Avoidance	People sometimes avoid information, even when it is free and could improve decision making [55].	Clinicians fear knowledge of suicidal thoughts will put them in a challenging situation so avoid that information.	TIMELY	Quick Referral	People are more likely to do something that is quick than something that takes a long time [31].	Referring patients who screen positive should be a quick process.
	Time sludge	Slow processes stop clinicians from completing practices [31].	Screening conversations take too long and frustrate clinicians.		Convenience	Clinicians can incorporate convenient short processes into existing clinical schemata [56].	Normalize a brief sensitive suicide screen, covering thoughts, plans, and intentions.

Next steps for comprehensively integrating implementation science and behavioral economics for suicide prevention efforts

Although there have been many attempts to implement effective suicide screening practices, coupling insights from both IS and BE is an as yet underutilized approach [19]. The EAST framework provides one useful way to integrate insights from these two fields but has its limitations. Insights may fit two or more categories; for example, making a task easy may also make it take less time. The EAST framework may also lead to omission of important approaches. For example, framing information as a gain or a loss can be an important technique in some behavior change strategies, but it does not fit into EAST [20, 41]. The EAST framework has allowed us to highlight key opportunities where BE can inform IS including the importance of simplifying screening documentation, providing feedback to clinicians, enabling comparison between clinicians and organizations, and ensuring referral is a quick process. However, narrow focus on the EAST framework risks overlooking wider barriers and other opportunities for intervention. While this framework provides one helpful way to organize potential links between BE and IS, we encourage readers to also consider other possible links.

Behavioral interventions fail in isolation without adequate consideration of system factors and multi-level approaches. Effective screening is only possible if there is sufficient downstream provision of assessment and evidence-based treatment. Not only is screening unethical without assessment and treatment, but the perception of ineffective or overwhelmed follow-up care undermines attempts to increase screening, and this perception can be hard to dispel even when system-level factors do subsequently improve [37, 42]. Similarly, individual nudges such as feedback or social comparison are insufficient without careful consideration of interventions to address challenges in the outer context and inner context which may not be amenable to behaviorally-informed interventions [43]. Thus, those wishing to design or test BE-informed IS approaches must ensure they are also attending to these outer-level factors.

Conclusion

Pediatric suicide is a significant population-level health concern. To comprehensively address this public health challenge, strategies guided by rigorously tested psychological theories and lessons learned from successful implementation are needed. Effective strategies to increase clinicians’ implementation of suicide screening may benefit from insights from implementation science and behavioral economics.