Behavioral Interventions for Stroke Prevention: The Need for a New Conceptual Model

INTRODUCTION

Social and behavioral factors, also termed social determinants of health, are increasingly established risk factors for incident and recurrent stroke, both ischemic stroke (IS) and hemorrhagic stroke (HS), yet improvement in addressing these factors remains insufficient.¹ There is a lack of clarity around who should share in accountability for reducing risk (e.g., patients, providers, health systems) and what interventions are practical, cost-effective, and scalable.² For the purposes of this review, we will use the definition of an intervention as a “set of actions with a coherent objective to bring about change or produce identifiable outcomes.”³ We first review the published literature to summarize the relevant research on previous behavioral interventions for prevention of stroke and other related conditions, the theoretical frameworks underpinning these behavioral interventions, and then propose a new conceptual model for more effective implementation of social and behavioral interventions for stroke prevention. Successful implementation will require adequately addressing the known inherent barriers to behavioral interventions and the ambiguity of financial responsibility and accountability among the various stakeholders. As new tools such as digital phenotyping, social network analysis, machine learning, and “gamification” have emerged for facilitating, measuring, and improving existing behavioral interventions, a promising new paradigm in behavioral change has emerged.

BEHAVIORAL INTERVENTIONS IN STROKE

Table 1 categorizes the various behavioral interventions by operational level, risk factor targets and tactics that range from the simple (e.g., provider referral for behavior change program) to the complex (e.g., multi-faceted, multi-domain intervention).⁴ Table 2 summarizes the prior interventions specific to stroke prevention covered in 3 major systematic Cochrane reviews. Because the evidence suggests that modifiable risk factors are often not effectively managed following a stroke or TIA, the first review sought to identify interventions for improving control of 6 major modifiable risk factors⁵ of blood pressure, lipids, atrial fibrillation, diabetes, body mass index, and general medication adherence. Twenty-six studies of reasonable quality involving 8021 participants with IS, HS, or non-specified etiology through April 2013 in USA, Canada, Europe, Asia, and Australia were identified for inclusion. The majority were conducted in primary care or community settings and were of 3–12-month duration. Eleven studies involved educational or behavioral interventions for participants and 15 studies involved predominantly organizational interventions. Changes to the organization of healthcare services were associated with meaningful improvements in systolic and diastolic blood pressure, blood pressure target achievement, and body mass index. Examples of organizational interventions include professional role revision (e.g., non-physician staff involvement in prevention clinics), multidisciplinary team collaboration (e.g., primary and secondary care team coordination), and integrated care services (e.g., disease and case management programs following screening, education, treatment, and surveillance protocols). The effects of these interventions on changes in blood lipids, diabetes management, medication adherence, or the occurrence of stroke and other cardiovascular events were equivocal. Overall, changes to healthcare delivery services that addressed patient education or behavior in isolation without changes to the organization of care delivery were not associated with clinically significant changes in modifiable risk factors for stroke. The second review analyzed the effectiveness of multi-dimensional, non-pharmacological interventions used in cardiac rehabilitation including aerobic training, dietary advice or nutritional strategies, verbal or written patient education, and/or lifestyle counseling delivered by health professionals or other personnel and compared these with usual care in preventing secondary vascular events and reducing vascular risk after index stroke or TIA.⁶ Only one study was identified that met inclusion criteria involving 48 participants in a 10-week pilot cardiac rehabilitation program for post-stroke patients, and showed that patients after stroke had significantly greater improvement in a standardized cardiac risk score. The third review focused on physical fitness training for mostly ambulatory stroke patients and included 58 trials with a total of 2797 participants at all stages of care after stroke up through February 2015. It included cardiorespiratory interventions (28 trials, 1408 participants), resistance interventions (13 trials, 432 participants), and mixed training interventions (17 trials, 957 participants).⁷ There was sufficient evidence to recommend incorporating cardiorespiratory and mixed training with walking within post-stroke rehabilitation programs to improve the outcome of speed and tolerance of walking, but no conclusions could be drawn about other outcomes and there was limited generalizability of the observed results.

Table 1.

Classification of behavioral interventions organized by operational level.

Intervention Types by Operational Level	Examples
Organization*	Targeting Structure and Process Revision of professional roles Multidisciplinary teams Integrated care services Knowledge/Information management Quality management Financial incentive
Individual/Group †	Identifying and Quantifying Individual Risk Factors Hypertension (self-reported history of hypertension or blood pressure ≥140/90; greatest relevance to ICH) Current smoking (greatest relevance to IS) Abdominal obesity or waist-to-hip ratio Diet (according to modified alternative healthy eating index score) Regular physical activity Diabetes (self-reported history of diabetes or hemoglobin A1c ≥6.5; greatest relevance to IS) Alcohol intake Psychosocial factors‡ Cardiac causes (e.g., atrial fibrillation/flutter, previous myocardial infarction, rheumatic valve disease, or prosthetic heart valve; greatest relevance to IS) Dyslipidemia (including apolipoproteins; greatest relevance to IS) Targeting Primary Behavioral Change§ Dietary choices Exercise/Physical activity Smoking cessation Reduction in alcohol intake Medication management‖ (e.g., statin, antihypertensive, antiplatelet, and/or antithrombotic agents) Adherence to recommendations Behavior Change Methods and Tactics# Counseling (e.g., motivational interviewing, psychotherapy) Skills training (e.g., self-regulation, self-management, goal planning, mindfulness, cognitive-behavioral therapy) Incentives ∘ Extrinsic (e.g., financial or resource-based incentive) ∘ Intrinsic (e.g., enactive/performance-based attainment or mastery) Contracting Vicarious experience Exhortation Strategic platform (e.g., print, telephone, computer, mobile device, in-person) Modifying physiological states (e.g., pharmacologic intervention to address withdrawal and mood symptoms for coping) Choice environments
Multi-faceted	Multi-level (i.e., organizational and individual/group) Mixed organizational intervention Mixed individual/group intervention (e.g., multi-target, multi-modal, multi-tactic)

^*Taxonomy of organizational changes to improve patient care

^†For patients or, less often, providers; self-directed or dyadic (not organizational)

^‡Defined as composite measures of psychosocial stress which assess characteristics like the combined stress of home and work, life events, and depression

^§Domain through which the intervention is expected to modify behaviors and risk factors to improve desired outcomes

^‖Directed at changing the prescribing behavior of providers

^#Example tactics/mechanisms (not comprehensive list) to influence behavior by affecting sources of expectations

Table 2.

Cochrane systematic clinical trial reviews of behavioral interventions to reduce recurrent stroke or targeting individual risk factor reduction.

Citation	Total Trials Included; Setting; Intervention Types	Study Duration	Outcomes of Interest	Participants; Significant results at final follow-up
Lager, et al.5	26; primary care or community; 11 educational or other behavioral, 15 organizational	3–12 months	Risk factor control in secondary stroke prevention: systolic and diastolic blood pressure, blood lipids, atrial fibrillation, diabetes management, body mass index, medication adherence	n=8201 Organizational interventions compatible with meaningful improvements in blood pressure and body mass index changes in blood lipids, diabetes management, medication adherence, or the occurrence of stroke and other cardiovascular events were imprecise and equivocally supportive of benefit and harm Education or other behavioral interventions without health organization change were generally not associated with clinically significant changes in modifiable risk factors for stroke 4 of 21 ongoing trials at the time of review subsequently concluded and none had significant impact on clinical outcomes
Mackay-Lyons, et al.6	1; cardiac rehabilitation program for stroke patients; multi-modal	10 weeks	Vascular risk factor modifications	n=48 Small reduction in standardized cardiac risk score, but limited conclusions about other clinical outcomes 2 of 5 ongoing trials at the time of review subsequently concluded (improved cardiac risk score, n=24; improved exercise capacity only, n=65)
Saunders, et al.7	58; rehabilitation program for stroke patients; mixed physical activity interventions	4–24 weeks	Physical activity modification	n=2797 Study quality, variability, and lack of date prevented conclusions about clinical outcomes

WHY DO BEHAVIORAL INTERVENTIONS MATTER?

Neurology is transforming from a predominantly diagnostic field to one with potent therapeutics. Despite major advances in acute stroke treatment,⁸ new and recurrent ischemic and hemorrhagic stroke represent a major public health burden in the US and worldwide.^{9, 10} Prevention is fundamental to decreasing disease burden and its sequelae, and much work has been devoted to producing and disseminating evidence-based guidelines for primary and secondary stroke prevention.^{11, 12} While acute stroke interventions have been increasingly implemented, they impact only a small fraction of stroke patients, and progress in improving the cardiovascular and cerebrovascular health of the population lags far behind.⁹ The mismatch of poor health outcomes and high spending on health care services within the United States exists across many conditions, but stroke is a critical target given the influence of potentially modifiable risk factors and its known contributions to dementia.¹³ While regional variation exists in the influence of most stroke risk factors and contributes to worldwide heterogeneity in stroke frequency and case-mix, 10 potentially modifiable risk factors are collectively associated with ~90% of the population-attributable risk of stroke in each major region of the world across ethnicity, sex, and age.¹ Across 188 countries, much of the preventable disease burden due to lifestyle stems from active smoking, physical inactivity, and excess intake of salt, sugar, and alcohol,¹⁴ with hypertension, abdominal obesity, diabetes, heart disease, and dyslipidemia contributing powerfully. Health care providers have generally regarded improving health behaviors and reducing psychosocial stress as beyond their sphere of influence or training, yet recent progress in behavioral sciences, health technology, and healthcare spending priorities have set the stage for stroke providers to meaningfully join the fight to reduce the global burden of cerebrovascular disease.

Of the potentially modifiable stroke risk factors, efforts to increase exercise are considered a “best buy” in public health¹⁵ and therefore have been a focus of interventions given the potential virtuous cycle of reduced risk for obesity, diabetes, major vascular events, and vascular dementia. However, to date they have largely been unsuccessful.¹⁶ Common interventions to increase physical activity include (1) community-based media campaigns, social support interventions, or physical activity classes, (2) school-based education interventions, (3) community-wide policies and programs, and (4) activity monitoring with technology and feedback.^{3, 17} Despite more countries implementing physical activity surveillance systems and national physical activity campaigns, global physical activity levels have not increased since 2012.¹⁶ Evidence of the benefits of behavioral interventions is growing, but scaling effective interventions to the population level has been challenging.^{3, 18} We explore the potential reasons for these challenges and propose an alternative framework for proceeding.

PROPOSED FRAMEWORK

While health systems have focused heavily on the supply side of healthcare delivery, growing cost pressures will likely limit future investment. Innovative and cost-efficient solutions are sorely needed. Interventions based on social cognitive theory are more patient-centered, focus on the demand side of health care, and offer principles and predictors that can be used to inform and motivate people to adopt healthy habits.¹⁹ Social cognitive theory suggests that behavior is determined by the combination of expectations and reinforcements. Specifically, expectations of self-efficacy influence the regulation of human motivation and behavior by operating together with goals, outcome expectations, and perceived sociostructural facilitators and impediments (Figure 1).¹⁹ Laibson examined lessons learned from efforts to promote favorable consumer retirement savings behavior and adapted social cognitive theory to apply them to medicine, identifying the three core elements of motivation, barriers, and a “universal bulldozer.”²⁰ He highlights the intrinsic complexity of behavior change and the many positive and negative reinforcing internal pathways that interact with endophenotypes (i.e., inherent variations of behavioral and biological characteristics) along the path to the desired outcome, not unlike the complex excitatory and inhibitory synaptic connections that regulate basal ganglia activity (Figure 2). Motivation is at minimum the combined effect of expectations and reinforcements. Expectations include environmental cues, efficacy expectations, and outcome expectations. Positive and negative reinforcements can be subdivided as either extrinsic or intrinsic. Extrinsic (or “hedonic”) reinforcements are typically tangible, derived from external forces, and short-lived.²¹ Conversely, intrinsic (or “eudaemonic”) reinforcements are closely tied to psychological well-being and the perception of autonomy, mastery and purpose. They are less tangible but have greater impact on sustaining behavior over the long term.²¹ Barriers are not only impediments or costs to behavior, but also pre-existing and dynamic sources of negative expectations and reinforcement. Barriers are highly context dependent and can stem from biological factors (e.g., genetics, epigenetics, reactive health states, and physiology), other environmental influences, and the inextricable interrelationships between both. Environmental influences can be predominantly external (e.g., sociocultural, economic, and geographic), include the presence of behavioral triggers, and stem from the design and functioning of organizational systems. Environmental influences can also be internal states (e.g., physical and mental functioning, affective states of apathy, fatigue, anxiety, and impaired self-regulatory executive control) or tied to deeply rooted aspects such as the degree of acquired knowledge and health literacy tied to crystallized intelligence and educational attainment. Lastly, an effective intervention will be one that is a “universal bulldozer” in that it that has the capability to overcome all relevant barriers in the cascade with precision and is cost-effective, scalable, and sustainable. This universal bulldozer must have the ability to address multiple barriers simultaneously through the combination of tailored, mixed, and multi-faceted tactics likely including structures for funding, incentives, engagement, and support. As each intermediate desired outcome is achieved, there is the potential for a virtuous cycle whereby the accomplishment increases motivation and serves as a source of increased expectation and positive reinforcement.

Figure 1.

Social cognitive theory conceptual pathways for how efficacy expectations can influence behavior directly and indirectly through outcome expectations, goals, and perceived sociostructural facilitators and barriers.

Figure 2.

Conceptual model for behavioral strategies to modify risk factors in primary and secondary prevention.

WHAT IS NEEDED MOST?

The two missing mechanisms needed most to make progress using the proposed framework are (1) the capability to identify and address all context-specific barriers for each individual and (2) a design strategy that simplifies the user experience of the at-risk individual by shifting the complexity away from the individual and onto the social and technical “software” and “hardware” that power the intervention.

The comprehensive identification of context-specific barriers is critical to maximizing an intervention’s precision. This design-informed approach will require identifying barriers to motivation that are traditionally unexplored by most healthcare providers, such as the level and type of interactive guidance that would be best suited to a person’s self-management capabilities and their motivational preparedness to achieve the desired behavior change.¹⁹ This can be thought of as a form of precision medicine that targets epigenetic factors such as psychological and social determinants of health rather than genes, but with no less rigor. Furthermore, an intervention addressing only some potential barriers would be considered a “partial bulldozer,” such as a smoking cessation intervention that educates smokers about the harmful health effects of cigarette smoking. A “universal bulldozer,” meanwhile, addresses all relevant barriers, such as a smoking cessation intervention that teaches health literacy, treats nicotine cravings, enrolls in a support group, monitors purchases of nicotine products, provides social and financial team-based incentives for tangible, challenging, and attainable smoking cessation goals, and more. When only some of the context-specific barriers are addressed, a lower complexity partial bulldozer will be easier to design and implement but will demonstrate lower efficacy (Figure 3). Therefore, despite the challenges for implementation, an intervention that is both precise and addresses all major barriers identified is the one which will be most consistently efficacious and sustainable. Sophisticated, computer-assisted interventions based on the principles of self-regulatory behavior in social cognitive theory already exist in other industries (e.g., retirement savings plans),²² and have recently been introduced at a smaller scale in chronic disease self-management for conditions like stroke.^{23, 24} These models have the potential to increase healthcare value (i.e., lowering cost while improving quality) by combining a clinically precise and individualized approach to overcoming barriers with population-based public health interventions (Figure 4).²⁵

Figure 3.

Combinations of precision and extent of barriers addressed incorporated into the design of behavior change interventions. Whereas intervention A is easiest to implement and has low efficacy due to low precision and partially addressing barriers to behavior change, intervention B is challenging to implement and has low efficacy due to low individual precision while attempting to address all barriers, and intervention C is challenging to implement and has low efficacy due to high individual precision yet only partially addressing barriers, intervention D has the highest efficacy but is also the most challenging to implement from the combination of high precision and addressing all barriers relevant to the individual.

Figure 4.

Example of a computer-assisted self-regulatory system for behavior change self-management.

KEY BARRIERS AND POTENTIAL SOLUTIONS

Funding and Accountability

In relation to the worldwide economic burden of stroke, the global median adjusted population attributable fraction of stroke associated with physical inactivity is 4.5%.²⁶ The impact of physical activity on stroke bears an economic burden of at least $6 billion out of the $53.8 billion related directly to healthcare costs.²⁷ The direct and indirect economic costs are paid mostly by high income countries, but the lower and middle income countries pay for a larger share of the disease burden from physical inactivity in the form of reduced health and increased mortality. The investment to reduce the burden of physical inactivity would seem a rational investment, yet there is a general lack of capacity and funding globally suggesting it has not been a top public health funding priority.¹⁸ The economic burden is paid mostly by the public, followed by private third parties and individual households.²⁷ The quality of a nation’s health is simultaneously a social and personal matter, yet it is not consistently embraced as a public or government responsibility in the way that the health of the economy is uniformly acknowledged to be (e.g., tracking of a nation’s economic indicators such as unemployment rate and gross domestic product).

As a result of direct health care expenditures and indirect productivity losses, the staggering $67.5 billion economic cost of physical inactivity worldwide is sufficient cause for alarm.²⁷ There is a glaring lack of established pathways to redirect financing toward outcomes of health, even with more public and political motivation.²⁸ To be clear, though, it is not the lack of sound policy prescriptions so much as a deficiency in collective will and efficacy, from the governmental level down to the individual, that prevents government agencies from realizing existing policies. Adopting national policies or action plans are not equivalent to stimulating or implementing change.¹⁶ Meaningful action also means addressing one of the more daunting challenges to overcome: limitations in political commitment and resources. These limitations can only be overcome through multiple long-term agreements, partnerships, and collaborations driven by advocacy.²⁹ Decline in smoking rates, for example, are a result of a successful social approaches. Despite being one of the most individually preventable causes of death, the collective movement to create smoke-free environments to decrease cigarette smoking was not accomplished solely by government agencies with the explicit responsibility to protect national health. Rather, this public health success was accomplished through collective social efforts.

Changing the practices of social systems that impair health and changing the habits of individuals are both required. To be successful, behavioral interventions for health must be funded and structured as an integral component of a societal commitment to prioritize the health, survival, and quality of life of its citizens at equal priority to their economic well-being.

Barriers to Implementation

Meaningful real-world, scalable implementation of behavioral interventions that have proven effective in highly controlled environments such as randomized clinical trials requires addressing both knowledge-action and process-outcome gaps. For example, reduced exposure to risk factors at the population level has been shown to be beneficial for cardiovascular and overall health, as with reducing salt and sugar content in the manufacturing of packaged foods.³⁰ There is great value in accomplishing even small subgoals which contribute to meaningful intermediate outcomes: for example, small physical changes in urban environments that introduce just ten minutes of moderate physical activity daily to two-thirds of inactive persons worldwide in accordance with international health guidelines can potentially prevent 3.6 million deaths annually.³¹

Barriers to Participation

Scalability can be achieved vertically (through organizational systems that support health behavior change) and horizontally (through dissemination), but always requires addressing barriers to participation. Investigators, policymakers, and practitioners need to identify the appropriate questions for each set of action-oriented stages required for successful scaling: effectiveness (what is the impact of the intervention), reach (is the target population being reached), adoption (is the organizational support readily available), implementation (is the intervention delivered properly), and maintenance (is the intervention sustained and scaled up).³² Large-scale multinational behavioral interventions require the facilitated development of global consensus targets and indicators by country (see www.ICHOM.org for example standard measurement sets), with national health organizations focusing on within-country multilevel and multisector interventions that include participation of both health and non-health sectors.

Research Gaps

A systematic review of randomized trials that used multimodal behavioral interventions for secondary stroke prevention suggests that while non-pharmacologic interventions are effective in reducing anxiety and the odds of cardiac events, they are not effective at reducing the odds of death, recurrent TIA, or stroke.³³ Randomized trials using a self-management intervention for secondary stroke prevention improved an intermediate outcome of medication adherence.³⁴ While randomized clinical trials are the gold standard for comparing outcomes between two alternative treatments, because most behavioral intervention trial designs have a high degree of heterogeneity^{5, 6} they are not well suited to pooled or meta-analyses and thus definitive conclusions even from negative trials are lacking.

Progress will require a greater number of behavioral intervention studies that are rigorously designed according to the relevant factors discussed above and for which pragmatic scalability and dissemination have been considered, and these studies must be prioritized by researchers, funding agencies, and scientific journals.³ Such well-designed studies should include (1) a defined theoretical basis, (2) standard intervention parameters (e.g., dosing, duration, frequency, timing, platform, and modality), (3) standard intervention taxonomy, (4) ease of replication, (5) standard measures of generalizability, (6) standardized outcome measurement,³⁵ and (7) coordinated scientific oversight with larger samples and longer term outcome assessments. Progress is especially needed in low-income and middle-income countries³ whose citizens are disproportionately impacted by non-communicable diseases due in part to unhealthy behaviors.

Tools and Technology

Understanding and measuring behavior

Better tools for active and passive surveillance and evaluation of patients of all age groups and settings are needed for more objective measurement of desired behaviors and a greater understanding of intervention efficacy.¹⁶ With the ubiquitous availability of smartphones and wearable devices, new opportunities exist to measure physiology and behavior in situ (known as digital phenotyping), rather than through periodic cross-sectional self-report assessments which are prone to recall bias and under-reporting.³⁶ These devices can passively monitor blood pressure, step counts, sleep patterns, changes in vocal characteristics, and other physiologic parameters as well as behaviors, such as distance traveled, time spent at home or in specific parts of the home, and the frequency, duration, and directionality of social contact. However, evidence suggests that technology strategies, when applied in isolation, cannot function as a universal bulldozer. They are unlikely to change behavior in high risk patients, and they can raise deep concerns about patient privacy, data storage capacity, and information overload for clinicians.³⁷ Technology strategies are, at their best, facilitators of health behavior change and require both patient and provider engagement.

Influencing behavior

Engaging media and social relationships

With the greater availability of audiovisual content on demand, the thoughtful use of compelling and engaging media becomes a potential source of widespread health promotion messaging that can be targeted to the individual viewer. However, there is limited value to motivating change in the absence of resources and community supports to implement those changes, and without proper oversight and coordination this messaging can be inaccurate, misleading, misinterpreted, or damaging to existing efforts.³⁸ Media is a source for providing expectations and can promote reinforcement to influence behavior by immersing individuals in a network with socially contagious behavior.¹⁹ With the widespread embrace of social media applications across various age groups and growing within aging populations,³⁹ these paths of influence become particularly relevant. Social applications such as Facebook, LinkedIn, Twitter, Instagram, Pinterest, and Snapchat are all channels that can modify expectations through increased visibility of role modeling, social positive or negative reinforcements, and compelling text, links, images, and videos through posts and hashtags.⁴⁰ Highly social and individualized interactive algorithm-based systems are readily scalable; however, as previously stated, low-cost tools will be ineffective without addressing the other individualized barriers.³⁷

Behavioral economics and gamification

Behavioral economics is an area of study that incorporates into economic models insights from psychology such as loss aversion, present bias, social comparisons, preference for variety and inattention.⁴¹ Behavioral economists refine traditional principles of modern economics by studying how people make successful and unsuccessful attempts to pick their best feasible option. These insights can be leveraged as tactics in the design of “choice environments,” and as elements in complex computerized systems that support tailored interventions.

“Gamification” can be defined as the thoughtful integration of game dynamics into a system to drive participation, which often involves leveraging many behavioral economic insights.⁴² One instance of applying gamification would be a workplace physical activity intervention designed as a contest where co-workers are given activity monitors to wear, broken up into teams, and the team with the most steps in a month wins a social or financial prize. There are several recent examples of gamification to change behavior in health care,⁴³ but none has had as profound an impact as the mobile device-based Pokémon Go.⁴⁴ The Pokémon Go user experience design eliminates almost all barriers to its use (e.g., simple enrollment, simple and visually appealing interface, captivating narrative, geographically-tailored engagements) leading to powerful leveraging of expectations and reinforcements to motivate millions of users of all ages to walk through the real world in their local community participating either as individuals or in larger social groups in order to “catch” augmented reality creatures called Pokémon. There is even a small clinical trial testing the association of Pokémon Go and increased physical activity underway.⁴⁵ A systematic review of randomized trials of diabetes self-management using mobile eHealth technology, videogames, and virtual environments found mostly small sample sizes and short duration endpoints.⁴³

Timeframes for Measuring Value

A precise “universal bulldozer” to improve population-based secondary stroke prevention will require aligning systems and processes around value-based care delivery and working across multiple industry sectors to provide meaningful and enduring reinforcements. The Coordinated Approach To Child Health program required ~20 years to translate from evidence to practice,³ and broad adoption of retirement savings plans required over 40 years of continuous iterations and development.⁴⁶ Measuring health value concretely and transparently can enhance goal visibility and thus assist with reinforcing collective self-efficacy and outcome expectations through behavioral interventions.⁴⁷ Generating international consensus-driven and condition-specific measurable outcomes that are meaningful to patients is a critical first step in behavior change, and the implementation of standardized global outcomes for stroke such as the Stroke Standard Set³⁵ are a viable start.

CONCLUSIONS

If improving and preserving brain health is the core business of neurology, then we must be bolder and more creative about how this is accomplished. Despite the challenges, creation of a cost-effective and scalable behavioral intervention with a sophisticated engine that is precisely designed to deeply engage the user through a simple yet compelling experience is a worthy investment that can reap great dividends. The blending of social cognitive theory with emerging powerful digital tools in the medical and social spheres offers an unprecedented opportunity for effective interventions. Novel partnerships between social scientists, physician investigators, media content creators, user experience designers, public health leaders, and policymakers are needed to achieve the desperately needed progress.