Overcoming Psychological Barriers to Treatment Adherence: The Case of Antihypertensive Medication

Abstract

The goal of this article is to examine the modifiable psychological causes of adherence to antihypertensive medications and to propose a potential approach for studying behavioral interventions to improve medication adherence. This article briefly discusses the importance of antihypertensive adherence and the most prevalent patient-level factors associated with nonadherence. The authors examine psychological barriers to adherence, which are often the causes that are most easily modifiable with person-level interventions. The article then reviews the work of the Science of Behavior Change Program and the Human Behavior Change Project and discusses how this work can help to develop a standardized, mechanistic approach to intervening on patient-level causes of nonadherence. Specifically, the article discusses how researchers and practitioners can develop behavior change technique interventions addressing underlying psychological mechanisms of behavior change for antihypertensive medication adherence.

INTRODUCTION

The goal of this article is to explore how we can hope to address cardiovascular disease (CVD) by focusing on health behaviors. Using the example of adherence to antihypertensive medication, the authors propose to guide the reader through our approach using the following roadmap in Figure 1.

Figure 1:

Theoretical Roadmap for our proposed approach.

Antihypertensive adherence

Human behavior is one of the most important determinants of health, accounting for nearly half of all deaths in the United States.^{1, 2} Approximately 25% of these preventable deaths are attributed to CVD, with hypertension being one of its most relevant risk factors.^3-5 In this setting, pharmacological therapy plays a critical role in managing hypertension, in turn, decreasing the risk of CVD events. Individuals who are adherent to antihypertensive medications have a 38% reduction in hazard of incident CVD,⁶ while nonadherence is associated with a 48% increased hazard of all-cause mortality.⁷ However, nonadherence to antihypertensive medications is common, with 35% to 72% of individuals with hypertension not taking their medication as prescribed.^{6, 8, 9} This poor adherence to antihypertensive medications is associated with increased rates of CVD, hospitalization, healthcare utilization, and healthcare costs.^10-12 Improving adherence to medications should be a primary target for physicians, patients, policy makers, and payors seeking to improve health.¹³ To achieve this goal, researchers have investigated various interventions aimed at improving patient adherence to medications.^14-16 Despite the importance of increasing adherence to antihypertensive medications and the increased research focus on nonadherence, low levels of medication adherence are still a concern.¹³

Understanding the barriers to antihypertensive adherence

There are multiple ways for patients to be nonadherent to their prescribed medication. These include not initiating medication at all (e.g., never filling the first prescription), taking medication incorrectly (e.g., missing doses, refilling medication late), and non-persistence (e.g., stopping medication before physician recommendations).¹⁷

There is an extensive literature investigating the causes of antihypertensive nonadherence. The causes are complex, but the most commonly identified factors are patient demographics (e.g., age, sex), treatment-related factors (e.g., class of antihypertensive medication, number of antihypertensives prescribed), medical factors (e.g., hypertension severity, medical comorbidity, psychological comorbidity, medication side effects), health-system based factors (e.g., patient-provider relationship, communication about diagnoses/medication), financial factors (e.g., socioeconomic status, lack of medical insurance, medication cost), and psychological factors (e.g., knowledge about hypertension diseases symptoms and consequences, Self-Efficacy, health beliefs, social support).^18-25

An important distinction among these causes for nonadherence are whether they are modifiable or unmodifiable. For example, certain patient demographics, such as age, can be strong predictors of health behaviors and health outcomes but are often unmodifiable ^{26, 27} Other factors, such as socioeconomic status or lack of medical insurance, may also require substantial life changes or systemic changes to healthcare policy to change behaviors or outcomes. On the other hand, modifiable psychological factors like hypertension knowledge, Self-Efficacy, health beliefs, and social support are often some of the most important causes of nonadherence.²⁴ In fact, some studies suggest these psychological factors may be more strongly associated with adherence behaviors than demographic factors.²⁸

Identification of modifiable factors has led to increased focus on patient-centered interventions to improve adherence.²⁵ The focus of this approach is to work with patients to identify the causes of their nonadherence and then develop a person-centered approach to intervene on these causes. By focusing on the underlying psychological causes of nonadherence, it is possible to deliver interventions directly to patients and make targeted, substantial changes in adherence behavior.

Within psychological factors related to nonadherence, it is also possible to divide modifiable and unmodifiable reasons for nonadherence into intentional and unintentional reasons.^{29, 30} Being nonadherent for intentional reasons is defined as making an active, reasoned decision to take or not take medication.³¹ Intentional reasons which have been commonly endorsed include not believing medications were needed, knowledge about the consequences of hypertension, and fear of complications/side-effects.^{22, 31, 32} Among the most commonly identified unintentional reasons for nonadherence are having difficulty taking medication, forgetting to take medication, medication cost, complexity of treatment, and medical comorbidity.^{22, 31} Though these are the most common, there are many intentional and unintentional reasons for nonadherence which can be modified.

Each of these causes (modifiable and unmodifiable; intentional and unintentional) will require appropriately tailored and different intervention approaches to change behavior (e.g., a reminder intervention for unintentional forgetting versus an information intervention to improve intentional health beliefs). To effectively select the appropriate intervention for a particular cause of nonadherence, it is thus essential to understand how and why an intervention works (i.e., the mechanism).

The literature summarized above highlights that nonadherence to antihypertensive medications has a variety of root causes and distinguishing factors, with modifiable psychological factors being significant contributors to behavior. This variety makes altering adherence behavior complex, as each factor may respond to an intervention differently. Further, patients may present with multiple reasons for nonadherence both individually and in combination. To improve adherence, it is necessary to understand why nonadherence is occurring and address those reasons directly with specific, granular, targeted intervention approaches. This article proposes this understanding can be achieved in the following manner: 1) examine the mechanisms underlying adherence; 2) develop targeted interventions for nonadherence using information from behavioral intervention ontologies and 3) systematically study how targeted interventions influence mechanisms for adherence using randomized controlled trials (RCTs).

Mechanisms of Behavior Change and Science of Behavior Change

Mechanisms of behavior change (MoBCs) are the mechanisms by which behavioral interventions can influence behavior.³³ The study of MoBCs is based on identifying why behavior exists and how best to change it.³³ The consideration of MoBCs also identifies the underlying psychological processes which help to increase or decrease adherence. ^34-37 In fact, one could think of MoBCs more specifically as modifiable reasons for behavior. For example, Self-Efficacy, defined as beliefs about one’s ability to successfully perform a behavior, has been identified as a key MoBC in behavior change theories³⁸ and has been found to be integral to medication adherence.^39-41 Low medication adherence Self-Efficacy, defined as a low level of belief in an individual’s ability to consistently take medication as prescribed, is a potential modifiable reason for nonadherence. Once the mechanisms and the causes of adherence have been identified, it is possible to act directly on these MoBCs, thereby changing health behavior. This theoretical basis is integral in precisely prescribing the appropriate intervention to influence behavior, including medication adherence. Despite the potential benefits of studying mechanisms and MoBCs specifically, only 3% of intervention studies funded by the National Institutes of Health examine intervention mechanisms.⁴²

Recognizing the importance of identifying MoBCs and the lack of research into mechanisms of interventions, the NIH established the Science of Behavior Change (SOBC) program across 17 NIH centers and programs in 2009.³³ The SOBC approach aims to address this basic scientific problem, and advance behavior change research by implementing a mechanisms-focused, experimental medicine approach to behavior change research.³³ Following this approach, researchers: 1) identify the putative MoBC; 2) engage this MoBC; and then 3) assess how MoBC engagement produces behavior change. This approach provides a systematic, rigorous, and reproducible method for identifying the MoBCs that mediate successful behavior change. Further, by addressing how an intervention works, the SOBC approach is able to link the work of basic and applied behavioral science.⁴³ In addition, the SOBC approach is designed to increase transparency, rigor, and reproducibility through the use of common terminology, methods, and measures.⁴³ The goal is to identify common MoBCs and uniform methods of measurement for these mechanisms.³³

Behavior Change Techniques

Complementing the work of the SOBC is the research led by Dr. Susan Michie and the Human Behavior Change Project (HBCP)⁴⁴ This group aims to link interventions with MoBCs, but also works to decompose behavioral interventions into components. These components, called behavior change techniques (BCTs), represent the “active intervention ingredients” of a behavioral intervention, similar to the active components of a prescription medication.⁴⁵ Through use of literature review and expert consensus, Dr. Michie and colleagues have identified 74 BCTs that represent the smallest active components of an intervention.^{45, 46} The benefit of breaking down behavioral interventions into smaller components is that it allows for standardization and supports the accumulation of knowledge regarding the success of interventions across multiple research studies.³⁸ Further, this research was expanded to identify which potential MoBCs were engaged by each BCT, utilizing a synthesis of 277 research articles.⁴⁶

Using the combination of theory- and practice-based research, RCTs already have operationalized effective multi-component BCTs for influencing health behavior.^{47, 48} Multi-component BCT interventions have demonstrated modest efficacy for increasing positive health behaviors in areas such as smoking,⁴⁹ physical activity,^{48, 50, 51} healthy eating,^{48, 50, 51} and medication adherence,^{47, 52-55} including antihypertensive adherence.^{56, 57}

Combining MoBCs and BCTs into an intervention approach

The authors propose to address the complicated and varied causes of nonadherence to antihypertensive medications, particularly those that are modifiable psychological barriers to adherence, using the standardized approach provided by the SOBC and the HBCP. Prior research has already illustrated the importance of identifying the root causes of nonadherence at the individual patient level and developing interventions to address these causes.¹³ What is missing from this conceptualization is standardization and linkage to behavioral theory. By utilizing the extensive work conducted by the SOBC and the HBCP, it is possible to utilize validated BCT interventions paired with standardized MoBCs to understand which intervention components are most effective for improving patient adherence.

Combining these approaches could lead to the following conceptual model (Figure 2). Researchers and clinicians have the potential to use BCTs, well-defined active ingredients for behavioral interventions, to address individual MoBCs, defined as potential modifiable causes for nonadherence, to address nonadherence to antihypertensive medications along multiple paths.

Figure 2: Conceptual Model with multiple mechanisms.

From Carey RN, Connell LE, Johnston M, et al. Behavior Change Techniques and Their Mechanisms of Action: A Synthesis of Links Described in Published Intervention Literature. Ann Behav Med. Jul 17 2019;53(8):693-707. doi:10.1093/abm/kay078.

To better explore how this might work, a more concrete example using the available science is provided below.

An Example Intervention Model

Theoretical Approach

In order to explore this concept, a literature review has been conducted to generate a hypothetical model of an intervention utilizing BCTs and putative MoBCs that would improve adherence to antihypertensive medications. To select BCTs to include, the full range of BCTs that have been found to change positive health behaviors, in general,^{46, 48, 50, 51, 58-60} and medication adherence, in particular, has been considered.^{46, 52, 61-64} From this list, four BCTs that are most likely to improve antihypertensive medication adherence have been identified. These four BCTs are some of the most common^{56, 61, 63} and effective intervention components for improving antihypertensive medication adherence.^{12, 65-67}

The BCT of Problem Solving encourages trial participants to identify barriers to their medication adherence and develop solutions, helping them to manage their medications. For example, individuals might be encouraged to think of ways to make taking medication easier, such as placing their medication next to their toothbrush so they remember to take medication after brushing their teeth. The BCTs of Feedback on Behavior and Prompts/Cues help with nonadherence by providing feedback on whether participants have taken their medications and reminding participants to take their medication as prescribed. For example, Feedback on Behavior may involve saying “You took your medication as prescribed five out of seven days over the last week”. A Prompts/Cues intervention might be a text message stating “Remember to take your medication”. Information about Health Consequences provides educational messages reminding participants of the importance of adhering to their antihypertensive medication. For example, sending a message stating “Taking your antihypertensive medication consistently can help reduce the likelihood of a stroke”. Table 1 provides an overview of the BCTs, including definitions and example messages for each BCT.

Table 1.

Key BCTs for an example intervention to improve adherence to antihypertensive medication

BCT Name	BCT Description	BCT Example Message
1.2 Problem Solving	Analyze, or prompt the person to analyze, factors influencing behavior and generate or select strategies that include overcoming barriers and/or increasing facilitators.	“Think of some reasons why you might not take your medications tomorrow. What could you do to make it easier for you to take your medication tomorrow?”
2.2 Feedback on Behavior	Monitor and provide information or evaluative feedback on performance of the behavior (e.g., form, frequency, duration, intensity).	“You took your medication as prescribed yesterday.”
5.1 Information About Health Consequences	Provide information (e.g., written, verbal, visual) about health consequences of performing the behavior.	“You can have high blood pressure and not feel sick. It’s important to keep taking your medication even if you don’t feel like you need to.”
7.1 Prompts/Cues	Introduce or define environmental or social stimulus to prompt or cue the behavior. The prompt or cue would normally occur at the time or the place of performance.	“It’s time to take your medication.” (delivered 15 minutes prior to the time the patient usually takes medication)

Note: BCT descriptions come from the HBCP “Theory and Technique Tool”⁶⁹

The BCTs described each help with important patient-related causes of nonadherence, such as patient expectations, health literacy, forgetfulness, and lack of feedback.⁶⁸ Further, these BCTs address both intentional (e.g., patient expectations) and unintentional causes of antihypertensive adherence (e.g., forgetfulness).

To select the key MoBCs required a review of all 26 MoBCs from the Michie group.⁴⁶ The three MoBCs selected are: Hypertension Knowledge; Self-Efficacy; and Behavioral Automaticity:

Hypertension Knowledge is defined as a participant’s awareness of the symptoms of hypertension, the potential consequences of hypertension, and how hypertension can be treated. Low levels of Hypertension Knowledge about medications previously have been associated with nonadherence to antihypertensive medications,^{65, 67, 70, 71} and likely will be influenced by the BCT Information about Health Consequences.⁴⁶

As stated previously, Self-Efficacy (also known as an individual’s beliefs about their capabilities to perform a certain behavior) has been identified as a key MoBC for many identified behavior change theories,³⁸ and has been found to be integral to medication adherence,^39-41 especially to antihypertensive medications.⁷²

Finally, the MoBC Behavioral Automaticity (also known as Cueing) also has been found to be associated with adherence to antihypertensives,⁷³ and has been shown to engage the Prompts/Cues BCT.⁴⁶ Behavioral Automaticity can be defined as the level of performing a behavior automatically without conscious thought. Behavioral Automaticity is also more commonly known as “habit” i.e., a behavior performed automatically. Table 2 provides an overview of how each of these MoBCs can be defined and classified.

Table 2.

Key MoBCs for improving adherence to antihypertensive medication

MoBC Name	MoBC Definition	MoBC Domain from HBCP Ontology	MoBC Domain Definition
Hypertension Knowledge	Knowledge about hypertension, including defining hypertension, what blood pressure is, and what the consequences of hypertension are.74	Knowledge	An awareness of the existence of something
Self-Efficacy	People’s beliefs about their capabilities to engage in a behavior or achieve a goal.75	Beliefs About Capabilities	Beliefs about one’s ability to successfully carry out a behavior.
Behavioral Automaticity	This is a measure of whether a behavior is conducted automatically, without cognitive attention. Also known as a habit.76, 77	Behavioral Cueing	Processes by which behavior is automatically triggered from either the external environment, the performance of another behavior, or from ideas coming to mind.

Note: MoBC descriptions come from the HBCP “Theory and Technique Tool”⁶⁹

These key MoBCs selected, both because of their association with antihypertensive adherence, and because of their association with the BCTs used in this study, will help identify the mediating mechanisms of the intervention. Figure 3 shows the established associations between our four identified BCTs and these three putative MoBCs from a synthesis study conducted by Dr. Michie and her colleagues;⁴⁶ the darker the red, the stronger the association based on the lower p-value. By comparison, associations with no color (e.g., white) are not strong enough to support being included in an intervention. Thus, associations with darker colors were selected. However, it should be noted that for the purposes of their review, Carey and colleagues did not utilize the traditional significance cutoff (p<0.05) and highlighted p-values less than 0.30 as being worthy of further attention and future research.^{46, 78}

Figure 3: P-values for the Association of BCTs and MoBCs from Carey 2019.

From Carey RN, Connell LE, Johnston M, et al. Behavior Change Techniques and Their Mechanisms of Action: A Synthesis of Links Described in Published Intervention Literature. Ann Behav Med. Jul 17 2019;53(8):693-707. doi:10.1093/abm/kay078.

It is also necessary to make sure these MoBC measures are assessed using standardized, validated questionnaires previously used for health behavior and adherence research. For example, it is possible to measure the MoBC Self-Efficacy using the Medication Adherence Self-Efficacy Scale Revised, a 13-item measure of medication adherence Self-Efficacy, previously used in a variety of populations to measure Self-Efficacy for adherence to antihypertensive medications.^79-82 Similarly, it is possible to assess Hypertension Knowledge using a 10-item hypertension knowledge scale,⁷⁴ previously used in several trials to increase medication adherence and lower blood pressure.⁸³ Finally, it is possible to assess Behavioral Automaticity using the Self-Report Behavioral Automaticity Index,⁸⁴ a 4-item measure of the automaticity of health behaviors, previously used for trials examining medication adherence.^{73, 85}

Synthesizing these components together allows us to address medication adherence using the following theoretical model (Figure 4). The paths identified in Figure 4 are those identified as being worth attention in previous syntheses of BCTs and MoBCs.⁴⁶ Further, each of these MoBCs have supported associations with the outcome and well-researched assessment methods. The model shows how Information about Health Consequences and Feedback on Behavior can contribute to adherence through the MoBC of Hypertension Knowledge. Moreover, it illustrates how Feedback on Behavior and Problem Solving can influence medication adherence via the MoBC of Self-Efficacy. Finally, the model displays how Prompts/Cues influence adherence via Behavioral Automaticity (i.e., habits). Therefore, the model summarizes how four BCTs delivered individually or in combination have the potential to influence adherence via three key MoBCs.

Figure 4: Conceptual Model.

Note: The BCTs with the strongest association with particular MoBCs from Figure 3 (depicted using dark red shading) were selected as BCTs for this hypothetical intervention.

It should be noted that Figure 4 displays only one of many possible conceptual models that can be used to improve health behaviors such as antihypertensive adherence, as theoretical and practical concerns can influence model specification. For example, Prompts/Cues is the sole BCT identified to influence Behavioral Automaticity, but other BCTs are theoretically plausible.^46,69 Moreover, the BCTs specified in this model were also BCTs that our research team has had experience delivering.^86-90 Therefore, our research group would feel comfortable utilizing these particular BCTs in an intervention.

An additional consideration is ensuring that participants are not overwhelmed by the intervention by delivering too many BCTs. For example, sending twelve text messages might potentially boost the overall intervention effect but might reduce the effectiveness of each BCT. Similarly, selecting too many MoBCs might be difficult to analyze. Model specification requires careful consideration and thought. Any model selected should balance established behavioral theory, previous empirical research, and practical considerations for intervention delivery.

Statistical Evaluation

To better understand how this model may be evaluated statistically, it is worthwhile to examine one of the potential mediational pathways depicted in Figure 4 as an example: Figure 5 shows how each of the individual mediational pathways for this model may be assessed statistically, using the BCT of Information about Health Consequences (termed ‘Information’ in figure 5) and the psychological MoBC of Hypertension Knowledge as an example.

Figure 5: Example Latent Growth Model.

Note: This model shows how an Information intervention can influence change in Hypertension Knowledge over a 6- week period which, in turn, influences change in medication adherence over the same 6-week period. As is traditional in latent growth models in psychology, observed variables are shown with rectangles while latent variables are denoted with ovals.

More specifically, Figure 5 presents an example of how latent growth models could be utilized to study how the BCT of Information about Health Consequences acts on medication adherence via Hypertension Knowledge. While many approaches could be used to explore this question, latent growth modeling is chosen as an example because it presents an established statistical methodology that has often been used to study changes in mechanisms of behavior.^91-93 Latent growth models are a particular type of latent variable models,⁹⁴ which are common in psychology and the social sciences.⁹⁵ Underlying the idea of latent variable models is the idea that there are observable data points, which represent a larger, unobservable latent variable.^{95, 96} In Figure 5, the idea is to model an underlying construct (e.g., change in Hypertension Knowledge over time) using measurable indicators of that construct (e.g., multiple survey measures of Hypertension Knowledge measured every week for six weeks). This model represents how an Information about Health Consequences BCT intervention can cause change over time via the mechanism of Hypertension Knowledge, which in turn influences changes in health behavior over time. In contrast to traditional models examining statistical mediation,⁹⁷ this approach examines change in multiple variables simultaneously rather than sequentially. In addition, latent variable models incorporate measurement error into the analysis of how variables change over time.⁹⁵

To fully evaluate the conceptual model shown in Figure 4, it would require expanding the example model depicted in Figure 5 to include all 4 BCTs influencing change in all 3 MoBCs, which in turn influences medication adherence. This is difficult to visualize given the sheer volume of latent and observed variables, which would need to be depicted, as well as the path arrows indicating the associations between each of these variables. Though difficult to visualize, latent variable models are capable of examining the conceptual model shown in Figure 4 while accounting for multiple measurements and potential measurement error. Thus, latent growth modeling in particular, and latent variable models in general, may be powerful tools to incorporate into analyses of behavioral interventions.

With the conceptual model shown in Figure 4 and the analytic approach depicted in Figure 5, it is possible to see how an intervention could target three unique, modifiable, psychological causes for poor adherence to medications (i.e., Hypertension Knowledge; Self-Efficacy; and Behavioral Automaticity). By studying an intervention in this manner, it is possible to study not only what key components are essential for changing behavior (i.e., BCTs), it is also possible to study how interventions succeed via MoBCs.

Steps for the Future

Arguably, by utilizing the behavioral intervention ontologies developed by the Human Behavior Change Project (HBCP) and the mechanistic focus of the SOBC, there is the potential to empirically examine the MoBCs by which antihypertensive medication adherence can be influenced and to evaluate the effects of BCTs designed to target these MoBCs. Our hope in proposing this approach is that researchers and clinicians can standardize what components of BCT interventions are most effective and how these interventions are effective across multiple putative MoBCs.

There is even the potential of personalization for this intervention approach. For example, this could be achieved by identifying the most important causes of nonadherence among particular patients and then providing them with tailored BCT interventions, assessing how these interventions influence key MoBCs over time.

The authors have previously utilized this empirical approach to study the BCTs which influence physical activity via the key MoBC of Self-Efficacy,⁸⁷ thus hoping to expand this approach to other health behaviors utilizing established BCTs and MoBCs.

Limitations

Several limitations should be noted. First, as with any intervention, it is possible that BCTs (delivered individually or in combination) will not universally influence adherence and may not influence adherence via the proposed MoBCs. Though previous research has identified links between BCTs and MoBCs,⁴⁶ there is no guarantee that the identified associations can be replicated in all populations, for all intervention modalities, and for all health behaviors.

Second, it must be acknowledged that since health behaviors like adherence often have multiple root causes, multiple potential MoBCs for adherence must be considered in any approach. Conducting multiple mediation models with sufficient statistical power can be a complex process,⁹⁸ potentially requiring large sample sizes for cases where effect size estimates are not accurately defined for all paths.

Finally, though BCTs interventions may be designed to address the most common modifiable MoBCs for nonadherence, it is unlikely that all reasons for nonadherence will be addressed by any one intervention. There are many societal factors or factors existing outside the patient, which are unlikely to be fully addressed by this approach. There are also unmodifiable causes for nonadherence which may significantly influence behavior. Despite the potential limitations, it is still important to address and improve the elements of nonadherence that can be influenced.

Conclusion

Nonadherence to antihypertensive medications remains a critical and under-addressed target for behavioral interventions. Initiatives such as the SOBC and the HBCP have already provided an appropriate framework for addressing the modifiable psychological causes of antihypertensive nonadherence through structured and evidence-based frameworks. The proof will be in how RCTs incorporating MoBC measurement and BCT interventions can influence health behaviors such as medication adherence. From the extensive body of literature, it is possible to improve our theoretical understanding of what intervention components researchers and clinicians use to improve health behaviors (i.e., the BCTs) and how these interventions work (i.e., the MoBCs). This approach may enable clinicians and researchers to more effectively change health behaviors, ultimately improving clinical outcomes.

Key Points:

• Psychological barriers to treatment adherence are important targets for behavioral interventions to improve cardiovascular disease.

• The mechanistic work of the Science of Behavior Change program and the Human Behavior Change Project provides an important framework for identifying the mechanisms of health behavior and the components of interventions to change behavior.

• Using the example of adherence to antihypertensive medication, it is possible to identify methods for intervening on poor adherence behaviors and how to target the underlying mechanisms of adherence.

• Utilizing behavioral theory to inform the design of health behavior interventions could be an important approach for improving adherence to positive cardiovascular health behaviors.

Clinics Care Points:

• Patient adherence is essential for improving cardiovascular health and preventing cardiovascular disease.

• So far, the modifiable, psychological causes for health behaviors or how to target interventions to improve these behaviors are not fully understood.

• Drawing on existing research that highlights the mechanisms underlying behavior (such as Self-Efficacy) and the components of interventions (such as encouraging Problem Solving) offers a behavior change approach grounded in behavioral theory.

• Using targeted interventions that address the underlying mechanisms contributing to behaviors such as adherence has the potential to improve cardiovascular health.