Key Highlights and Action Items

This series of workshops and meetings, sponsored by OBSSR and the NIH-Wide Adherence Research Network, aims to develop a more complex understanding of health behavior maintenance to promote and sustain health outcomes. The first workshop focused on conceptualizing the processes of behavior maintenance, and the second focused on measurement, monitoring, and innovative tools. Workshop 3 focused on intervention approaches, designs, and strategies to support behavior maintenance. 

The first session of Workshop 3 highlighted exemplar interventions that have been used to promote behavior maintenance, and presenters highlighted specific strategies designed for behavior maintenance across behaviors and settings. In the second session, an interactive panel discussion elucidated gaps in and opportunities for advancing the understanding of behavior maintenance in behavioral interventions. 

State of the Science for Promoting Behavior Maintenance in Diabetes Management/Weight Loss: A Tale of Two Cities

• Engaging the community and understanding community needs, preferences, and resources is essential to developing and sustaining effective interventions. Maintenance programs can be implemented across community organizations with varying resources. Creation of community hubs, or central resources within a community, can help provide infrastructure for sustained communication with community-based organizations and providers, thereby monitoring effectiveness of the approach and identifying needed modifications. Sustainability should be considered at the onset of any research project, particularly when engaging the community, to ensure that resources and programs can be maintained.
• Implementation science is focused on identifying multilevel barriers and facilitators to providing and employing evidence-based treatments. Community-based interventions need to identify and engage multiple levels (e.g. individual, family, community, health care system, policy) for effective behavior maintenance strategies. As distinct from targeted individual-level behavior change strategies, Systemwide interventions appear key to sustainability: infrastructure and policy-level interventions can be highly effective in promoting and supporting sustained behavior maintenance. 
• From the perspective of a clinician scientist, the question was addressed regarding how clinicians can promote and sustain behavior maintenance in the ongoing practice of clinical care. Clinicians often can identify the need for behavior change but struggle to support their patients in behavior change and maintenance because of contextual factors often outside their scope of practice. System-level biases that affect a patient’s ability to change and maintain behavior should be acknowledged and addressed in clinical settings. Severe or acute stressors – medical or lifecourse - can cause lapses or change in behavior; some patients have other critical needs that must be met first. Understanding these stressors and needs—and providing support for those needs—is key to promoting behavior maintenance. 
• Better understanding of actual behavioral components and underlying mechanisms of action of behavior maintenance could help clinical researchers test intervention strategies.

Mechanisms for Behavior Maintenance in the Context of Intervention Development: Presentations of Exemplar Interventions

The following considerations were brought up in discussions regarding development of effective behavior maintenance interventions. Workshop participants elevated these conditions/factors with potential to effect mechanisms of action.

• Researchers should specify the mechanisms that an intervention strategy (or a set of strategies) are designed to change, and attention should be paid regarding whether the mechanism is being targeted to elicit initial changes in behavior, maintained changes in behavior, or both outcomes. Some strategies may be more effective at eliciting initial changes in behavior, whereas others may be more effective at eliciting sustained behavior maintenance.
• Behavior is not static, but fluctuates over time, reflecting numerous dynamic influences. These influential factors occur at and effect multiple levels. Effects vary across individuals and levels. These considerations should be accounted for in studies testing behavior maintenance strategies and specific mechanisms of action. 
• Intervention timing in terms of prevention or disease course may also be a significant factor to be considered in the development of behavior maintenance interventions. For example, an intervention might target behavior maintenance after initial establishment of a health behavior or remediate a slip, lapse or discontinuation.
• Personal health concerns in and of themselves are often not sufficient to motivate behavior maintenance—social, cultural, and environmental context must also be considered and can be harnessed in intervention development. 
• Peer and social networks can be leveraged to foster accountability, which, in turn, strengthens behavior maintenance. Positive social support for maintaining behavior can also be a mechanism of action, but maintaining the relationships that afford this support long term can be challenging. 
• Behaviors that are tied to group affiliation/identity and group-based activities have the potential to enhance behavior maintenance. Strategies that promote identity integration (i.e., linking personal values/group affiliations to behavior change processes) is a promising approach to boosting adherence/adoption of behavior. 
• Health behaviors that occur regularly – perhaps as regards time or location -- are promising targets for interventions. However, it is important to clearly define the characteristics that underly the regularity of the behavior and identify strategies that can capitalize on those features. 
• Nudges, or changes in the way choices are presented or information is framed are often low in cost, scalable, and can help guide or motivate decision-making. Routines can help people become more aware of their behaviors and can serve as “memory traces” that help people sustain or resume healthy behaviors after they have a temporary lapse in execution. 
• Adaptive interventions have the potential to promote behavior maintenance over time, and timely targeted adaptations to intervention strategies might be crucial. Automated approaches can be effective as technological capabilities continue to advance. An example facilitated by use technology is automated “nudges” when lapses are electronically noted.  Once the skills and frequency of a health behavior have been achieved, criteria for what constitutes a lapse should be defined a priori in intervention development. If you have a clearly specified indicator of a lapse, you can design the intervention to monitor for and respond to it.
• A potential target for adaptive interventions is intervening on “all or nothing” thinking in response to barriers to behavioral execution in the context disruptive stressors.

Research Gaps and Future Directions: Panel Discussion

• The field needs consensus on the differences in criteria for and definitions of the terms “behavior maintenance” and “habit”. Not all behavior maintenance is habitual. Factors influencing, driving or which are barriers to execution may overlap or be discrete. Identifying person- and context-dependent determinants of behavior can inform real-time technology-enabled interventions (See comments on timing and nudges).
• Sustaining health interventions and programs can serve to promote health equity. Testing community engagement and involvements critical to supporting individual level behavior maintenance has the potential to inform policies and resource allocation.
• A key area for future research includes defining deviations or lapse in behavior and reliable and valid indicators for deviations from maintenance. How can we adapt this construct across various domains (e.g. physical activity, diet, medication use)? Can we predict the risk of a lapse before it occurs? Approaches to detect risks for deviation are needed. The use of technology allows for data capture in real time to inform behavior maintenance and deviations from maintenance; this also offers new opportunities for diverse intervention approaches.
• There does not appear to be one “gold standard” approach to achieving health behavior maintenance. When testing interventions, contextual characteristics such as the need for all or none/abrupt behavior change vs. gradual approaches should be considered.  Approaches may need to be tailored for specific individuals, diseases, and behaviors. 
• Some motivations and mechanisms of action can be targeted to help promote downstream healthy behaviors that are indirectly related but still affect desired health outcomes. More research is needed to quantify these relationships and processes.