Facilitators and Barriers to Adherence and Engagement in a Lifestyle Intervention for High-Risk Older Adults: A Multi-Trial Analysis
PMCID: PMC12986953
PMID:
Abstract
Background : High adherence and engagement are critical components of successful lifestyle interventions for health outcomes. Gaps persist in understanding the multifaceted factors influencing older participants’ ability to adhere and engage with health behavior interventions. Methods : Using data from two pilot randomized clinical trials evaluating a nutrition program in older adults at increased risk for dementia, we evaluated participant-level factors influencing adherence and engagement with the recommended diet and program. Additional qualitative themes were used to provide context. Both trials are registered on ClinicalTrials.gov (Pilot A: NCT0417176 on 23 March 2021 and Pilot B: NCT06121986 on 30 October 2023). Results : Better income, cognitive functioning, emotion regulation, quality of life, and independence in activities of daily living were associated with better dietary adherence, while being non-White, living in a rural area, having higher depressive symptoms, worse health symptoms, and worse sleep quality were negatively associated with adherence. Higher education, cognitive function, anxiety, and previous weight challenges were associated with better program engagement. Conclusions : Adherence and engagement were impacted by a combination of individual factors, including cognition, mood, physical health, as well as the broader socioeconomic context. Our findings highlight the ways psychological and social determinants of health may impact adherence to lifestyle interventions.
Full Text
Suboptimal adherence to lifestyle interventions results in reduced therapeutic efficacy, diminished quality of life, unfavorable health outcomes, and escalated healthcare expenditures [1]. While ensuring adherence and engagement to interventions is challenging [2], evidence suggests that enhancing adherence and engagement to treatment regimens and healthcare recommendations can significantly contribute to better health outcomes, especially for individuals with chronic health conditions [3]. Despite this growing body of work, much of the existing literature focuses on intervention efficacy rather than the contextual factors that influence long-term engagement, leaving important gaps in understanding why adherence varies across individuals and populations.
Adherence and engagement are not shaped by individual choice alone. Instead, they are influenced by broader social, economic, educational, environmental, and healthcare-related conditions that affect an individual’s capacity to initiate and sustain lifestyle change. These conditions align with the social determinants of health (SDOH), where the circumstances in which individuals are born, live, learn, work, and age influence health behaviors, functioning, and quality-of-life outcomes. SDOH are commonly organized into five domains: Economic Stability; Education Access and Quality; Healthcare Access and Quality; Neighborhood and Built Environment; and Social and Community Context [4]. Although this framework has been widely applied to explain disparities in health outcomes, its application to adherence and engagement in lifestyle intervention, particularly among older adults, remains limited.
Later life represents a period in which multiple social determinants of health intersect and intensify, shaping engagement with health-promoting behaviors. Older adults represent a population for which lifestyle interventions are both highly relevant and pose unique challenges. Aging is often accompanied by increased rates of chronic disease [5], physical and functional limitations [6], and changing roles in society [7], all of which may interact or influence engagement in behavioral interventions. However, many studies examining adherence combine age groups, or treat older adults as a broad category, which can limit insight into how age-related social and contextual factors shape adherence and engagement. As a result, there is a need for research that explicitly examines adherence and engagement within older adult populations through a framework that accounts for late-life changes in social determinants.
Within the SDOH framework, education access and quality play a central role in shaping health literacy, self-management, and engagement with lifestyle interventions. Individual differences and personal characteristics may help to characterize which older adults are most responsive to lifestyle interventions by reflecting underlying social and structural conditions that shape health behaviors. Examining this relationship can offer valuable insights for refining and customizing interventions to improve acceptability, scalability, and efficacy. Educational attainment, a key indicator of education access and quality, has consistently been associated with greater adherence to various intervention modifications and treatments [8,9]. However, these factors are often examined independently, rather than as interconnected determinants embedded within broader social contexts.
Social and community context, including social positioning and lived experience, further shapes adherence to lifestyle interventions. The relationship between age and adherence to lifestyle interventions remains inconsistent, with some studies suggesting greater adherence among older adults [10], while others report higher adherence among younger age groups [8,11]. Racial disparities in adherence and retention to physical activity and diet interventions have also been identified [12], with some researchers finding poorer adherence amongst African Americans compared to white individuals [13]. Additionally, adherence rates in lifestyle intervention programs are often lower among women compared to men [14,15]. Together, these findings highlight the importance of considering broader social and structural context when evaluating adherence.
While this study is grounded in the SDOH framework, individual-level psychological factors also play an important role in shaping adherence and engagement. Self-efficacy [16], motivation [16], stress [17], quality of life [18], and cognitive functioning [19] play important roles in sustaining behavioral change. For example, the Self-Efficacy Scale for Adherence to the Mediterranean Diet study found that adherence increased when participants reported greater confidence in their ability to adhere to the diet, alongside positive outcome expectations, autonomous motivation, affective balance, and life satisfaction [20]. Similarly, within a Lifestyle and Cardiovascular Risk Modification program, high perceived self-efficacy regarding eating habits emerged as a significant predictor of adherence [10]. Successful intervention outcomes have also been associated with higher levels of autonomous motivation, self-efficacy, and self-regulation skills [21], while lower perceived stress and higher quality of life have been linked to greater engagement in lifestyle modifications [22,23]. However, their role within the broader SDOH framework in later life remains unexamined.
Social relationships and daily environments shape participation in lifestyle interventions [24]. Factors such as social well-being, relationship status, living situation, and occupational status all play a role in an individual’s overall health profile [25] and capacity to engage in lifestyle interventions. Among older adults, adherence to dietary interventions has been shown to be lower among both employed and unemployed individuals compared to retired counterparts [11]. Conversely, individuals who report greater emotional support, such as close friendships, demonstrate higher adherence to lifestyle intervention programs [26]. Older adults who live with family members or in structured care environments have been found to exhibit healthier dietary patterns [27]. Even so, social context remains insufficiently integrated into models of adherence among older adults.
Physical health constraints and functional capacity intersect with neighborhood and built environment and healthcare access and quality. Health symptoms (e.g., fatigue, pain) and functional limitations can hinder participation in dietary and physical activity components and increase attrition. Among individuals with chronic conditions, fatigue, poor fitness [28], and joint pain [29] are commonly cited barriers to maintaining exercise and diet routines. In contrast, greater independence in activities of daily living and higher baseline physical activity levels are associated with improved adherence and greater likelihood of achieving intervention goals [28,30,31]. Poor sleep quality may further reduce motivation and adherence [32]. These findings highlight the role of functional capacity in shaping adherence within lifestyle interventions.
Pilot A (N = 58) was a two-arm, block randomized pilot clinical trial investigating the use of motivational interviewing (MI) and cognitive behavioral therapy (CBT) techniques to enhance adherence to a Modified Mediterranean ketogenic diet (MMKD) through the Improving Cognitive Aging through Nutrition (ICAN) program. One arm contained an education-only component (N = 29), whereas the other incorporated MI-CBT techniques embedded within the program’s sessions (N = 29). Therefore, this pilot aimed to evaluate whether MI-CBT strategies enhanced adherence to an MMDK within a group setting, compared to the group with an education-only component. A detailed description of Pilot A methods, the tointervention, and primary outcomes have been previously reported [33].
Pilot B (N = 65) expanded on the findings of Pilot A using a 2 × 2 factorial design in which participants were assigned to one of four study arms: (1) MMKD ICAN program + ongoing support group (N = 16), (2) MMKD ICAN program (N = 17), (3) Mediterranean diet (MED) ICAN program + ongoing support group (N = 15), or (4) MED ICAN program (N = 17). Based on the results demonstrating that the MI-CBT arm was associated with better adherence, retention, and engagement from Pilot A [33], all arms included the MI-CBT components. Of note, 3 of the 65 participants were assigned to a group non-randomly (due to factors unrelated to study arm, e.g., group time/day availability), and one participant was withdrawn from the study prior to beginning the program due to a dementia diagnosis not identified at screening. It is also important to emphasize that the current analyses did not explore differences between the four arms, as the primary focus is on non-program factors affecting adherence to the program more broadly.
The ICAN program, used in both pilots, consisted of one-hour sessions, except for the first session, which lasted 90 min. Sessions were delivered via HIPAA-compliant Zoom with the first session being held in person in Pilot B. Nutrition information and psychoeducation components were presented using pre-recorded video slides. During sessions, participants were provided with detailed guidance on identifying, purchasing, and planning meals based on each dietary plan, and participants had full autonomy for purchasing and preparing their own meals. All participants received a workbook containing handouts aligning with the weekly session content (e.g., nutrition, goal setting, macronutrient tracking) and a cookbook. Weekly online surveys were collected throughout the intervention sessions. This intervention package was developed using the Information-Motivation-Behavioral skills model [34]. Sessions were led by two trained facilitators who guided participants through educational content, structured skills practice, and opportunities for discussion and questions. Facilitator training included passing nutrition knowledge checks, reviewing all program manuals and presentations, completing an MI-CBT workshop, and completing additional relevant readings and peer-practice sessions. To maintain intervention fidelity, all sessions were recorded, reviewed, and rated by a licensed psychologist and certified nutrition coach using the ICAN fidelity checklist. This checklist was designed to evaluate the consistent use of MI and CBT skills, accurate nutrition guidance, as well as appropriate use of session prompts and materials. Facilitators were supervised by a licensed psychologist and registered dietitian.
Cognitive Function was assessed using the Telephone Montreal Cognitive Assessment (T-MoCA), with scores converted to the MoCA-30 scale using the equipercentile equating method [35]. Scores of 26 or above were considered normal, whereas scores between 18 and 25 suggest possible MCI [36].
Memory Complaints were measured by the 7-item Memory Complain Scale (MCS) [37]. Each item was rated on a three-point Likert scale (0–2), yielding a total score ranging from 0 to 14 (higher indicates worse memory concerns).
Depression was evaluated using the 9-item Patient Health Questionnaire (PHQ-9) [38,39]. Participants rated the frequency of depressive symptoms experienced over the past two weeks on a four-point scale ranging from 0 (“not at all”) to 3 (“nearly every day”). Total scores range from 0 to 27.
Anxiety. The 10-item Geriatric Anxiety Scale (GAS-10) was utilized to assess anxiety symptoms among older adults [40,41]. Each item was rated on a four-point scale ranging from 0 (“Not at all”) to 3 (“All of the time”), based on the frequency of symptoms experienced over the past week. Total scores range from 0 to 30.
Stress. The 10-item Perceived Stress Scale (PSS-10) [42,43] assessed stress during the last month using a 5-point Likert scale from 0 (“never”) to 4 (“very often”). Four positively worded items (items 4, 5, 7, and 8) were reverse coded so that higher scores indicate higher levels of perceived stress. Total scores range from 0 to 40.
Emotion Regulation. The Positive and Negative Affect Schedule (PANAS) [44,45] was employed in Pilot A, which consists of 20 items measuring positive and negative feelings over the past week. Sum scores ranging from 10 to 50 were calculated separately for positive and negative affect. In Pilot B, participants’ emotion regulation strategies were assessed using the 10-item Emotion Regulation Questionnaire [46,47]. Items were rated on a 7-point scale from strongly disagree (1) to strongly agree (7). Two subscale scores were computed: cognitive reappraisal (six items) and expressive suppression (four items).
Self-Efficacy was evaluated using the 10-item Self-Efficacy (GSE) Scale [48,49]. Participants rated how true each statement was for them on a four-point scale from not at all true (1) to exactly true (4).
Previous Eating Disorder was assessed using the SCOFF questionnaire [50,51], a five-item yes/no screening tool designed to identify individuals at risk for eating disorders. Higher total counts suggest an increased likelihood of having an eating disorder.
Quality of Life was measured using the 13-item Older People’s Quality of Life questionnaire (OPQOL) [52], range = 13 (worst)–65 (best). Responses were recorded on a 5-point Likert scale from 1 (“strongly disagree”) to 5 (“strongly agree”).
Interpersonal Support was assessed using the 12-item Interpersonal Support Evaluation List (ISEL), which evaluates perceptions of social support including appraisal support, belonging support, and tangible support [53,54]. Each item was rated on a 4-point scale ranging from 1 (“definitely false”) to 4 (“definitely true”). Six negatively worded items were reverse coded, and a total sum score was calculated.
Functional Status. In Pilot A, the Functional Status Questionnaire (FSQ) [55] was utilized to evaluate participants’ physical, psychological, social, and role functioning during the past month, and was divided into six subscales. Average scores for each subscale were calculated based on valid items and linearly transformed to a 0–100 scale, with higher scaled scores representing better functioning.
General Symptoms. In Pilot B, participants reported how often they were bothered by a list of 30 general symptoms in the past 2 weeks on a five-point scale from 0 = “not at all” to 4 = “very much” [56]. A sum score ranging from 0 to 120 was calculated.
Pain was measured by 24 items adapted from the Roland–Morris Disability Questionnaire (RMDQ) assessing functional disability related to lower back pain [57]. In our studies, we expanded the items to general pain. Participants indicate whether each statement applies to them (“Yes” = 1, “No” = 0), with a total score range of 0 (no pain) to 24 (severe pain).
Sleep Quality. The Pittsburgh Sleep Quality Index (PSQI) [58] was used to assess participants’ sleep quality. Seven component scores including subjective sleep quality, sleep latency, sleep duration, sleep efficiency, sleep disturbance, use of sleep medication, and daytime dysfunction were calculated based on participants’ corresponding survey responses, with each component ranging from 0 to 3. A global sum score ranging from 0 to 21 was calculated, with higher scores indicating poorer overall sleep quality.
Mediterranean Diet Adherence. Adherence to the MED was measured using the 14-item Mediterranean Diet Adherence Screener (MEDAS) [59] in Pilot B. Each item was scored 0 or 1, based on whether participants met specific dietary criteria. Total scores range from 0 to 14.
Audio recordings from each study interview were transcribed and subjected to reflexive thematic analysis, as described by Braun and Clarke [60]. The analysis followed Braun and Clarke’s six-phase approach: familiarization with the data, generating initial codes, searching for themes, reviewing themes, defining and naming themes, and producing the report. An inductive, data-driven approach was used to allow themes to emerge directly from the transcripts rather than being guided by pre-existing theoretical framework.
Two coders independently conducted an initial review of all transcripts. Each coder generated preliminary codes and identified patterns related to facilitators and barriers. Following independent coding, the coders met to compare coding structures, discuss discrepancies, and refine interpretations to enhance analytic consistency. Disagreements that could not be resolved through discussion were reviewed by a third investigator to reach final consensus. Coding, organization, and interpretative analysis were facilitated by NVivo 20 software [61].
Table 1 summarizes the sample characteristics for Pilot A (N = 58) and Pilot B (N = 65). Notably, significant differences were observed in cognitive function, memory complaints, physical activity, group cohesion, and all outcome measures. Compared with participants in Pilot A, those in Pilot B exhibited greater cognitive impairment, more memory complaints, and lower levels of physical activity at baseline. With respect to outcome measures, participants in Pilot B demonstrated lower diet adherence, but higher engagement in terms of session attendance and food log submission.
Complete results are shown in Appendix A Table A1, Table A2, Table A3, Table A4 and Table A5. Economic stability factors, including household income and employment status were associated with adherence and engagement, particularly in Pilot B. Household income was positively correlated with both session attendance (r = 0.26, 95% CI [0.02, 0.49]) and food log submission (ϕ = 0.28, 95% CI [0.04, 0.52]).
Several sociodemographic factors were associated with dietary and program adherence across the two cohorts. In Pilot A, Black participants reported lower adherence; however, in the more racially diverse sample from Pilot B, no racial differences in adherence were identified. These discrepant findings may suggest that, while adherence may at times be shaped by racial or ethnic background, inclusion of a representative sample may eliminate such biased findings. Similarly, inclusion of a more geographically representative sample in Pilot B more clearly demonstrated the adherence challenges faced by rural participants who tended to describe more challenges related to accessing healthy foods and higher food costs. These findings are consistent with prior research showing that factors affecting adherence to a MED included limited access to healthy foods and cost [62]. Household income was positively associated with engagement behaviors such as attendance and food log submission, and participants with greater financial resources or more flexible schedules (e.g., retirees) described fewer obstacles to meal preparation and self-monitoring. Educational attainment was also associated with more consistent food log submission, which may have implications for measurement of engagement and adherence for future trials. Specifically, individuals who are more academically inclined may be more likely to engage with components of programs that are similar to academic practices (e.g., class attendance, homework completion). Although it was notable that income and education were not associated with dietary adherence.
Psychological and cognitive factors were also associated with variation in dietary adherence and program engagement. Better cognitive function and fewer memory complaints were linked to higher adherence in Pilot A and to higher session attendance in Pilot B. Although there was a negative association between cognitive function and ketone adherence in Pilot B, this may reflect the higher proportion of participants with MCI in that cohort or may align with modifications to the intervention for Pilot B to specifically support memory retention for individuals with MCI. Consistent with previous studies, higher depressive symptoms were related to lower perceived adherence in Pilot B [63]. Meanwhile, higher anxiety and stress were unexpectedly associated with greater session attendance. Given that the average level of anxiety and stress in the sample was relatively low, this finding may point to the potential benefits of eustress or anxiety in motivating engagement for some participants [64]. Patterns in emotion regulation also corresponded with adherence behaviors, with expressive suppression associated with higher ketone adherence and cognitive reappraisal linked to higher MED adherence but lower food log completion.
Although there were relatively fewer social factors related to adherence and engagement, one important relationship was evident in both Pilots. Specifically, participants who lived with supportive others showed higher perceived adherence than those living alone. This association was consistently reflected in the qualitative data, where participants described how shared meal preparation and encouragement from spouses, family members, or roommates supported consistency, whereas those living alone often reported difficulty sustaining motivation and preparing meals for one. Prior lifestyle programs have already demonstrated benefits of taking these types of approaches [65]. Together, these results suggest that social support and context likely play significant and meaningful roles in sustaining adherence. Programs may therefore benefit from incorporating partner- or peer-based support strategies, meal planning tools for single-person households, and structured accountability features tailored to participants with varying levels of prior experience with dietary change.