The aim of this scoping review was to systematically chart studies assessing psychosocial factors as moderators of behavioral change in primary CVD prevention targeting adult populations. The further aim was to summarize current knowledge and identify potential knowledge gaps.
The overall findings from this review indicate that a wide range of psychosocial factors may be of importance in lifestyle interventions, and may therefore have implications for developing or improving lifestyle interventions in the future. However, it is difficult to draw more detailed conclusions due to the heterogeneity among the studies.
A total of 35 studies were included. The most common lifestyle behaviors targeted with interventions were physical activity, followed by diet, whereas fewer studies evaluated the intervention effects on alcohol consumption and smoking cessation. Only one study evaluated a composite CVD risk score as an outcome. Overall, most studies reported moderating effects of a psychosocial factor on the outcome in the intervention group. Other studies reported a moderating effect only in the control group, which can still indicate the relevance of the psychosocial factor in relation to the intervention. For example, in Pfeffer et al. (2019), high intention strength predicted increased physical activity only in the control group, which can be interpreted as the intervention helping those with low intention strength to overcome this deficit. We charted the components of lifestyle interventions present in the different studies (Table 7), but given the heterogeneity among the studies, no visible patterns were found between specific types of interventions and the moderating effect of the psychosocial factors.
Evidence within categories of psychosocial moderators
The studies on self-efficacy and motivation show mixed results. This is in line with the results of a review on factors associated with attendance and attrition in weight management interventions [18] which also showed inconsistent results for self-efficacy. Most studies in this category showed a moderating effect such that higher self-efficacy or motivation was associated with larger intervention effects. In contrast, Elbert et al. (2017) [35] showed that those with low self-efficacy benefited more from an intervention aiming to increase vegetable consumption among university students.
Of the eight studies investigating social support and relationship quality, seven reported a moderating effect. However, in Westland et al. [48], low social support was associated with increased physical activity in the intervention group. Most of the studies included in this review investigated samples with a majority of women, however, the studies investigating the role of social support and relationship quality were exceptions in this regard, as four of the studies in this category had roughly equal gender distribution. Overall, the results are in line with previous qualitative studies that have identified social support as a perceived facilitator of adherence to lifestyle modification programs [22] and of the maintenance of changed health behaviors over time [21]. These results are also in line with a 2017 review by Leung et al. [18] which concluded that social support is associated with a higher level of attendance in weight management interventions.
Overall, the studies investigating the moderating effect of mental health indicate that aspects of poor mental health (e.g., depression, anxiety, and perceived stress) are associated with less favorable outcomes in lifestyle interventions. In two studies in which no moderating effects were found, the prevalence of depression among participants was low. It is possible that those with better mental health also have higher self-efficacy, motivation and restraint behavior, factors that have been associated with weight loss [66]. In addition, those with higher levels of perceived stress may be less likely to regulate their eating behavior [67]. Depression, when studied in isolation, did not moderate the effect of an intervention aimed at changing lifestyle behaviors [52, 53], but this may only apply to people with less severe depression. In one of these studies [52] those with major depression were excluded and in the other study [53] the prevalence of depression was low. Furthermore, people with depression can be less likely to participate in and complete a lifestyle intervention program [20]. It is therefore difficult to determine whether depression is a moderator of the success of a lifestyle intervention when those at the higher end of the depression spectrum are not included in studies. Comparatively, a study that included a higher proportion of people with depression, half of whom were on treatment found a significant moderating effect [54].
Taken together, the studies on personality and emotion indicate positive moderating effects. More specifically, the personality traits neuroticism, perseveration, and conscientiousness as well as optimism and self-compassion were positively related to intervention effects, whereas impulsivity was inversely related to intervention effects. Previous research on the association between personality and health behavior in general provides support for these findings. For example, studies indicate that conscientiousness is positively related to health behaviors [68], such as healthy eating [69, 70] and physical activity, and is negatively related to sedentary behavior [71]. With regard to neuroticism and emotional instability, some previous studies indicate that these constructs are negatively related to health behavior, but the findings are mixed [72]. Recent research suggests that “healthy neuroticism”, defined as the interaction of neuroticism and conscientiousness, is positively linked to health-related behaviors [64]. This is in line with the findings reported in Burnos et al. [57] in which participants high in neuroticism, perseveration (a facet of conscientiousness) and emotional reactivity improved in terms of health practices. As for the personality traits of introversion and extraversion, the findings of Sakane et al. [60] indicate that introvert participants have a larger intervention effect in terms of weight loss compared to more extrovert participants. This is in line with the findings of Sutin et al. [71], reporting that extraversion was positively associated with BMI in a large community sample. However, other studies have reported none [73, 74] or negative [75] associations between weight and extraversion.
Regarding studies on cognition, all studies indicated moderating effects of cognitive factors, more specifically health literacy, self-affirmation and cognitive processing style. This review included two studies that showed that higher health literacy enhanced the intervention effect to improve physical activity and dietary behaviors [62, 63]. The literature indicates that health literacy also may enhance self-efficacy and perception of risk [76], which may further contribute to behavior change in lifestyle interventions. Further, those who perceived their health as poor were more likely to change their behaviors [63]. Of note is that attrition was high in this category, with three of the studies losing roughly half of the sample between baseline and follow-up.
Methodology and sampling
Given the heterogeneity within the field of research in terms of e.g., types of intervention, outcome measures, and psychosocial moderators, a scoping review was conducted to broadly summarize available research and identify knowledge gaps. Several important methodological issues were observed across the included studies, the first of which concerns the assessment of both psychosocial factors, and the outcome variables. (1) There are often many risk factors involved leading up to a CVD. Still, most of the studies focused on a single risk factor. In fact, only one study used a composite score assessing risk of CVD. (2) Five studies assessed moderating effects using sub-group analyses rather than interaction analysis, and thus these results should be interpreted with caution. (3) The outcomes were assessed through self-reports in most studies, which might bring some uncertainty to the reliability of the results. (4) The quality of the assessment of the psychosocial moderators is of particular importance for this review. Despite the availability of comprehensive, accessible, and validated instruments to measure psychosocial constructs, about half of the included studies did not assess psychosocial variables using validated instruments. Consequently, the use of unvalidated instruments may introduce measurement error, which in turn may lead to erroneous conclusions regarding the effect of a psychosocial factor as a moderator. Thus, moderating effects of psychosocial factors could either be inflated, or remain undetected, potentially affecting clinical recommendations and possibly the effectiveness of future interventions.
Other issues concerned the samples investigated. (1) The samples in many of the studies were homogenous in terms of demographic factors such as age, sex, and place of residence. As in all studies, this makes the sample well defined and can provide control for certain background factors, but it also contributes to the risk of type-II error and makes it more difficult to generalize results to a general population. (2) In about half of the studies, the sample consisted of participants who were either considered to be at elevated risk of CVD, or had at least one risk factor of CVD present, e.g., obesity, hypertension, sedentary lifestyle, smoking, or heavy drinking. Though it is important to study these kinds of samples in order to facilitate and investigate behavioral change amongst those most at risk, a majority of those who experience a cardiovascular event for the first time were low to moderate in CVD risk prior to the event. Therefore, interventions aimed at the general population are also of significant importance. (3) The focus of most studies has been on middle aged to older people. Only eight studies investigated samples with a mean age younger than 30 years. Of note is that all studies using alcohol consumption as an outcome were conducted on samples with a mean age < 30 years. 4) The gender distribution among the included studies was skewed, as 24 of the studies had a sample with a majority of women. 5) Most studies were conducted in Europe or North America, making it difficult to draw any general conclusions for other geographical areas. 6) In 10 studies participants were recruited through advertisements which excluded potential participants not being exposed to the advertisement. More importantly, the participants in these studies needed to have actively sought out and entered the trial. Therefore, these samples are likely to include individuals with higher motivation than a randomly selected sample. 7) Only 14 of the studies suffered attrition rates less than 20%, and in eight studies the attrition exceeded 40%. Drop-outs from health-promoting interventions are more likely to be at higher risk and suffer worse outcomes. Therefore, the attrition rates may have consequences for the generalizability of the results.
Only five of the included studies fulfilled the criteria of being an RCT with a follow-up time of 12 months or longer and had a sample size of more than 250 participants, which in addition assessed the psychosocial moderator with a validated instrument and carried out an interaction analyses to evaluate the moderating effect. Three of these studies [49, 51, 52] investigated the moderating role of depression, whereas the other two [42, 43] evaluated the moderating effect of social support or relationship quality.
Knowledge gaps and implications for future research
Taken together, the results from this review show a large heterogeneity in terms of study design, assessment of outcomes, potential moderators, and the samples being studied. Our findings suggest that self-efficacy may be a robust effect modifier of lifestyle interventions targeting physical activity regardless of age, intervention types or follow-up duration. However, the considerable heterogeneity among the studies limits our ability to draw definitive conclusions for most of the psychosocial factors assessed, making it difficult to translate these findings into practice. This review also highlights important gaps in the current evidence base. In particular, more research is needed regarding psychosocial factors which are not easily modifiable, e.g., depression, and regarding interventions aimed at alcohol and smoking cessation. Furthermore, we find that there is currently a lack of RCTs using (1) large samples, preferably from the general population, (2) reliable and valid instruments for assessing psychosocial variables, and (3) objectively measured outcomes, either exclusively or as a complement to self-reports. More studies fitting these criteria are needed in order to gain better understanding for the role of psychosocial factors in interventions for CVD prevention.
Compared to the number of studies on the effect of various CVD prevention and lifestyle modification programs, there is a lack of studies on the role of psychosocial factors. Despite this review finding mixed results regarding the moderating effect of psychosocial factors on lifestyle interventions, we conclude that it is important to consider these factors when planning and evaluating lifestyle interventions. However, caution is warranted when tailoring interventions based on assessed psychosocial factors. Differences in study populations and types of interventions must be carefully considered. We also observe a positive trend in an increasing number of studies investigating psychosocial moderators in relation to CVD prevention. As seen in Table 1, the number of studies that meet the criteria for inclusion in this scoping review shows more than a three-fold increase from the first decade of the century, to the second, and 13 of the included studies were published between 2020 and 2025. This may reflect an increasing awareness of the importance of considering psychosocial factors when developing and evaluating CVD prevention programs.
Within this field of research, there is a lack of studies on younger adult samples. Most studies included in this review comprised samples with participants of middle-age or older, and only nine studies included participants with a mean age below 30 years. Older age is a well-known, strong risk factor for CVD [77]. However, atherosclerosis, the main underlying mechanism for CVD, can start at a young age [78,79,80], as early as in a person’s 20 s [79], or even in their teens [23]. Hence, strategies for CVD-prevention should be implemented at an early age, and knowledge of psychosocial factors that may moderate adherence to lifestyle interventions differently across age groups becomes important [81].
Results from RCTs evaluating effect modification have been difficult to replicate [82]. This may be due to some studies not being powered to perform moderation analysis. Also, unequal distribution of the moderator between intervention groups may exist even after randomization, which may influence the results. Therefore, more RCTs that assess effect modification a priori are needed as well as studies that evaluate psychosocial factors as moderators of smoking and alcohol interventions.
Despite the growing popularity of online interventions, only eight of the studies investigated interventions of this format. Nevertheless, it was the most common type among studies that examined cognitive factors as potential moderators. Given their effectiveness in promoting lifestyle changes – such as increasing physical activity, improving dietary habits, and reducing smoking and alcohol consumption [83] – further research is warranted to clarify the role of psychosocial factors in the context of digital interventions specifically.
This current review focuses on psychosocial factors as moderators of effects of lifestyle interventions. However, there is another important aspect of psychosocial factors that this review does not consider, namely their role as mediators. Hence, an important topic for future research is to synthesize the evidence of psychosocial factors as mediators of lifestyle change in primary CVD prevention, in order to understand the mechanisms of health-behavior change.
Practical implications
Identifying factors that may moderate intervention effects is a crucial step in tailoring interventions to become as effective as possible. It is therefore essential to use appropriate statistical tests to distinguish between true differences in the intervention response and chance findings. Erroneous effect modification conclusions may lead to harm when individuals are excluded from an intervention based on a wrongful notion that they may not benefit. Additionally, resources may be wasted when individuals receive an intervention that is less effective in changing their behaviors.
Previous reviews have investigated factors that influence adherence or uptake to lifestyle recommendations or intervention focused on adults that have CVD [19, 20, 22]. This review adds valuable insight into psychosocial factors that may hinder or enhance effects of lifestyle interventions in persons without diagnosed CVD. Early identification of people who may not benefit optimally from a specific lifestyle intervention is essential to adapt strategies to enhance the effectiveness of prevention efforts, and to strengthen equality in health care. In this perspective, it is important to consider whether the psychosocial moderator is modifiable or not. Moderators that are not easily modifiable would call for certain preventive strategies for certain groups. For example, the findings from the current review suggest that poor mental health, and low social support, is associated with smaller intervention effects, which calls for specific strategies for certain groups. Other moderators, such as self-efficacy, motivation and risk perception, could be addressed with the design of the intervention. When designing individually tailored interventions, researchers could consider, if practically feasible, for example a screening of relevant psychosocial factors before the start of the intervention. The purpose of the screening would be to identify psychosocial factors that may influence the success of the intervention. For instance, motivational interviewing and mastery experiences (i.e. performance accomplishments) might be used as components in the intervention to strengthen motivation and self-efficacy [84, 85]. For participants with poor mental health, tailoring may involve integrating mental health support and reducing cognitive demands of the intervention – e.g., by providing short, clear descriptions, and visual aids – since mental ill-health can impair attention, memory, and executive functioning [86].
We hope that systematic reviews with meta-analyses may be conducted in the future to investigate the strengths of the moderating effects, and in more detail chart these effects in relation to specific outcomes and to specific types of interventions. In order for that to happen, more high-quality studies are needed that a priori aim to assess moderation effects. Therefore, we would encourage those researchers designing new lifestyle interventions to consider including baseline measurements of relevant psychosocial factors using validated instruments.
Limitations
This scoping review has several limitations that need to be addressed. Our aim was to include as many relevant psychosocial factors as possible in the search terms. However, it would have been practically impossible to include all possible psychosocial factors in our search. Therefore, we may have overlooked studies on the moderating effect of psychosocial factors that were not included in our search strings. In fact, two of the studies included were not identified in the searches despite fulfilling all our inclusion criteria, and were instead found through citations in articles already included.
In conducting a scoping review, our aim was to include all research on the moderating effect of psychosocial factors in interventions aimed at behaviors associated with increased CVD risk. Therefore, our search strategy was broad and resulted in a large number of unique hits. However, when scanning over 10,000 abstracts manually, studies of interest may be overlooked. In order to minimize the risk of overlooking studies for inclusion, reviewers flagged and discussed any questionable studies until consensus had been reached.
Because the psychosocial factors were person-factors, it cannot be excluded that factors correlating with the psychosocial factors can account for the moderating effects. In addition, there are likely overlaps between the categories of psychosocial moderators we created in order to systematically analyze the results. For example, some studies investigated aspects of mental health as a moderator, but mental health may also be associated with social isolation, or cognitive impairments, making it difficult to rule out effects from other psychosocial factors that were not explicitly assessed. Another point to consider is the possibility of publication bias, which occurs when results from published studies systematically differ from unpublished ones. In this scoping review, we have not assessed the likelihood of publication bias and to what extent this has affected our results. However, publication bias may lead to an exaggeration of the moderating effects of psychosocial factors.
Finally, although a majority of the studies were RCTs which adjusted for possible confounders, moderator analyses were seldom planned a priori and therefore residual confounding may remain.
Summary and conclusionsagarmid A. Rayyan-a w
The purpose of this scoping review was to systematically chart studies that investigated psychosocial factors as moderators of lifestyle intervention for primary CVD prevention. This review highlights the heterogeneity amongst the studies in terms of e.g., the psychosocial factors investigated, types of intervention, and outcome of interest. Though our results suggest that various psychosocial factors may moderate the outcome of lifestyle interventions, this heterogeneity limited our ability to draw more detailed conclusions or identify any general trends. In order to inform intervention strategies and to improve primary prevention, more high quality RCTs are needed that a priori aim to assess moderation effects, using appropriate statistical methods and validated instruments.