Background

Eating a variety of nutrient rich foods and regularly being active contributes to positive psychological and physical health outcomes [1,2,3,4]. However, many adults do not meet dietary [5,6,7] or activity [8,9,10] guidelines. A 2020 systematic review of vegetable intake in 162 countries suggested that 88% of adults consumed less than the recommended 240 g of vegetables per day [5]. Furthermore, over a quarter of adults fail to meet the recommendations to participate in 150 min of moderate-intensity or 75 min of vigorous-intensity physical activity per week [10]. The discrepancy between ‘ideal’ practice of health behaviors and reality suggests a need for more effective strategies to support the practice of beneficial diet and activity behaviors [5, 10].

Lifestyle interventions are one avenue to support behavior change [11, 12]. Historically, these interventions have tended to focus on diet or activity related practices, though it is recognized that sleep and sedentary habits are also important [7, 13]. Lifestyle interventions can be rooted in a variety of frameworks and methods [14, 15], making it difficult to determine what components or ‘active ingredients’ [14,15,16] within interventions offer merit for behavior outcomes [17,18,19]. Literature has called for greater clarity in intervention reporting [14, 16], including the use of common terminology to describe what is occurring within an intervention to support behavior change, or how it’s delivered [14, 20]. One strategy to overcome this is to explore how participants within lifestyle interventions feel they are supported or hindered to change their behaviors [15]. By identifying perceived facilitators and barriers across diverse interventions and consolidating patterns from participants’ experiences [21,22,23,24], strategies to promote positive behavior change, regardless of intervention framework, aims or terminology can be revealed [15].

Interventions targeting behavior change do not act in isolation. Instead, they overlap with participant’s personal characteristics and structures in place surrounding a participant [15]. These include social contracts, physical spaces and societal norms [12]. Using an adapted Socio-Ecological Model [12], these factors can be separated into three broad categories: Individual level factors (i.e., intrapersonal factors), environmental factors (i.e., interpersonal, community and policy factors) and intervention factors (i.e., institution). Individual level factors include a participant’s personal motivation underlying their behavior change such as their knowledge, skills, attitudes, or perceptions about change. The environment level includes aspects of the physical environment (i.e., infrastructure) and systems within it (i.e., social influences or norms). Finally, the third category (i.e., intervention level) pertain to aspects of an intervention itself (i.e., its components, delivery, and location). Though broader factors, such as the “policy level” also play a role, they are often viewed as outside of an intervention or an individual’s control [12]. For this reason, focusing on more controllable aspects of an intervention and individual motivation can help interventionists make more actionable decisions about intervention design to improve participant adherence [15].

Understanding participant perceived facilitators or barriers to diet or activity interventions can inform their design (i.e., what) and delivery (i.e., how). More efficacious interventions in turn may facilitate greater uptake and maintenance of health protective behaviors [15]. Thus, the purpose of this systematic review is to explore participant perceived barriers and facilitators to diet and/or activity changes in lifestyle interventions at the individual, environment and intervention levels.

Methods

This systematic review was registered and can be accessed at PROSPERO (ID: CRD42021262918). It has been conducted in accordance with the PRISMA Statement [25].

Search strategy and selection criteria

A literature search was conducted in five databases (i.e., MEDLINE Ovid, PubMed, Web of Science, the Cochrane Library and Google Scholar). Google scholar was searched informally to help identify additional literature. Qualitative studies describing participant’s barriers and facilitators to adherence in diet and/or PA interventions were identified. Truncation and appropriate Boolean operators were used (Table 1). Since some interventions seek to change diet or activity habits for weight loss, we included the keyword ‘weight loss’ in the search syntax to expand the results. This way, an intervention flagged for ‘weight loss’ that aimed to change diet and/or activity behaviors would be captured. The databases were searched between January 2005 to October 2021. This period was chosen as the prevalence of adults living with overweight and obesity has been relatively stable over this time.

Table 1 Search syntax used in PubMed, Goggle Scholar, MEDLINE Ovid, Web of Science Core Collection, and Cochrane Library
Full size table

Inclusion and exclusion criteria

Qualitative studies (e.g., interviews, focus groups or open responses writing data) that reported on participants perceived barriers or facilitators to behavior change during a lifestyle intervention were eligible. Lifestyle interventions were defined as those focused on changing diet (improving eating behaviors, diet quality) and/or activity (increasing frequency, type and duration) behaviors. Behavior change was defined in terms of participant perceived adherence to changing their diet and/or activity behaviors. In this way we considered ‘successes’ from the individuals’ perspective in their own words. All studies were conducted among adults (18–65 years). We chose to focus on adult populations as youth are often still confined to parental control over their behaviours, creating a situation where adherence is reliant on a third party [26]. Studies that included participants older than 65 years were eligible if the mean reported age was below 65 years.

Studies were excluded if they were conducted in populations living with dementia, cognitive impairment, physical disability, arthritis, HIV or undergoing cancer treatment. This approach was taken to control for underlying pain-related medication use that can impact diet or activity behaviors (e.g., pain from arthritis preventing activity). Studies conducted in pregnant women were also excluded. Letters, editorials, Masters and Doctoral theses were excluded. Systematic reviews of qualitative studies were back checked by hand for potentially relevant studies.

Review section and data extraction

The primary outcomes were perceived barriers and facilitators to participants’ adherence to diet and/or activity interventions. Two separate searches were conducted. In the first search (January 2005 to March 2020), three researchers removed duplicates and screened study abstracts divided by database (HB, MS, TC). One researcher then screened the full texts (MS) to determine eligibility in consultation with a second researcher when uncertainty existed (TC). This same researcher performed data extraction (MS). In the second search (April 2020 to October 2021), one researcher searched all data bases, removed duplicates and screened study abstracts (AD). Full text review was then conducted by the same researcher in consultation with a second researcher (AS). The second researcher then preformed all data extraction (AS). The second search was done to account for disruptions in the original study timeline caused by COVID-19. Figure 1 illustrates the flow chart for the study selection. Extracted details included study design, population (i.e., age, sex, ethnic or weight eligibility criteria), aim(s), methods, and findings (i.e., barriers and facilitators) and can be viewed in Table 2.

Fig. 1
figure 1

Flow diagram of included papers for review (n = 35)

Full size image
Table 2 Extracted information from all included studies (n = 35)
Full size table

Data analysis

Two researchers created the code book using inductive coding after reviewing all extracted themes from studies identified in search one (AD, CB). To do this, extracted themes from studies in search one were separated into three broad categories borrowed from SEM (i.e., individual, environment and intervention levels) [12]. Then, inductive codes borrowing language from published work were used to name sub-themes with the personal, community and intervention spheres. This included using terminology from behavior change taxonomies (BCTs) [62], behavior change theories (e.g., motivation, knowledge, attitudes) [11] and definitions of what and how components of interventions [20]. Sub-theme names were not deductively applied, but instead used to guide final decisions as all themes inductively emerged from the extracted data itself. Guidelines do not currently exist on how to consolidate emergent themes across studies. As a result we opted to rooted our analysis in behavioural science terminology as other work has suggested that this approach can help facilitate relevant understanding and application within the field of behavioural science [63]. After the codebook was established, it was independently applied by two researchers to all studies identified in search two (AD, AS). Discrepancies in theme names or coding of studies was triangulated with a third researcher (CB).

Quality assessment

The methodological quality of all papers was assessed using the Critical Appraisal Skills Programme (CASP) checklist for qualitative research [64, 65]. This appraisal tool consists of ten questions. In the appraisal, responses to each question were checked off as ‘yes’, ‘no’ or ‘can't tell.’ Afterwards, a ‘somehow’ group was added. An overall quality score for each article was then assigned as follows: 1. ‘Yes’ assigned one point; 2. ‘Somewhat’ assigned a half point; and 3. ‘No’ or ‘can’t tell’ assigned zero points for each of the 10 questions. The maximum achievable score was 10 points [64]. The methodological quality of all included studies was evaluated by teams of two researchers in each of the two searches (AD, AS). Discrepancies were discussed until consensus.

Results

Titles and abstracts were reviewed for 28,072 papers (25,162 search one, 2,910 search two). Seventy-eight duplicates (54 search one, 24 from search two) were removed. After screening, 156 papers were identified for full text review (129 search one, 27 search two) (Fig. 1). The eligibility criteria were applied resulting in a final sample of 35 papers (24 search one, 11 search two). Ten of these looked exclusively at diet-focused interventions [27,28,29,30,31,32,33,34,35,36]. Of these, one reported only on facilitators [30]. Fourteen studies exclusively at activity behaviors [37,38,39,40,41,42,43,44,45,46,47,48,49,50] with one reporting only facilitators [42] and one only barriers [46]. Eleven studies reported barriers and facilitators in mixed interventions [51,52,53,54,55,56,57,58,59,60,61].

Study designs

Semi-structured interviews (n = 22) [27,28,29, 32, 33, 35, 37, 39,40,41,42, 46, 51,52,53,54, 56,57,58,59,60,61], structured interviews (n = 1) [31], unspecified interviews (n = 2) [36, 48], focus groups (n = 8) [30, 34, 38, 43, 44, 47, 50, 55] and open written responses (n = 2) [45, 49] were used.

Study populations

Ages of included participants varied from 17 to 82 with an overall mean age of 49.7 years (diet: 50.9 years, activity: 49.1 years and mixed: 49.9 years) in 30 studies. Five studies did not present data to calculate a mean age [33, 39, 47, 52, 57]. Forty-six percent of studies were conducted in North America, four diet [29, 30, 34, 36], eight activity [37,38,39,40,41, 43, 47, 48] and four mixed [53,54,55, 57]; 31% in Europe, three diet [28, 31, 35], four activity [42, 44, 45, 49] and four mixed [51, 58,59,60]; 9% Australia, two diet focused [27, 32] and one mixed [61]; 9% East Asia, one activity [46] and two mixed focused [52, 56]; 3% Africa, one activity focused study [50] and 3% South America, one diet focused [33]. Seventy-four percent of studies (n = 26) included both males and females [27,28,29,30, 32, 33, 35, 38,39,40,41,42, 44, 45, 48, 49, 51,52,53, 55,56,57,58,59,60,61]. Of these, only 14% (n = 5) had more male participants than female and none of these were diet focused interventions. Twenty-six percent (n = 9) of studies recruited only females [31, 34, 36, 37, 43, 46, 47, 50, 54]. Sample sizes ranged from 6 to 117 participants (mean: 28.1). On average, ages in activity interventions were higher (diet: 22.3 years, activity: 35.1 years, and mixed: 24.6 years).

Review quality

The CASP scores ranged from 5.5 to 9.5. This indicates that included studies were primarily of moderate to high quality [64, 65]. The lowest scoring domain was recruitment. Many studies did not describe participant characteristics and only one study included details on saturation. Several studies also failed to include information on researcher-participant relationships.

Primary findings

Facilitators and barriers that arose across all three level were often an inverse of each other. For example, having a positive attitude was a facilitator whereas having a negative attitude was a barrier. At the individual level, three themes (attitudes, concern for health and physical changes) emerged. Four more at the environment level (social support, social accountability, changeable aspects of the community and unchangeable aspects of the community) and two at the intervention level (delivery and design and content) also arose. One additional theme called ‘fostering self-regulation through BCTs’ emerged as a facilitator only at the intervention level. A description of the emergent themes and their sub-themes from all studies can be found in Table 3. The remaining sections will discuss how themes differed across different intervention types (diet, activity and mixed). 

Table 3 Emerging facilitators and barriers participants felt impacted their ability to adhere to intervention guidelines for diet and/or activity behaviors
Full size table

Individual level

Individual attitudes played a large role in motivating behavior adherence. Participants talked about how their ‘desire for knowledge [27, 29, 30, 32,33,34, 41, 50, 52, 55,56,57,58,59, 61] (e.g., interest in learning or gaining knowledge), ‘positive initial mindset’ [29, 31, 33,34,35,36,37, 40,41,42,43,44,45, 47, 47,48,49,50, 53,54,55, 58] (e.g., optimism for changes or commitment to intervention goals) or experiencing ‘changes in self-perception’ [30, 33,34,35, 38, 40,41,42,43,44,45, 47, 48, 50, 52, 55,56,57,58, 61] (e.g., increased self-efficacy or feeling pride with achieving goals) had benefits on their perception of the intervention or its guidelines, which fostered adherence. In one study, having a positive attitude towards the intervention was conferred due to the notion that it was ‘pa[id] for’ [56]. Participants also discussed how their positive attitude towards an intervention was influenced by different desires in diet interventions compared to activity interventions. For example, in diet interventions participants focused on changes in their physical skills and abilities through wanting to gain knowledge of how to eat better or learning new food-related skills (e.g., new healthy recipes) [27, 29]. This contrasted motivation of participants in activity interventions, where the focus surrounded wanting to change aspects of they viewed themselves (i.e., their self-perception through changes in self-esteem [50, 52, 61], self-worth [33, 61] or mood [33, 34, 50, 52, 54, 56]) or how others viewed them (i.e., by forming an exercise identity [50, 56]).

Having a negative attitude towards the intervention or its guidelines hindered adherence. Participants revealed that ‘negative self-perception’ [29, 31, 33, 38, 43,44,45,46, 49, 51, 52, 58] (lacking self-efficacy or motivation to change behaviors), too many ‘competing priorities’ [29, 31, 33, 34, 38, 40, 43,44,45,46,47,48,49, 52, 54,55,56, 59, 60] (e.g., lack of time), ‘feeling overwhelmed’ [29, 31, 43, 47, 49, 51, 52, 56,57,58, 61] (e.g., previous failure in changing behaviors or not knowing how to start) or being ‘unwilling to change’ [29, 40, 43, 45, 48] (e.g., not interested) negatively impacted their attitudes and likelihood of adherence. In interventions with a diet component, ‘inner food cues’ like hunger, food cravings or emotional eating arose as a unique competing priority [27, 29, 31, 32, 34, 35, 51, 56, 61]. No equivalent discussion of feeling uncontrollable urges to be active or inactive were discussed in activity interventions. Other competing priorities common across all intervention types included ‘work outside of the home’ [29, 33, 38, 40, 44, 48, 52, 54, 55, 60], ‘work inside the home’ [27, 29, 33, 34, 38, 43, 46,47,48,49, 55] (e.g., caring for others) or unforeseen life events [27, 31, 46,47,48, 50, 61] (e.g., moving or travel).

Concern for ‘current health’ [27, 30, 35, 38, 40, 48, 50,51,52, 55, 56, 58,59,60,61] (e.g., positive changes) and ‘avoiding future diagnosis’ [27, 33, 42, 43, 48, 49, 53, 54, 58, 59] (e.g., chronic disease development) facilitated adherence to diet and/or activity interventions. As a barrier, health concerns for ‘physical state’ [27, 29,30,31, 33, 38,39,40, 45, 46, 48, 50, 51, 60] (e.g., illness or injury) or ‘feeling low’ [27, 29, 35, 43, 45, 47, 48, 52, 56, 60] (e.g., fatigue, low mood or depression) prevented behavior changes. No differences between the different types of interventions emerged.

Observations of physical changes signalled to participants that an intervention was working. This helped reinforce commitment to continue to pursue behavior guidelines. Observed changes in ‘body shape’ [29, 30, 32, 35, 38, 42,43,44,45, 47, 48, 50,51,52,53,54,55,56, 56, 59, 61] (e.g., weight loss or body image) or ‘brain-body connections’ [35, 38,39,40, 42,43,44,45, 48, 50, 55, 58,59,60] (e.g., feeling stronger or having more energy) acted as facilitators. As a barrier, participants exclusively talked about how failing to see changes in weight or having a ‘focus on weight’ often led to dissatisfaction in progress, hindering behavior maintenance [32,33,34,35, 44, 47, 52, 53, 56, 58]. Though discussed in all interventions, those with an activity component were more often mentioned to contribute to changes in brain-body connections through increased perceptions of physical well-being or abilities to preform activities (e.g., greater mobility or abilities to perform activity) compared to diet (e.g., increased energy levels only).

Environment level

Social support was the most frequently identified facilitator and barrier, talked about by participants in all interventions. This included support ‘within the intervention’ [28,29,30,31,32,33,34,35,36, 38, 39, 43, 44, 48, 49, 51,52,53,54,55,56, 58,59,60,61] (e.g., other participants, intervention staff and health experts like dietitians, trainers or doctors), ‘within the home’ [27, 31, 33, 34, 36,37,38,39,40,41, 43, 47,48,49, 51, 52, 55, 56, 59, 61] (e.g., family) and ‘outside of the home’ [30, 31, 34, 36,37,38,39,40, 43, 45, 47, 51, 56, 58, 59] (e.g., at work or with peers). In activity interventions having a pet or someone to by active with [41, 43, 45, 49] (e.g., co-participation) also supported behavior changes. No mention of having someone to preform dietary behaviors with like eating, cooking or grocery shopping was mentioned.

‘Social accountability’ also arose as a key influence. As a facilitator, feeling ‘participation guilt’ [28, 29, 35, 38, 39, 41, 44, 47, 50] (e.g., not wanting to let the research team down), wanting to ‘be a role model’ [27, 30, 35, 41, 45] or ‘change for others’ [34, 51, 53, 56, 58, 61] (e.g., feel accountable to change for family and friends) promoted diet or activity adherence. Contrasting this, ‘opposing norms’ or social contracts was talked about as a barrier [27, 28, 30, 31, 43, 47, 52, 55,56,57, 60]. In diet interventions opposing norms on the types or quantities of foods that are typically prepared (e.g., family or cultural norms), consumed with others (e.g., baked goods at coffee shops with friends) or a part of celebrations (e.g., holidays) were frequently discussed. Only two studies mentioned opposing norms in activity studies with both suggesting pressures to change to meet the typical convention surrounding body shape for women [43, 47].

‘Changeable’ (e.g., built environment and cost) and ‘unchangeable community aspects’ (e.g., weather) were mentioned mostly by participants regarding activity behaviors. As both a facilitator and a barrier, aspects of the built environment like access to nature [41, 45], nearby stores or recreation sites impacted activity practices [41, 44, 45]. No mention of infrastructure to cook or nearby food outlets was suggested by participants as a facilitator, but it did arise as a barrier [31, 57]. Cost was exclusively mentioned as a barrier in all types of interventions [27, 31, 37, 40, 57, 58, 61]. Finally, unchangeable aspects such as the weather to support outdoor activities, especially walking, was discussed only in activity studies [38, 39, 41, 44, 45, 48, 49].

Intervention level

Having ‘nearby locations,’ [33, 37, 58] ‘inclusive spaces’ [32] (e.g., stigma free), ‘flexible delivery routes’ [30, 33, 44, 50, 55, 58, 59] (e.g., timing of counselling sessions or incorporated some on site and at home components), ‘opportunities for social support’ [27,28,29,30, 33, 34, 38, 43, 44, 47, 48, 50, 53, 55, 56, 58, 60, 61] (e.g., with peers and with professionals) and ‘support after the intervention period ends’ [43, 53, 58] (e.g., post program info or resources to help with the transition to doing things without interventionists) were suggested intervention facilitators. Their inverse including ‘far away sites’ [31, 32, 38, 40, 54], ‘opportunities for stigma’ [54] (e.g., onsite weighing), ‘one size fits all’ [38, 39, 52,53,54, 57, 60, 61] (e.g., rigid structure or timing of sessions), ‘limited social engagement’ [31, 47] (e.g., with participants) and ‘intervention reliance’ [29, 34, 38, 39, 47, 51, 55, 60, 61] (e.g., delivery that was solely reliant on intervention for monitoring or spaces to facilitate behavior uptake) were reflected as barriers. In interventions where participants viewed interventionists as ‘experts’ (e.g., fitness instructors, diet specialists, research team members), adherence was facilitated [28, 30, 34, 49, 51, 53, 56, 58, 61]. This was largely attributed to perceived knowledge and their trusting the interventionists [27, 34, 35, 38, 43, 48]. It was also influenced by interventionists having a recognizable title (e.g., dietitian) [30], or through interactions where the interventionists modelled or provided corrections to an individual’s actions in physical activity interventions, for example [53, 55, 56, 58, 61]. Conversely, when participants did not feel that interventionists were knowledgeable, they did not perceive them as experts and were less open to follow their advice [51, 52, 55]. Comments surrounding inclusive spaces and opportunities for stigma was mentioned in a single diet focused study [54]. All other themes were consistent across intervention types.

Intervention content was viewed as a facilitator when it was ‘perceived credible’ ​[27, 34, 35, 51, 52] (e.g., trust information provided or who delivered it), had ‘clarity in messaging’ [31,32,33, 39, 52, 55, 61] (e.g., clear guidelines or goals), and allowed for ‘tailoring’ [27, 34, 56, 58] (e.g., flexibility in food choices or activity type). In diet focused interventions, content that included ‘lifestyle management’ [30, 34, 35, 55, 61] such as information on physical activity was also suggested to support adherence. No mention of dietary information supporting adherence for activity was mentioned in any studies. As barriers, ‘distrust’ [32, 51, 52], ‘lack of tailoring’ [27, 28, 31, 32, 35, 38, 55, 57, 61], ‘unclear messaging’ [37, 44, 54, 55] and ‘lack of activity information’ [34, 55, 61] arose.

Fostering self-regulation through incorporation of different BCTs in interventions was perceived by participants to have positive impacts on their adherence. This included ‘feedback and monitoring’ through self-monitoring [29, 30, 32, 35, 39, 41, 42, 45, 49, 53, 57, 58] (e.g., using pedometers, diaries) and professional monitoring [28, 34, 49, 51, 53, 56, 58, 61] (e.g., counselling support or check-ins). ‘Goals and planning’ through goal setting [41,42,43, 49, 51, 57], planning ahead [29, 34, 40, 58]. Demonstrations [30, 56] through grocery store tours or trainer demonstrates for activity) and forming habits [29, 33, 34, 38, 39, 41,42,43, 45, 49, 53, 55, 56, 58, 59, 61] through habitual practice of activities like walking or meal planning. Use of ‘tangible rewards’ to self-reward arose in one activity study [49]. All other themes were consistently stated regardless of intervention types. Presence or absence of BCTs was not recognized as a barrier.

Discussion

We explored in a systematic review of 35 lifestyle (diet and/or activity) interventions participant perceived facilitators and barriers to behavior adherence at the individual, environment and intervention level. By consolidating these similarities and differences across intervention types, our findings add to the literature by suggesting actions that interventionists can implement now to help overcome common barriers. This has the potential to improve intervention design or implementation and in turn may increase participant adherence, leading to improvements in health and wellbeing. A summary of relevant themes and there call to action for intervention developers can be found in Fig. 2.

Fig. 2
figure 2

Calls to action for interventionists based on common facilitators and barriers in lifestyle interventions at the individual, environment and intervention level 

Full size image

Individual level

At the individual level, common facilitators for both diet and activity interventions centred predominately around psychological factors (desire for knowledge, positive mindset, self-perception), self-regulatory skills (overcoming barriers with perceived lack of time or feeling overwhelmed) and observations of physical changes (weight, physical health or sense of wellbeing). This supports findings from a 2015 systematic review suggesting that psychological or self-regulatory skills and body image act as effective mediators of behavior change in lifestyle interventions [66]. However, our findings also revealed that focusing on weight can hinder behavior changes when expectations are not met.

Unrealistic expectations surrounding weight changes can lead to discouragement [67]. It also prevents an understanding of the importance of health behaviors on internal changes like blood pressure. For example, in one study in this review the authors found that participants empathized how positive changes in their body composition or energy were not motivating unless other people acknowledged them [47]. Counter-acting a focus on weight may be one strategy to help prevent discouragement with a lack of outward physical changes in lifestyle interventions [47, 68]. Other interventions within this review support this approach as changes in energy [35, 38, 42, 44, 45, 48, 50, 55, 59], confidence [30, 33,34,35, 40,41,42,43, 47, 52, 55, 56, 58, 60, 61], physical abilities [40, 42, 43, 45, 48, 55, 58, 60] or general health [38, 39, 44, 45, 50] acted as facilitators. This creates a twofold call to action for interventionists. First, a greater understanding of participants personal goals prior to intervention onset and second, increasing participants’ understanding that changes beyond weight are valuable [47, 68]. This matches findings from a recent systematic review looking at barriers and facilitators in 13 community-based physical activity interventions. In this study, researchers concluded that strategies such as negotiated planning and fostering individual buy-in are critical for intervention success [15]. Furthermore, community engagement strategies like those used by Bryne (2019) may provide one approach to better understand participant goals of a target group from the start to help shape outcomes in intervention design phases [69].

The most common barrier across all interventions was competing priorities and time constraints [27, 29, 31,32,33,34,35, 38, 40, 43,44,45,46,47,48,49,50,51, 54,55,56, 59,60,61]. Though not surprising, supporting participants to feel like diet or activity changes can be adopted as a part of a regular routine may facilitate behavior change [70]. This matches suggestions from participants in this review discussing how components in interventions that encouraged habit formation [29, 33, 38, 39, 41,42,43, 45, 49, 53, 55, 56, 58, 61] and incorporated planning [29, 34, 40, 58] were beneficial. Interventions should focus on helping participants find ways to easily incorporate guidelines into their schedule as feasible habits. This can be achieved through aspects like tailoring content [27, 34, 56, 58] or flexible intervention delivery (i.e., location or timing) [30, 33, 50, 55, 58, 59]. Interventionists could also take an approach of helping participants predict barriers that they may face to pre-emptively have strategies in place when anticipated barriers arise [15].

Environmental level

Participants in all interventions discussed the importance of social support from various entities. This is supported by other literature [71,72,73]. Interestingly, social support as a barrier looked different in diet compared to activity interventions. Lacking social support in dietary interventions was perceived to cause social isolation. For example, participants felt that they had to oppose typical norms surrounding eating in social settings like at the workplace, during holidays or social gatherings [28, 31, 52, 55, 56, 60] and when feeding other people [27, 28, 30, 31, 55, 57]. In contrast, in activity interventions, a lack of social support prevented action. For example, not having a companion for exercise [34, 43, 46, 47, 53, 56]. Based on these differences, targeted strategies based on targeted behavior type may be needed. Diet interventions could focus on breaking down discomfort opposing norms in eating with other people [74] or when cooking for others [75]. In activity interventions, social support strategies could utilize group activities [47, 50, 58, 60] or create opportunities for companionship (e.g., walking with other people or pets) [45, 49]. Helping participants identify solo activities that they enjoy is also likely important to prevent intervention reliance [76]. This could include strategies for indoor activities as well to avoid barriers suggested by participants regarding weather [38, 39, 41, 44, 45, 48, 49].

The built environment, including infrastructure, access to active spaces, healthy food and cost were common barriers across interventions [27, 31, 37, 40, 41, 44, 45, 57, 58, 61]. Though these aspects are difficult to address within interventions themselves, they are still noteworthy. Creating an ‘implementation plan’ may help reveal barriers at the community level that could hinder intervention implementation or adherence by participants [15, 77]. Context specific plans may help interventionists identify barriers and control for them in intervention design. For example, subsidizing or covering external costs associated with behaviour change like the cost of healthy food, gyms memberships or physical activity equipment. This has proven to be beneficial for compliance and consequently in improvement of intervention outcomes in low sodium diet [78,79,80,81,82] and activity interventions [83]. However, this type of approach can create intervention reliance [29, 34, 38, 39, 47, 51, 55, 60, 61] and should be evaluated for its potential to limit sustainable change once an intervention ends. Early identification of possible barriers outside of an individual’s control at the intervention level through an implementation plan may help shape design, helping overcome more systematic barriers from the start [15, 77].

Intervention level

A common theme at the interventional level was a lack of support for participants once an intervention ended [29, 34, 38, 39, 47, 51, 55, 60, 61]. This could have been attributed to participants discussion of reliance on intervention tools or experts for monitoring [28, 30, 32, 34, 49, 51, 56, 58, 61], or losing accountability to interventionists [28, 29, 35, 38, 39, 41, 44, 50]. This creates an environment where external motivation fuels behaviors [76] and can be problematic. Many participants talked about wanting to have regular interactions with ‘experts’ like activity trainers or dietitians in interventions to guide behavior change [28, 30, 33, 34, 38, 43, 48, 53, 55, 56, 58, 61]. It is worth noting however, that participants included within this systematic review and by others [15] quickly point out that when ‘experts’ in an intervention are not perceive as skilled, these interactions act as a barrier [51, 52, 55]. To overcome this, it is key for trained, credible interventionists to guide participants through behavior changes while fostering development of self-regulatory skills [50, 61]. One option to support this is by tapering off intervention supports. It could also include providing resources that can be used long after an intervention period has ended [28, 43, 58]. This approach has been suggested regarding intervention implementation (i.e., stepwise implementation) in a similar vein [15]. Complementary delivery routes that include mobile health (mHealth) may help satisfy this need as they can continue to be used autonomously by participants after an intervention period [84,85,86]. Future work should explore if the inclusion of mHealth can supplement traditional in-person interventions to better support participants once an intervention has ended.

Participants talked about how activity helped facilitate dietary behavior change, but not the other way around. It is not clear why this arose. Some literature has suggested positive benefits for more holistic interventions that focus on multiple behaviors [87]. Therefore, there is a need to consider if combining behaviors, including exploration of other health behaviours like sleep, has benefits on long-term behavior change compared to interventions that target one behavioral realm.

Finally, key facilitators and barriers among all interventions surrounded a lack of personalization to unique needs, goals, interests and schedules. This suggests that tailoring of interventions can support incorporation of new diet or activity behaviors. Individualized interventions where participants have bought into an intervention [15] have been shown to be more effective [88,89,90,91,92]. Goal setting [93,94,95] or self-monitoring [96,97,98] may be examples of effective BCT to help personalize interventions while prompting self-regulation [93,94,95, 99], as these BCTs have shown promise as facilitators of intervention adherence [27, 30, 32, 37, 39, 41,42,43, 45, 49, 51, 53, 56,57,58, 61, 96]. However, these BCT have their own set of challenges [93, 100] and research is needed to understand when different BCT offer merit. To do this, we echo calls for clear identification and classification of BCT within interventions first to facilitate greater exploration as to when different BCTs work [14].

Limitations

This review is not without limitations. First, the search strategy did not include targeted MeSH (medical subject heading) terms. This may have contributed to some studies being missed in each database depending on how they were indexed. However, it is more likely that a larger number of articles were returned and screened using this approach and instead, semi-related studies were reviewed (and excluded). The review protocol is also limited by having a single researcher lead full text review in consultation with a second researcher instead of having two independent researchers review all texts. Furthermore, caution needs to be present when interpreting these findings and making extrapolations to different sex, gender, age or cultural groups as the results above are largely representative of the opinions of females over the age of 40 years from North America. Many studies also ranked low in their quality of recruitment methods and did not outline if saturation was reached. This could imply that themes from studies with smaller samples in this review are not exhaustive. Lastly, though we rooted our analysis in a modified SEM (i.e., individual, environment and intervention levels), we did not acknowledge the nuances in how factors can act at multiple levels. For example, ‘cost’ can be rooted in preferences of what to spend money on (individual level), the cost of living (environment level) or failure of an intervention to provide certain supports (intervention level). In this paper, facilitators and barriers were viewed as mutually exclusive at one level, though they can be intertwined, which may have resulted in an oversimplification of the findings.

Conclusion

Incorporating strategies to mitigate barriers participants face within lifestyle interventions at the personal, environment and intervention levels may help promote behavior adherence. This includes: 1. Understanding participant unique goals and de-emphasizing weight-related outcomes; 2. Providing opportunities for diverse social companionship; 3. Anticipating personal and intervention level barriers in advance of intervention onset; 4. Preventing intervention reliance by fostering self-regulatory skills (i.e., rooting in BCT); and 5. Tapering off intervention supports. Greater adherence to intervention guidelines may support the uptake and maintenance of new diet or activity habits, supporting lifelong health.