Introduction

Obesity represents one of the most significant public health challenges of the twenty-first century [1]. In 2021, an estimated 64% of adults in the United Kingdom were living with excess weight, 26% of whom were living with obesity [2]. Excess weight can have a significant adverse effect on an individual’s health status and quality of life, associated with increased risk towards a plethora of long-term conditions and co-morbidities such as type 2 diabetes, cardiovascular disease, and some cancers [3,4,5]. Excess weight and obesity are reported as a leading cause of preventable mortality, placing a considerable strain on health systems and substantial associated economic costs [6,7,8].

The benefits of weight loss for individuals living with obesity have been extensively reported, ranging from individual health to global economic factors [9]. Ensuring effective strategies are established and implemented to support individuals to facilitate weight management is crucial. Weight management interventions incorporating specific behaviour change techniques have been associated with significantly greater weight loss than interventions where such techniques are absent; for example when coded to the refined behaviour change techniques taxonomy CALO-RE, in a univariate model, each additional technique used in the ‘comparison of behaviour’ domain (03: provide information about others’ approval; 04: provide normative information about others’ behaviour; 22: model or demonstrate the behaviour; and 28: facilitate social comparison) was associated with an additional −1.5 kg weight loss at 12 months [10]. Also, technique 22 ‘model or demonstrate the behaviour’ was associated with −2.7 kg greater weight loss at 12 months when controlling for the other three techniques in this domain [10]. Clinically significant weight loss (−5% of body weight) has beneficial impacts, however, weight regain is a common clinical concern, particularly when sessions are missed or the support ceases [11]. Consequently, engagement in, adherence to and subsequent behaviour change long-term is important for weight-loss maintenance [12]. However, evidence demonstrates a large diversity in approaches and content between weight management interventions, which can negatively affect engagement and lead to variations in the effectiveness [13, 14].

Remote digital health behaviour interventions for weight management have become increasingly popular, driven by increased access to and adoption of technology over recent years [15, 16]. Digital approaches to weight management have been shown to have comparable effectiveness to face-to-face delivery regarding weight reduction outcomes [17, 18]. The COVID-19 pandemic has provided a unique opportunity to deliver more cost-effective, convenient and scalable digital solutions [19, 20], particularly among underserved populations who have been disproportionately affected by obesity [21] and COVID-19 [22]. There are, however, increasing concerns that the movement towards digital solutions to weight loss may lead to digital exclusion and exacerbate pre-existing health inequalities [23], with ethnically diverse, older, and economically disadvantaged populations shown to be less likely to engage with remote interventions [24,25,26].

The engagement of digital behaviour change interventions has been conceptualised as both a behaviour (e.g., attendance dose, attendance frequency, attendance duration, depth of usage) and subjective experience, which draws on users’ cognitive and emotional states when interacting with the intervention (e.g., attention, interest and affect) [27]. The barriers and facilitators to engaging with remote weight management interventions remain unclear [23, 28, 29] and this is the focus of this review.

Methods

This systematic review used the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) checklist guidelines [30, 31]. The protocol is registered in PROSPERO [CRD42021253439] [32]. This review was guided by methods highlighted by the Cochrane Handbook for systematic reviews [33]. Ethical approval was granted by The University of Bedfordshire Institute for Sport and Physical Activity Research Ethics Committee (REF: 2021ISPAR009).

Inclusion criteria

The Population, Intervention, Comparison, Outcomes, and Study design (PICOS) framework was used to define the eligibility criteria [34]. The following criteria were selected for inclusion.

Population

Studies including Adults (>18 years old) living with excess weight using the adjusted BMI obesity category thresholds to be inclusive of all ethnic groups (BMI ≥ 27.5 kg/m2) [35] and/or providers of care to this population following the NICE guidelines.

Intervention

Studies that included a synchronous (real-time), interactive, remotely delivered weight management intervention.

Comparison

Studies with or without a comparison group.

Outcome(s)

Studies that reported on the barriers and facilitators to engage with a remotely delivered weight management intervention.

Study design

Mixed-methods, quantitative and qualitative study designs where barriers and facilitators to engaging in a synchronous, interactive, remotely delivered weight management intervention were measured/discussed.

Exclusion criteria

The following criteria were set for exclusion using the PICOS framework.

Population

Studies with participants <18 years of age or with a BMI < 27.5 kg/m2 were excluded.

Intervention

Studies that were not a health intervention with a weight management focus, alongside those that were solely autonomous or delivered only face-to-face, were excluded. Studies that reported predicting factors (such as demographic predictors) without specific barriers or facilitators were also excluded.

Comparison

No exclusion specification was set for this criteria.

Outcome(s)

Studies not specifically reporting barriers or facilitators to engagement in a remote weight management intervention

Study design

Studies not published in the English Language and systematic reviews, conference papers, protocols, case studies, opinion and editorial letters, and unpublished works were excluded.

Search strategy

A systematic literature search strategy of 12 electronic databases including grey literature (MEDLINE, PubMed, CINAHL, PsycINFO, ProQuest, Scopus, Web of Science, SPORTDiscus, ISRCTN, NHS Evidence, ClinicalTrials.gov, and CENTRAL) was formulated, agreed upon between all authors and conducted from inception up to and including October 2023. The search strategy was co-developed with the research team and an experienced subject specialist librarian was derived from key terms, their synonyms, and MeSH terms, which were combined using Boolean operators (AND, OR) for terms relating to the population, intervention, and outcomes (see Box 1). Results were then filtered to include only the English language. Reference lists of any relevant/included studies were also hand-searched and screened for inclusion.

Study selection and data extraction

Retrieved articles were exported into Rayyan software [36] with duplicates being identified and removed. Titles and abstracts were then evenly distributed between all authors and were double-screened for eligibility using the inclusion and exclusion criteria, with any discrepancies discussed and resolved by three authors (JMW, JV, and AMC) to reach a consensus. The full-text screening was conducted similarly, with discrepancies resolved by three authors (JMW, JV, and AMC). A PRISMA flow diagram was used to report the screening process [31] (see Fig. 1).

Fig. 1: PRISMA flow diagram for study inclusion.
figure 1

This figure illustrates the different phases of the systematic review, specifically the number of records identified, screened, assessed for eligibility, and included in the final analysis.

Full size image

A data extraction template was developed using Microsoft Excel, which included general information (authors, publication year), country, study characteristics (design, aims), participant characteristics (sample size, age, gender/sex, ethnicity, socio-economic status, education), intervention features (descriptions of the intervention including duration, setting, provider, delivery format and procedure using TIDieR guidelines [37]), outcomes (including; consent/attrition/drop out, barriers, facilitators) which was completed by one researcher (EC) and checked by JMW.

Theoretical underpinnings

The use of behaviour change theory has been strongly recommended when designing, delivering and evaluating health interventions, particularly when assessing barriers and facilitators. The COM-B [38] (Capability (physical and psychological), Opportunity (physical and social), Motivation (reflective and automatic)—Behaviour) model assesses individuals, capability, opportunity and motivation towards a target behaviour which occurs through an interactive system at the hub of the Behaviour Change Wheel (BCW [38]). Using the Theoretical Domains Framework (TDF), the COM-B constructs can be understood at a more granular level. Capability consists of influences such as functional ability, knowledge, skills, memory, attention, decision processes and behavioural regulation. Opportunity consists of influences such as environmental context and resources, physical time, finances, accessibility, social influences and support. Motivation consists of factors such as beliefs about capabilities, beliefs about consequences, optimism, social/professional role and identity, intentions, goals reinforced habits, and emotions.

Data synthesis and analysis

The main analysis focused on conceptualising the barriers and facilitators to engagement, with deeper understanding though a COM-B/TDF ‘behavioural analysis’. Narrative analysis was conducted through the employment of inductive thematic synthesis to iteratively extrapolate the description of barriers and facilitators to engagement [39]. The coded data was then collated and investigated for overarching themes, which were then deductively mapped to the COM-B model of behaviour change [38]. This was conducted by one author (EC) and then independently reviewed by two other authors (AMC, JMW), and corrected for inconsistencies. All reviewers are experienced in behaviour change research with expertise using the Behaviour Change Wheel [38].

Quality assessment

The Mixed Methods Appraisal Tool version 2018 (MMAT-v2018) [40] was utilised for its versatility of application to assess a multitude of study designs, including qualitative, randomised controlled, nonrandomized, quantitative descriptive, and mixed methods, being the only tool to include specific criteria for appraising the quality of integrated qualitative and quantitative, mixed-methods studies. This was independently applied by two authors (YP, JV), assessing study quality and risk of bias by answering “yes”, “no”, or “can’t tell” to two screening questions and five key questions based on the study’s methodological approach. Once completed, ratings were individually assessed by a third author (JMW) to resolve discrepancies and reach a consensus. All studies satisfied the two initial screening questions of the MMAT-V 2018 and were generally of high quality, with all studies satisfying more than 50% of the appropriate appraisal criteria. Very few studies had criteria deemed as not satisfied, and where it occurred, it mostly related to an underlying reason, such as not being possible due to being a feasibility or pilot study. An overall numerical MMAT-V 2018 score for each study was not calculated as the score does not highlight the problematic areas of the study in terms of quality [41]; instead, a more in-depth explanation of each specific criterion has been reported [40] (see Table 1).

Table 1 Study quality appraisals.
Full size table

Results

The search identified 36,561 titles and abstracts. After removing duplicates, the remaining 29,036 titles and abstracts were screened for potential inclusion. Following initial screening there were 656 full-texts reviewed, which identified a total of 39 studies [42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80] which met the inclusion criteria (see Fig. 1).

Study characteristics

Fifteen of the included studies (41%) were randomised controlled trials [44, 50, 56, 57, 59, 61,62,63,64,65,66,67, 74, 77, 80] with the other quantitative designs including non-randomised experimental [60, 69], pre-post [67, 75, 78, 79], cohort, cross-sectional [42, 43, 71] and feasibility/pilot studies [45, 47, 49, 52,53,54, 58]. Seven studies (19%) utilised only qualitative designs, including focus groups [70, 72], interviews [51, 55, 76], and a mixture of both [46, 48]. Seventy per cent (n = 27) of all studies were conducted in the United States [43,44,45, 47,48,49,50, 54, 56,57,58,59,60,61,62,63,64, 66, 67, 71,72,73, 75, 77,78,79,80] with the remaining undertaken in England [42, 65, 76], Australia [67, 70], Norway [46, 55], Scotland [52, 74], Wales [51], Finland [69] and Switzerland [53]. Within the studies, the intervention sample sizes ranged from 10 to 2818. The mean age ranged from 31.5 to 72.9 years. Intervention studies mostly included participants with a mean age between 40 and 49 years (n = 14) [42, 44, 56, 59, 62,63,64, 68, 69, 73,74,75, 77, 80] and 50 and 64 years (n = 12) [45, 50, 54, 57, 58, 60, 61, 65,66,67, 78] with a few studies targeting younger (mean age 20–40 years) [79] and older (mean age of >65 years) cohorts [47, 49, 52].

Participants’ ethnicity varied across studies; however, it is important to note that ethnicity was not reported for eleven (30%) of the articles included [46, 48, 52, 55,56,57, 63, 69, 70, 76]. Of the studies that did report ethnicity, only eight were identified as having an ethnically diverse sample, defined as 50% or more of the whole sample as being from an ethnic minority group [43, 50, 51, 58, 64, 68, 72, 77]. Most studies included all adults regardless of sex; however, several studies specifically targeted females [45, 67, 71, 79, 80] and males [65]. Ten studies reported employment status, with five of these revealing the sample had an employment rate of less than 70% [50, 51, 55, 67, 78]. Twenty-four studies reported educational attainment, with two studies having a sample where less than 20% had a college degree [50, 59].

The ‘remote’ intervention modes of delivery varied across all of the studies which included the use of individual/group-based telephone calls [50, 54, 57, 58, 60, 61, 67, 68, 73, 78], video-conference calls [42, 47,48,49, 51, 56, 59, 62, 63, 66, 69, 70, 80], interactive websites [52, 53, 63, 71, 72, 74,75,76, 81], mHealth smartphone applications [43, 44, 50, 66, 74, 77], social networking sites [45, 79], emails [62, 63, 65], online chat messaging/web-forums [55, 62,63,64,65] and personalised/interactive text messaging [61, 67, 68, 70, 77, 80]. Many studies delivered the intervention through a combination of modes such as telephone calls alongside personalised text messaging [61, 67, 68, 70]. Some studies delivered the intervention through mHealth smartphone applications with the addition of telephone calls [50], video-conference calls [66], text messaging [77] or interactive websites [74]. Sixteen (43%) studies incorporated a remote monitoring component as part of the intervention [44, 45, 47,48,49, 52, 54, 56, 59, 60, 66, 72, 74, 75, 77, 80], including wearable technology to self-monitor physical activity [44, 45, 47, 49, 50, 52, 59, 72, 75, 77] alongside platforms to record and self-monitor diet, weight [44, 54, 56, 66, 72, 75, 77] and ‘lifestyle’ behaviours [59]. Several studies included telehealth remote monitoring [48, 60] or WIFI-enabled smart scales [50, 61, 77, 80] which exchanged data synchronously to a third party for monitoring. All intervention and population characteristics are presented in Table 2.

Table 2 Study intervention characteristics.
Full size table

Engagement with a remote weight management intervention

Patterns related to the engagement with remote digital weight loss interventions, classified as the target behaviour for this study, were examined. This focused on the uptake (recruitment to the study/intervention), the duration of usage (retention/adherence) and interaction (automated recorded data including logins and data entry, attendance) with the intervention.

Several studies required participants to be eligible to have access to a high-speed internet connection [47, 59, 62,63,64, 72, 75, 76, 81, 82] and/or required them to have access to the technology that the intervention was being delivered through, which in the majority of cases was through smartphones [44, 50, 74, 77, 79] and/or laptop/computers (which in some instances required specific operating system requirements) [50, 62,63,64].

Most studies that reported a recruitment rate (n invited/n consent) achieved a consent rate of 40% and above. The consent rate was 80% for one study [59]; followed by six studies at 60–79% [42, 44, 45, 47, 52, 59] and seven studies at 40–59% [49, 54, 61, 64, 74, 78, 80]. One study, which was based on delivering weight management sessions via video conference, revealed that older participants and/or those from an ethnic minority group were significantly less likely to consent compared to white younger participants [42]. The most common reason for lack of consent was a lack of interest [45, 47, 49, 64, 73] or perceived need [67] to participate in a remote weight management programme. Inconvenience, alongside competing responsibilities [44, 47, 49], preference for face-to-face support [42, 51, 54], lack of digital skills or technology capability [42, 61], lack of access to the internet [42] or technology, [67] refusing to use their smartphone for the intervention [67] and anxiety about technology [49] were also shown to be barriers to consent. Four studies revealed a consent rate below 40% (0–19% [51, 65, 79], 20–39% [73]). The low levels of consent in these studies, where specified, mostly related to decline (no reason specified) and/or non-response [65, 79]. In one study, 37% declined as they opted for a dietitian-led weight management group through face-to-face sessions rather than synchronously using videoconference [51].

The majority of studies achieved high retention (n enrolled/n completed), with fourteen studies achieving a retention rate of 90% or above [45, 50, 52, 54, 59,60,61, 64, 68, 69, 75, 77,78,79] and four which achieved full retention (100%) [45, 54, 61, 77]. These studies all had interactive components; for example, two were delivered via telephone consultations with the ability to log and track behaviour [54, 61], one study was delivered in hybrid (face-to-face/online) format through a social network website [45], and the other involved a mHealth App alongside personalised text messaging [77]. The inclusion criteria on some studies required participants to have the ability and skills to use the technology, such as sharing data [59] and/or have the willingness to use the technology required as part of the study [68] or already using the platform the intervention was being delivered through (i.e., social networking site) [79]. Where studies used a pre-requisite to either have familiarity or current use of the technology the intervention was being delivered by, there was a trend towards higher retention, with all studies achieving a retention rate of >90%.

Only one study had a retention rate below the 70% threshold [44] (63%). This study had a predominantly ‘middle-aged’ (M = 44.9; SD = 11.1) female (78%) sample with high levels of ethnic diversity (49% African-American). The delivered intervention included intensive biweekly diet and exercise counselling sessions delivered by a nutritionist coach (months 2–6) alongside a weight loss mHealth app, which provided real-time feedback, self-monitored food intake and activity, and tracked progress with weekly weigh-ins encouraged with opportunities for social networking and support. Tracking data identified that intervention usage was highest in the intensive counselling plus smartphone group, with 70% of all sessions attended compared to intensive counselling only (58%).

Behavioural diagnosis: COM-B analysis of influences on engagement with remote weight management technologies

The findings revealed a wide range of influences on engagement with remote weight management interventions. When mapped to the COM-B model [32], all constructs in the system were identified via themes related to: psychological capability (n = 9); physical capability (n = 3); reflective motivation (n = 19); automatic motivation (n = 11); physical opportunity (n = 8); and social opportunity (n = 11). The schematic framework is depicted in Table 3. This is explained in more detail below.

Table 3 Barriers and facilitators to engagement with remote weight management interventions, deductively mapped to the TDF domains and COM-B constructs.
Full size table

Physical capability (TDF domain; skills)

Physical and/or functional limitations [71] were presented as a barrier which would impact on engagement with a digital intervention in a cross-sectional survey with women who had completed cancer treatment. Findings from a qualitative study with older adults (aged ≥65 years) identified that sensory limitations (e.g., hearing, vision) [48] would impact their ability to engage with a remote intervention.

Psychological capability (TDF domains: knowledge; cognitive and interpersonal skills; memory, attention and decision processes; behavioural regulation)

The knowledge that comes from familiarity and prior experience [43, 44, 50, 51, 61, 71, 74] alongside having the digital competency and technical skills to use the technology [42, 45, 48, 60, 66] were revealed to be important factors that influenced engagement with a remote intervention. Having the acquired knowledge and skills on how to perform a health behaviour [51, 59, 71] and the knowledge and skills on how to set goals and regulate behaviour [74, 78] were also identified as important. Facilitators also included self-monitoring and tracking behaviour [44, 46, 52, 59, 65, 71, 72, 76], shaping knowledge [46] and receiving biological [46] and tailored feedback [46, 59, 67, 70, 73]. Conversely, cognitive limitations (memory loss and forgetfulness) [48, 74], particularly among older populations, negatively impacted how some participants engaged with a remote intervention. Literacy ability (reading and writing) or lack of was shown to influence the use of a remote intervention, reported in one study [55]. In this study, the majority of the sample were only educated to high school level (61%) with around one-third identified as being unemployed.

Reflective motivation (TDF domains: beliefs about capabilities; beliefs about consequences; optimism; social/professional role and identity; intentions; goals)

As with traditionally delivered interventions, the desire and motivation to lose weight [43, 44, 53, 71] alongside the perceived confidence in the capability to change and adapt health behaviour [46, 51, 71, 74, 78] were all found to be important facilitators that influenced engagement of a digital intervention. Some participants declined an intervention as they had a preference for a face-to-face mode of delivery [42, 51] rather than using a remote method. This was particularly noted in studies that included weight management programmes which have been traditionally delivered in person, and/or they were given the option to attend in person. Perceptions related to the ease of use and simplicity [44, 48, 49, 60, 65, 66, 70, 72, 74, 76], perceived usefulness of technology and/or intervention [47,48,49, 58, 60, 72], perceived capability to use technology [61, 62] and the willingness to learn or adopt the technology [44, 74] were all shown to be important facilitators to engage with a remotely delivered weight loss intervention. Perceived confidence to interact with others using technology [61, 67, 71, 74] was also highlighted as being important. This was particularly noted in interventional studies which used group-based interactive elements (video-conferencing, online forums).

Autonomy to choose the platform and technology that the intervention was delivered on [72] and role and identity [46] were also viewed as a facilitator alongside setting goals and planning [46, 59], having motivational discussions [46, 59, 68], and increased autonomy and personal responsibility [51]. In contrast, several studies revealed that the perceived time burden to engage in the intervention [45, 59, 65, 71, 78] was a barrier to continued engagement; however, this mostly related to interventional studies which requested participants to use daily wearable technology alongside frequent self-monitoring of behaviour. Perceptions related to concerns surrounding the privacy of the technology [55, 57, 71], unrealistic expectations [53] and perceived performance or effort expectancy [72] that was involved in taking part in the intervention were shown to negatively impact engagement.

Automatic motivation (TDF domains: reinforcement; emotions)

Emotional factors, particularly anxiety were revealed as a barrier to engagement in remote interventions. This related to embarrassment or anxiety of self-disclosure, [45, 48, 55] alongside anxiety using technology [49]. In contrast, some participants found remote interventions a reduced threat (less stressful and daunting) when compared to face-to-face interactions [51]. One study required participants to approach others to be their ‘peer’ helper which some participants revealed they felt uncomfortable doing [74]. With facilitators, praise (social reward) [46], rewards (material incentives) [46, 53], and natural consequences [46] were all shown to positively reinforce and influence engagement. User dissatisfaction with technology/intervention [60, 66], fear of failure [53] and isolation following the end of the intervention [68] were all shown to be barriers.

Physical opportunity (TDF domains: environmental context and resources)

The convenience of engaging in the intervention remotely, with less need to travel and less impact on competing demands (e.g. childcare, work) [45, 48, 51, 52, 56,57,58,59, 62, 65, 67, 70, 71, 76, 78, 79] alongside reduced cost compared with face-to-face [51, 56, 57] were viewed as the most common facilitators that influenced participants desire to engage with a remote weight-based intervention. Having good levels of functionality and usability of equipment [49, 60, 65, 72, 74, 76, 79], including the opportunity to track, monitor and share behaviour [44, 71, 74, 76] were important facilitators. Technical issues (malfunctions, unresponsive, errors, incompatibility) were revealed to be important barriers that influenced continued engagement in some of the interventions [51, 59, 60, 66, 68, 69, 74]. Many of the interventions required accessing or using specific technology devices or platforms which required particular specifications [48, 49, 51, 52, 56, 60, 65, 67, 72, 76], alongside a high-speed internet connection [43, 48, 51, 61, 65, 66]. Therefore, providing access to these was shown to be integral to consenting to take part in the intervention.

Social opportunity (TDF domain: social influences)

General social support and engagement was identified as one of the most common facilitators [45, 46, 48, 51,52,53, 55, 57, 59, 61, 70, 71, 73,74,75,76, 78], including peer [45, 47, 51, 62, 69, 74] and group support and encouragement [51, 63, 74, 77, 80]. Several studies reported that positive social comparison with peers/others was also viewed as a facilitator for remote interventions [51, 71]. A common facilitator identified among nine studies to engaging with remote interventions was increased accountability [44, 48, 57, 59, 60, 62, 67, 70, 72].

Encouragement and support by a professional or specialist (weight loss/exercise) [43, 44, 46, 49, 53,54,55, 65, 68, 74, 78, 80] was found to be particularly important. However, a survey conducted with primary care patients living with obesity from an ethnically diverse background viewed encouragement and support with healthcare providers [43] less favourably. Maintaining a positive relationship with the professional [54, 65, 72, 80] alongside the provision of non-judgemental care [70, 72] and emotional support [67, 73] were viewed as important. Hybrid interventions, i.e., those that were delivered remotely but had some face-to-face contact, were also viewed more favourably [42, 43, 47, 54, 60,61,62, 64, 65, 67], alongside interventions with regular check-ins [44, 53, 54, 58, 60, 65, 67, 72, 74, 78] and follow up support when the intervention has ended were also viewed important [68, 74, 78].

Discussion

From the 39 articles included in this review, 57 themes were coded; representing 18 barriers and 39 facilitators to remote engagement with weight management interventions. These were then mapped to all six constructs of the COM-B model to provide a detailed ‘behavioural diagnosis’ of areas to optimise and address in future research and practice.

For behaviour to occur, individuals must have the knowledge and skills to engage in or refrain from the desired behaviour [83]. Not having the digital skills to partake in a remote intervention was often noted as a reason for disengagement. In relation to this, even when interventions provided participants with the technology to access the sessions, individuals were still sometimes reluctant to engage due to not knowing how to use the technology [49], which also caused anxiety. Concerns surrounding privacy were also highlighted as a barrier to engagement [55, 57, 71]. Research in other topic areas with remote delivery has also identified that technology anxiety is a barrier to participation when using such methods [84,85,86,87]. Therefore, increasing individuals’ knowledge and skills in using technology, including privacy assurances, before an intervention is imperative if remote delivery methods are to be used hereafter and reduce feelings of anxiety. A related facilitator was participants being familiar with the technology, thus reinforcing the need to ensure individuals have the correct level of digital skills required to engage. This facilitator has been highlighted in previous research on remote engagement with interventions as being a crucial component to not only engagement but also intervention satisfaction, and subsequently adherence [88,89,90]. Increasing skills through education and training would also align with the finding that those with a higher education had greater engagement, and those with a lower education had greater disengagement. This is supported by a range of previous research which highlights a positive association between education level and digital literacy in studies that have examined remote health interventions [91, 92]. Previous research has displayed that lower digital literacy skills have also been positively associated with older age, which is another barrier this review highlighted to engaging in remote weight management interventions [93, 94].

The results of this review also suggest that remote engagement enables individuals with functional or mobility issues to access services which they would usually be unable to. This corresponds with the results of a multitude of other studies examining remote interventions in health [95,96,97,98]. It is noteworthy that individuals with audio or visual impairments exhibited limited physical capability to engagement. This particularly becomes a barrier for those with visual impairments when there may be certain visual cues from peers or professionals that are not being received via remote intervention. Previous research that was conducted during COVID-19 among individuals with visual impairments suggested that some visual cues were missed and more detailed descriptions were required for any infographics used, thus making communication more difficult for both provider and participant [99]. Opposingly, other research has indicated that some video platforms can support individuals with visual impairments in real-time through methods such as screen readers [100, 101]. Henceforth, the mode of delivery should be carefully considered by intervention providers to ensure appropriate support for individuals with physical capability limitations.

The social opportunity provided, allowing connectedness with peers was the most frequently stated facilitator to engagement. It was also observed that a lack of social support and peer-connectedness was a barrier to engagement. This aligns with other research in the areas of health interventions that indicates increased social support has a positive effect on engagement and adherence [90, 102]. Not only this, but research also suggests that social support within health interventions improves the outcomes targeted by the intervention [103, 104].

The results of this review highlight the importance of recognising health inequalities across ethnic minority groups. Relative to Caucasians, research has evidenced the health inequalities of individuals from ethnic minority groups, with black adults in particular being at greater risk of obesity [105]. Many studies demonstrate the disproportionate effects of overweight and obesity among ethnic minority groups, even though there is evident underrepresentation of ethnic minority groups within weight management studies [106, 107]. It is noteworthy in this study that 30% of the included papers did not report ethnicity; and of those that did report ethnicity, only 25% were identified as having an ethnically diverse sample (≥50% from an ethnic minority group). An inadequate representation of individuals from ethnic minority groups has been demonstrated in previous reviews, with only 33% from ethnic minority groups (18% African American, 9% Hispanic/Latino, 5% Asian and 1% Native American) and 8% categorised as “Other” [108]. Thus, future research should ensure to not only report the ethnicity of the sample but to include a more ethnically diverse and representative sample of the research population. Regarding effectiveness, a related systematic review [109] has demonstrated that eHealth approaches display feasible short-term efficacy with modest weight loss in comparison to control groups for those from ethnic minority groups, albeit that this was from a small study collation (N = 6). Accordingly, even greater emphasis should be placed on the inclusion of an ethnically diverse sample in future research.

A common barrier to remote interventions related to both motivation and the physical opportunity to engage, linked to digital exclusion. The mitigation of this has been previously proposed through the use of assessing digital literacy and providing face-to-face demonstrations before the intervention, increasing confidence to use technology platforms, utilising user-friendly platforms, and increasing access to good internet connection, software and technology [42, 85].

Strengths and limitations

This is the first known study to systematically review the barriers and facilitators that influence adults engaging in a remote weight management intervention, drawing on a COM-B behavioural analysis. This review consists of a large sample size with several methodological approaches, therefore increasing the generalisability of the study results. This study accounted for the adjusted BMI for ethnic minorities by including studies with a BMI ≥ 27.5 kg/m2. It included interventions provided in many countries, using numerous delivery techniques and to a wide-ranging sample. However, the findings from this review should be interpreted with the following limitations in mind. Studies that reported demographic predictors which in some cases could be considered either a barrier or facilitator to engagement were excluded. Minimal grey literature was found and, it was beyond the scope of this review to determine relationships between influences and intervention effectiveness, therefore changes in BMI have not been reported. Another limitation, due to the focus of the review being barriers and facilitators to uptake and engagement, this study did not extract the impacts of engagement and retention on weight loss, or the potential implications from anti-obesity medications.

Conclusion

This review has provided a novel contribution to the literature to help understand factors that influence adults’ engagement with remote weight management interventions, addressing a gap in the literature [110]. It has provided a synthesis of studies in this area, the majority of which were deemed high quality, and furthered the interpretation of findings by mapping thematic influences to the COM-B model [38] of behaviour change. This, in turn, offers insight that can be used by intervention designers to optimise engagement in future weight management interventions. Alongside this, it is recommended that when developing, delivering or evaluating future weight management interventions, the barriers presented from the COM-B analysis in this review should be considered and mitigated where possible to lack of engagement and health inequalities. Furthermore, it is important to optimise ‘what works’, therefore, future interventions should also utilise as many of the facilitators identified in this paper as possible to provide an evidence-based foundation for remote weight management interventions. As weight loss itself was not a focus of this review, and therefore not reported in some of the papers included due to their study design (e.g. qualitative), future research should consider the effects of intervention retention on weight loss. Future research should also investigate the effects of anti-obesity medications on engagement and retention to such interventions.