Main Findings
To the authors’ knowledge, this is the first attempt to systematically review the literature of weight management interventions using eHealth specifically in people from a low SES background and living with overweight and obesity. It is important because low SES individuals are disproportionately affected by overweight and obesity [13]. The main findings are that eHealth interventions specifically designed for low SES groups are scarce with only 4 low risk of bias studies meeting our inclusion criteria, comprising a total of 373 participants. eHealth interventions aiming to reduce/maintain weight or increase physical activity varied. They included interactive websites or voice responses, periodic communication and discourse via telephone, social media, text messaging, or eNewsletters. All studies reported a significant effect of their respective eHealth interventions on weight loss. Generalisations should be made with caution however as the review revealed only USA-centric studies with predominantly female participants. Given that SES spectra are not invariant across nation states nor equally distributed between biological sex [40], and overweight and obesity affects males more than females in the UK [6], future eHealth studies specific to the UK and that include both sexes are required.
Effect on Weight Loss
Intervention duration was relatively short (1–3 months, with one exception of 12 months and follow-up at 18 months), yet all interventions demonstrated statistically significant weight loss during the intervention. In the longer intervention case, the effect was sustained at 18 months [39]. There was a significant effect of interventions on physical activity which improved at 3 months in two cases [36, 37]. Collectively, these findings are welcome and supports the premise that eHealth interventions are a successful approach for people with low SES. Our findings are in keeping with an earlier narrative systematic review (6 studies, n = 4,899 [32]). It observed that eHealth weight management interventions had a positive effect on weight loss in participants who identified as being part of an ethnic minority group. Given that ethnic minorities are also associated with higher risk of deprivation and obesity [41], there is further evidence eHealth is an efficacious approach for vulnerable groups within the general population.
Although in our review we found interventions led to statistically significant weight loss, these findings need to be interpreted with respect to a clinically significant weight loss. According to UK clinical guidance, 3–5% body weight loss is associated with clinically meaningful health benefits [31] and aiming for 30% of participants achieving 5% weight loss is a desirable service outcome [42]. Two studies reviewed [37, 38] reported 16% and 32% of participants achieved ≥ 5% of body weight loss respectively meaning a minority of low SES participants achieved a clinically significant weight and one did not meet the UK national guidance. There is a need therefore to develop successful interventions to achieve clinically meaningful weight loss in a greater proportion of participants.
Effect on Physical Activity
There is no doubt that there are economic, social, and political influences that are driving change in the amount of engagement in physical activity and exercise at the population level, seeing as uptake of global recommendations (e.g. [43]) remains low [44]. No study reviewed assessed the effect on physical fitness which is presumably because physical fitness is defined as a sub-set of physical activity [45]. It might also be due to the recognition of attitudinal differences towards exercise compared with physical activity in people with long-term conditions [46, 47]. Irrespectively, physical activity increased significantly as an effect of eHealth programmes in two studies included in the current review [36, 37].
Since optimising physical activity and exercise as a behaviour change is desirable to support and maintain weight loss and reduces the risk of non-communicable diseases [48], evidencing eHealth’s effectiveness in increasing physical activity for low SES participants supports targeting physical activity in the design of interventions for this group.
People with low SES face specific barriers to sustained physical activity changes such as the cost of gym membership, perceived neighbourhood safety, and availability of green spaces to be physically active in [21, 49]. Efforts to modulate these barriers should be included in the design of interventions. Improving self-efficacy is a positive predictor of increasing physical activity in low SES groups [50]. So, it was welcome that self-efficacy was included within the eHealth interventions in the reviewed studies by provision of tailored physical activity feedback, pedometer self-monitoring, and setting physical activity goals [36, 37]. But it was disappointing that neither were able to report whether physical activity changes were sustained after 1 month [36] and 3 month [37] intervention periods. A previous systematic review and meta-analysis with low-income participants identified that while eHealth interventions resulted in a small but significant increase in physical activity levels, the effect was modest compared to interventions involving the general population, and it was not maintained at 6 months [51]. Evidence supporting the relative effect of eHealth on physical activity levels in low compared to higher SES participants, and whether any increases are sustained, therefore remains elusive.
eHealth Interventions and Media
The reviewed studies supported behaviour change through increasing self-monitoring behaviours (e.g. interactive voice response (IVR) and text messages) and information provision (e.g. social media posts and eNewsletters). Three studies provided equipment to support self-monitoring of physical activity [36–38]. One provided access to a gym with reimbursement of travel costs for follow up visits [39]. Weight loss outcomes in this study were compelling and sustained at 18 months which suggests that providing financial support could be a significant behavioural modifier given that absorbing travel costs is a specific barrier identified in low SES groups. Access to gyms, walking groups and community involvement are effective strategies to prevent weight gain in low SES groups [22]. Thus, it is no surprise that interventions that consider environmental, social, economic and/or structural issues are more likely to improve outcomes across SES. In the development of future interventions, clinicians, researchers, and funders have an obligation to consider factors associated with low SES, such as insufficient financial agency to purchase interventions and self-monitoring equipment. At a national level, financial support for sustained public health could be provided as part of welfare systems. There is debate whether the advanced welfare tax burden that egalitarian societies sustain offsets health inequalities due to socioeconomic status compared to more neoliberal welfare states [52]. Our belief is that the investigations into the causes for health inequalities should continue and are welcome because they will provide testable theories that can explain, for example, how physical activity improvements due to eHealth interventions wane differently depending on SES, and why. These may well indicate that provision of sustained financial support programmes for eHealth as a public health intervention is indicated for subgroups of society, and if so, programmes should be duly scrutinised for their cost-effectiveness.
eHealth has the potential to improve health at local, national, and international levels by using the developing technology effectively. Counterintuitively though, an expanding eHealth landscape could widen social health inequalities because not all individuals are able to use eHealth well due to inequity and inequality in environmental factors, access, cost, and utilisation [23]. Inequality exists in the dissemination of intervention results to the public too. Due respect to the spectrum of health literacy in the public to whose behaviours the results are aimed at modifying is not always made. What’s more, our results identify the scarcity of studies that included low SES participants. This potential bias is vexing because individuals with low SES are at a greater risk of social health inequalities. There is therefore a clear need to focus eHealth interventions tailored to this group.
The delivery method of eHealth should be an important factor when developing interventions due to differing utilisation of technology across SES. eHealth that is not accessible, easy to use and/or targeted to the population may further the digital divide [53]. Using smartphones as the only access to the internet is high among low income groups [54]. This means the use of mobile technology and applications may be more appropriate and acceptable in this population. While the interventions revealed in this review could all have been practically accessed using a smartphone, only one study was specifically designed for smartphone use via direct text messaging – a modality which incidentally caused the largest mean change in body weight loss [37]. Three studies did not specifically describe the use of smartphone use or accessibility despite the potential this has in this population. Utilising or adapting eHealth for smartphone compatibility should be supported because it is a strong candidate to improve the efficacy of interventions while minimising health inequalities among low SES groups [55].
Uptake and Attrition
Uptake and attrition is a key challenge in investigating weight management interventions in individuals with low SES due to the complex behaviour change required [56]. Attrition rates were generally low in included studies compared to traditional weight management interventions where attrition rates can be up to 80% [56]. Bennett, et al. [39] reported the lowest attrition rate (5%), presumably due to the strict exclusion criteria removing any participants that were suspected of being “uninterested”. Griffin, et al. [37] observed the highest attrition rate (48.5%) among participants who identified as African American, and participants with the lowest education and incomes. This suggests there may be sub-groups within low SES along ethnicity, education, and income demographics and presumably their intersections. Understanding the reasons for the demographic differences in completing programmes is an important area for further research.
Engaging sub-groups in the development of interventions, and understanding their specific needs is likely to improve retention of participants and outcomes. Barriers to participation in interventional studies are well documented [57]. In addition to experiencing significant time demands to attend and travel to study appointments, people with low SES report mistrust of, and poor communication with, health professionals [58]. However, eHealth has the potential to overcome some of these barriers because it can offset time and costs and provides autonomy in selecting to participate at convenient times.
Strengths and Limitations
This systematic review was registered with an international systematic review register which is one of its strengths. It has been written following the PRISMA guidelines [28] and the protocol has been previously published [29]. We do however acknowledge the review’s limitations. Almost all participants (99%) within the included studies were female. This is not uncommon within weight management intervention [24, 26] but this significantly limits the generalisability of results. Furthermore, this systematic review only included adults. Given that the burden of overweight and obesity is growing, there is a need to identify how eHealth can be utilised across the lifespan including younger populations who have different digital habits. Included studies generally had short follow up, with only one study investigating outcomes past 3 months. Weight loss maintenance is desirable and studies that include wash-out periods, or prolonged targeted support, with medium to long term follow up are urgently needed.