To identify literature on WeChat-based obesity interventions up to April 14, 2024, we searched PubMed and China National Knowledge Infrastructure (CNKI, the largest academic database, aggregating more than 95% of officially published Chinese scholarly resources), using the following search terms: “WeChat,” “obesity,” “weight,” “BMI,” “waist circumference (WC),” “hip circumference (HC),” “waist-to-hip ratio (WHR),” “body fat (BF),” “skin fold thickness,” and Chinese equivalent characters: “weixin,” “feipang,” “tizhong,” “tizhongzhishu,” “yaowei,” “tunwei,” “yaotunbi,” “tizhi,” and “pizhehoudu” (Table S1 in Multimedia Appendix 1).

After eliminating duplicates and off-topic studies, all the papers that described WeChat-based obesity interventions were screened against the following inclusion criteria: (1) full-text original or empirical human research published in peer-reviewed journals (including trials with or without control groups, and blinded or nonblinded designs); (2) WeChat as primary intervention delivery platform (eg, dietary or PA guidance, dietary or PA diary logging, etc); (3) Chinese degree theses or dissertations from the CNKI dataset; and (4) objective anthropometric outcomes including body weight, BMI, WC, HC, WHR, body fat percentage (BFP), or skin fold thickness. The exclusion criteria included (1) studies using WeChat solely for participant recruitment or follow-up; (2) self-reported anthropometric parameters; and (3) study protocols, animal studies, reviews, patents, books, or inaccessible texts. This process was done by 2 authors (YNW and XXZ) who independently screened titles and abstracts and conducted full-text reviews to identify the final set of eligible papers for inclusion. Any conflicts during this process were resolved through team consensus discussions. The flow diagram was adapted from the official PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) template with modifications specific to our search strategy [43].

The following data were extracted from each study: authors, sample size, presence of control groups, WeChat functions used, interventions duration, outcomes measured, and intervention effectiveness. Intervention effectiveness was categorized as (1) “effective” if all predefined weight-related outcomes (eg, BMI, weight, WC, HC, WHC, BFP, etc) showed statistically significant improvements; (2) “partially effective” if only a subset of predefined outcomes demonstrated significant positive changes; and (3) “not” if none of the predefined outcomes demonstrated significant changes. Themes were synthesized from tile or abstract review followed by full-text review using a blinded, independent dual review process by 2 of the authors (YNW and XXZ). Conflicts were resolved through discussion, with a third author (JG) who would arbitrate unresolved conflicts.

A detailed quality assessment was conducted using the Quality Assessment Tool for Controlled Intervention Studies and Before-After (Pre-Post) Studies With No Control Group designed by National Institutes of Health (NIH) [44]. The assessment tool of controlled intervention studies is applicable to those with control group and intervention strategies, containing 14 items to evaluate the research. Before-after studies refer to those without control groups, instead comparing outcomes of baseline and the end. The tool for before-after studies consists of 12 items. All scoring criteria were evaluated as “yes,” “no,” “not applicable (NA),” or “not reported (NR).” For each item, we assigned 1 point for “yes” and “NA,” zero points for “no” and “NR.” According to the total score, studies were further rated as “good” (10 for controlled intervention studies; 9 for before-after studies), “fair” (5-9 for controlled intervention studies; 5-8 for before-after studies), and “poor” (zero to 4 for both). Two authors (YNW and XXZ) independently completed the quality assessment for each study. Conflicts were resolved via discussion, with a third author (JG) arbitrating unresolved cases.

Statistical analyses were performed using GraphPad Prism 8.4.0 (GraphPad). Descriptive statistics were applied, with categorical variables summarized as frequencies and percentages (n, %) to report study distribution across classifications, as well as the quality assessment outcomes of included studies.

From the initial 665 papers (67 from PubMed and 598 from CNKI), 12 duplicates and 58 Innovation Achievements of Chinese Science and Technology Project designs were excluded. During the first round of title or abstract review, 236 publications were excluded, in which 26 of them were nonscientific publications, 8 study protocols, 3 animal studies, 10 systematic reviews, 167 nonintervention studies, and 22 conference abstracts. Of the 359 studies undergoing full-text review, 315 (87.7%) were excluded due to the following reasons: focus beyond weight loss or obesity (170, 54.0%), WeChat not used for intervention delivery (45, 14.3%), inadequate WeChat intervention strategy details (8, 2.5%), pregnancy-focused studies (28, 8.9%), and self-reported or unverified measures (65, 20.6%). The flow diagram is shown in Figure 2. Eight (18.6%) studies [45-52] were in English with information about study population, control group presence, intervention methods, intervention duration, health outcomes, and intervention effectiveness summarized in Table 1. A total of 35 (35/43, 81.4%) studies [53-87] were in Chinese with study details shown in Table 2. Among the 43 studies finally included, 20 (46.5%) [53,57,59-63,65,70,72,74-76,79,81,83-87] of them were theses or dissertations. All the studies were conducted in mainland China. None of the papers reported using formative research for intervention development. Sample sizes ranged from 20 [64] to 15,310 [46]. Among the 43 papers, 33 (76.7%) [45-47,49-51,56-63,65-67,70-72,74-76,78-87] included a control group, and 26 (60.5%) [45,47,49-51,56-62,65,66,70,72,74,76,78-80,82-86] reported assigning participants randomly. The duration of interventions spanned 4 weeks [51] to 36 months [52].

Among the 43 papers ultimately included, 32 (74.4%) [46-48,50,53-57,59,62-64,66-77,80-82,84-87] used a single WeChat function for interventions, while 11 (25.6%) [45,49,51,52,58,60,61,65,78,79,83] used multiple functions of WeChat to implement their intervention studies. “WeChat group” and “Official Account” were 2 functions used the most, with 33 (76.7%) [45,49,51,52,54-61,63-70,72-79,81-83,85,86] and 11 (25.6%) [46,50,51,60-62,65,79,80,84,87] studies using them, respectively (Table 3). WeChat groups primarily facilitated communication between investigators and participants. Official Accounts were mainly used to deliver dietary or PA education and record daily diets and PA. Four (9.3%) studies [52,53,71,83] used Mini Program in their intervention strategies with similar but expanded functions as Official Account. Three (7.0%) studies [51,65,78] used one-to-one chatting functions enabling confidential consultations. In addition, 2 (4.7%) studies [58,79] used WeRun to conduct interventions. Meanwhile, another 2 used voice call [45] and live streaming [83] separately. A total of 4 (9.3%) studies did not clearly report the specific functions of WeChat they used in their intervention (refer to “Intervention” item in Tables 1 and 2 for more details).

In summary, 25 out of 43 (58.1%) studies [48,52-56,58,60,62,64-69,73-75,77-82,86] demonstrated significant improvement across all obesity-related measures. Of the studies remaining, 12 (27.9%) [45,46,49,50,59,63,70-72,83,84,87] showed partial effectiveness on the selected measures and 6 (14.0%) [47,51,57,61,78,85] reported no significant effects with a duration period no longer than 6 months. BMI was the most commonly used outcome measure for obesity, in which 27 (77.1%) [48,50,53,55,56,58,59,62-65,67,69-74,77-84,87] out of 35 [45,47-51,53,55-59,61-65,67,69-74,76-85,87] were significantly decreased after intervention. A total of 28 (65.1%) studies [45,46,48,49,52,54,55,58,61-66,68,69,71-73,75,77,79,80,83-87] assessed body weight, in which 23 (82.1%) [45,46,48,52,54,55,58,62-66,68,69,72,73,75,77,79,80,83,86,87] showed weight loss after intervention. A total of 30 (69.8%) studies [47,48,50,51,54,55,58,59,61-64,66,67,69-77,79,80,82-85,87] assessed WC as outcome, in which 18 (60.0%) [48,54,55,58,62-64,66,67,69,71,73-75,77,79,80,82] of them showed a significant reduction after intervention. For HC, a total of 14 (32.6%) studies [50,54,61-63,69,72,74,75,79,80,84,85,87] were considered it as outcome, in which 9 (64.3%) [54,62,69,74,75,79,80,84,87] showed significant decrease. Similarly, a total of 11 (25.6%) studies [48-50,60,61,63,65,71,80,81,87] assessed WHR as outcome, in which 7 (63.6%) [48,60,63,65,71,80,81] of them showed a significant decrease at the end of the study. Of the 14 (32.6%) studies [48-50,54,55,62,65,69,71-73,77,86] assessing BF, only 1 (7.1%) [55] of them used BF and all others used BFP. BF and BFP were shown to be impacted mostly by interventions compared with other measures in these studies because 12 (85.7%) [48,49,54,55,62,65,69,71,73,77,86] out of 14 [48-50,54,55,62,65,69,71-73,77,86] showed that they were significantly decreased (Tables 1 and 2).

Using the NIH Study Quality Assessment Tool, we assessed a total of 33 studies [45-47,49-51,56-63,65-67,70-72,74-76,78-87] that had a control group, with 4 (12.1%) [50,70,76,79] rated as good quality, 25 (75.8%) [45,47,49,51,56-63,65-67,71,74,78,81-87] as fair quality, and 4 (12.1%) [46,72,75,80] as poor (See Table S2 in Multimedia Appendix 1). The number of studies meeting each assessment item is shown in Table 4. Twenty-seven (81.8%) [45,47,49-51,56-63,65,66,70,72,74-76,78,79,82-86] of them were described as randomized, in which 16 (59.3%) [50,51,56-60,62,65,66,70,74,76,79,83,85] reported randomization methods that were regarded as adequate (see Table S2 in Multimedia Appendix 1, item 11 and item 22). Only 2 (6.1%) studies [50,76] reported using a double-blinded randomization process, and only 1 (3.0%) study [50] reported that outcome assessors were blinded. Among all the studies with control groups, 6 (18.2%) [49,72,78,80,82,86] did not report demographic information separately for intervention and control groups, and 4 (12.1%) studies [46,66,75,87] reported significant differences in some demographic or other factors related to the outcome being assessed (see Table S2 in Multimedia Appendix 1, item 6). Two (6.1%) studies [45,87] reported an overall dropout rate exceeding 20%, with exact values of 23.1% [45] and 28.4% [87] separately, and 4 (12.1%) studies [65,67,72,75] provided no information about dropout rate (see Table S2 in Multimedia Appendix 1, item 7). In addition, 3 studies reported high intervention adherence [47,72,86], while 3 others reported low adherence [46,61,87]. Besides, intent-to-treat analysis was used in data analysis in 3 studies to avoid bias caused by dropout [47,50,87].

A total of 10 out of 43 (23.3%) studies [48,52-55,64,68,69,73,77] used a before-after design without a control group. According to the rating, all of them [48,52-55,64,68,69,73,77] were rated as fair quality (Table S3 in Multimedia Appendix 1). The number of studies meeting each assessment item is shown in Table 5. All the before-after studies [48,52-55,64,68,69,73,77] stated research questions or objectives clearly and presented specific inclusion and exclusion criteria. Two (20.0%) studies [48,52] included participants who could represent the population who would be eligible, and 1 (10.0%) study [69] enrolled all the eligible participants meeting the entry criteria. Although some of the studies [48,53,54,64,73] provided information about sample size calculations and power to detect statistical differences, they claimed insufficient power as a limitation. All the studies [48,52-55,64,68,69,73,77] clearly and consistently described intervention strategies and used effective and objective outcome measurements. However, none reported whether blinding was used during intervention implementation or data analysis. All studies [48,52-55,64,68,69,73,77] prespecified outcome measures, which were applied clearly and consistently throughout the study process. Three studies reported a loss to follow-up rate exceeding 20%, with rates of 82.4% [52], 24.3% [53], and 55.5% [73], respectively. All studies [48,52-55,64,68,69,73,77] analyzed changes in outcomes of interest before and after the intervention and presented P values, but only 3 studies [52,64,68] measured outcomes multiple times before, during, and after the intervention. None of these studies was conducted at the group level; thus, it was not applicable to account for the use of individual-level data in determining group-level effects in the statistical analysis.

This is the first publication to describe intervention studies focused on weight loss or obesity control conducted via WeChat. We identified gaps in the current design and implementation of these digital intervention studies, such as inadequate tracking options and social support, lack of individualized goals, and absence of health care experts [33]. Overall, the results are promising, suggesting that WeChat may be an effective platform for supporting weight loss or obesity management. However, more rigorous, randomized controlled trials are needed to understand the effectiveness of interventions delivered using this platform, and to determine the best implementation strategies. Our research revealed several important findings.

None of the studies reported using formative research to design the intervention. Formative research is a process of collecting useful data about the targeted population and context before or during intervention implementation, during which behavioral change strategies are developed and refined to be maximally suitable [88]. This process helps with cultural contextualization and optimizes the impact of intervention strategies [89] by generating more effective interventions tailored to specific populations. For example, formative research on diabetes management and support indicated that intervention strategies for type 1 diabetes should be different from those for type 2 [90]. Thus, to better use WeChat in obesity intervention studies, formative research is imperative for developing further strategies to achieve weight management outcomes. Of the 6 studies [47,51,57,61,78,85] that reported no significant changes in obesity-related measures, all had a duration of less than 6 months. Intervention durations of 6-12 months and higher contact hours were correlated with more substantial changes [91,92]. Thus, future studies should consider extending intervention duration and enhance contact hours.

According to the assessment of studies using the Quality Assessment Tool designed by NIH [44], most studies had insufficient sample sizes. The intervention studies conducted in Chinese populations also had several common limitations. First, although most studies used randomization to assign participants to groups, blinding was rarely used (more than 95% did not report blinding during the intervention process). Similarly, information about participant adherence to interventions was limited, which impeded evaluation of effectiveness. In addition, our review of the inclusion and exclusion criteria found that most studies did not control for confounding effects caused by simultaneous participation in other intervention studies. These limitations might cast doubt on the generalizability of the findings.

WeChat, one of the most widely used social media platforms in China, has many functions that can facilitate obesity interventions [93,94]. In recent years, more intervention studies have used WeChat as a tool in China [41,93-95] because of its convenience and wide reach. The Official Account and WeChat group were the most common features used in the intervention trials included in this review. Both one-to-one chatting and WeChat groups facilitated immediate communication. Although this type of immediate communication might occupy researchers’ time, it can also provide quick assistance to participants, potentially reducing loss to follow-up and increasing adherence to interventions, which may lead to positive outcomes. Official Account and Mini Program are user-friendly ways to disseminate intervention-related information, such as health knowledge and PA courses, as well as to record and evaluate personal data on daily diet and PA. While these functions could help researchers deliver health-related information, they do not ensure that participants actually engage with this information.

Generally, WeChat platform shows promise for obesity interventions; however, its utilization for health promotion still needs to be optimized. For example, although researchers can send intervention materials to participants easily via WeChat, there is no guarantee that they will pay attention to the health messages received [75]. Thus, study coordinators should not only use WeChat to deliver health-related information but also try to improve participant engagement through compelling content and strategies to compete for attention with the vast amount of the information disseminated through digital platforms [75].

Our review has several limitations. First, methodological heterogeneity and variable reporting quality in Chinese language studies limited the direct comparability of outcomes, but we addressed this using narrative synthesis by study type (with or without a control group). Second, while the cultural specificity of WeChat-based interventions limits generalizability, this review lays important groundwork for cross-platform comparisons.

In summary, WeChat is a potential platform for delivering obesity interventions due to its wide use and various functions. Most of the reviewed papers reported effectiveness in terms of weight loss. However, intervention studies conducted via WeChat were limited in design and evaluation rigor. Therefore, future studies should include formative research for intervention development, use a randomized controlled blinded design, and use objective obesity outcome measures to increase the reliability of their findings. The main features of WeChat used in the included obesity interventions enabled chatting and information sharing without spatial limitation. Thus, further studies should explore multiple functions simultaneously to enhance participant engagement with health education materials, examine the impact of varying exposure levels, and increase participant engagement.