Participation in regular physical activity (PA) confers clinically significant improvements in musculoskeletal, functional, and mental health–related outcomes, with an extensive evidence base on maintaining energy balance, lowering the risk of cardiovascular and metabolic diseases, and improving overall quality of life (QoL) [1-4].

Midlife women undergoing menopause tend to have a more noticeable decline in PA levels, being more physically inactive than men across most countries [5,6]. In England, only 23% and 21% of women aged 45 to 54 years and 55 to 64 years, respectively, met the National Health Service aerobic and muscle-strengthening guidelines recommended for adults (aged 19-64 years) [7,8]. The UK National Health Service guidelines for PA recommend that adults aged 19 to 64 years take part in a minimum of 150 minutes of moderate to vigorous PA (MVPA), 75 minutes of vigorous activity per week, or an equivalent combination of both alongside muscle-strengthening activities (eg, body and weight lifting, yoga, and Pilates) twice a week [9]. Research suggests that a reduction in PA levels parallels the drop in estrogen during the menopausal transition, a factor that may contribute to decreased PA and the shift to more sedentary behavior among midlife women [10-12].

Midlife is also a period when the risk of chronic diseases increases, potentially because of the cumulative effects of unhealthy lifestyle behaviors [13] and, most directly, as a result of menopause-associated weight gain and increased risk of abdominal obesity [14-16]. During the menopause transition, women may experience an array of bothersome symptoms that may overlap or have a cascade effect, with hot flushes, night sweats, and vaginal dryness most frequently reported [17]. Other psychosocial and physical complaints include weight gain, sleep disturbances, mood swings, anxiety, fatigue, joint aches, sexual dysfunction, heart palpitations, and deterioration of QoL [18]. Increasing PA levels may reduce menopausal symptoms and improve QoL [19-22]. Evidence is currently mixed [23-25], but there are plausible biological mechanisms by which PA can alleviate vasomotor symptoms, for instance, by releasing neuroendocrine substances (eg, cortisol) that are involved in stress and thermoregulatory body responses. PA may also attenuate weight gain influenced by menopausal transition and aging, as well as other physical and psychological symptoms such as body pain, fatigue, poor sleep, and depression [5].

The use of mobile phone–based interventions may potentially encourage midlife women to increase PA. Mobile PA technology is defined as the use of wireless devices such as smartphones, tablets, wearable activity trackers (WATs), and PDAs to promote PA and provide a means for real-time monitoring [26,27]. Apps that run on mobile platforms typically form part of these interventions. In this review, we adopted an operating definition of mobile-based PA interventions by referring to the use of mobile app technology delivered through smartphones or WATs connected to partnering phone apps (eg, smartwatches or Fitbit) that can gather data and track progress remotely, with the aim of increasing PA participation in any form: aerobic (cardiovascular), resistance, endurance, or stretching exercise.

Compared with men, women are more likely to use smartphones and health apps daily [28], and 83% of adults aged 55 to 64 years owned a smartphone in 2021 [29]. Moreover, women may particularly favor mobile-based interventions that use flexible delivery modes as a motivator to overcome the risk of not allocating sufficient time to be physically active [30-32]. Unlike in-person training programs, mobile PA interventions may encourage women to overcome physical barriers (ie, lack of time because of multiple responsibilities [33-36]) and feelings of stigma, social discomfort, and self-consciousness linked with participation in group-based PA programs and gym attendance [35,37], for example, a fear of being judged for decreasing abilities [37].

The global market of PA apps was valued at US $1.1 billion in 2021, with a 46% increase since May 2020 in global downloads of fitness and health apps [38]. In 2017, there were >325,000 commercially available health and fitness apps on the market [39]; approximately 30% of them targeted PA [38]. Emerging evidence indicates the potential of these apps to promote PA uptake [40] even among older adults, contributing to healthy aging [41-43]. However, despite the popularity of PA apps, the published evidence of their effectiveness from recent systematic reviews in adults shows positive but mostly nonsignificant effects [44-47].

Incorporating behavior change techniques (BCTs) and theories in developing and implementing such mobile-based interventions is an essential ingredient to ensure their acceptability and effectiveness. Goal-Setting Theory and Social Cognitive Theory (SCT) argue that, for a behavior change to occur, goals should be specific, learning-orientated, attainable in the short term but sufficiently challenging, and linked to a longer-term goal [48,49]. Regardless, many PA apps on the market have limited BCTs, for example, the ability to be tailored to users’ needs and characteristics [40,50]. Recently, several content analyses have been conducted to determine the active ingredients of commercially available consumer-facing PA apps using the comprehensive BCT Taxonomy v1 (BCTTv1) [51]. Of the 93 BCTs in the taxonomy, Middelweerd et al [52] and Bondaronek et al [53] found that, on average, only 5 and 7 BCTs were used among 64 and 65 commercially available PA apps reviewed, respectively.

Furthermore, the key to successful digital behavior change interventions is potentially determined by the acceptability of the intervention and the level of motivation and user engagement [54,55]. Acceptability is a multifaceted construct that reflects how individuals consider an intervention to be appropriate based on anticipated and experienced responses to it.

Although the literature on the impact of mobile PA apps and WATs on adult and older adult populations is growing, to date, midlife women are largely neglected. In total, 2 pretest-posttest studies indicate that app- and web-based interventions may increase PA in this population [56,57] and may have advantages over conventional PA interventions [58]. However, to our knowledge, no review has synthesized current evidence on the contribution of mobile PA technology to changes in PA and menopause-related health outcomes among midlife women.

This mixed methods systematic review aimed to investigate and consolidate the existing evidence on the effectiveness, acceptability, and active behavior change components of mobile technologies for PA in midlife menopausal women. The following review questions were addressed: (1) How effective are mobile PA interventions in increasing PA levels in midlife women? (2) How effective are mobile PA interventions in improving menopause-related symptoms in midlife women? (3) How acceptable are mobile PA interventions for midlife women with menopausal symptoms? (4) Which BCTs are used across mobile PA interventions for midlife women during menopause?

A mixed methods systematic review of qualitative and quantitative studies was conducted following the 2020 PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [59]. The protocol was registered in PROSPERO (CRD42021273062).

Nine electronic databases—MEDLINE (Ovid), Embase, Scopus, CINAHL, Web of Science, SPORTDiscus, CENTRAL, PsycINFO, and the ProQuest Sports Medicine and Education Index—were systematically searched from January 1, 2007 (the year the first mobile app emerged on the market), to August 2021, updated in February 2022. Subsequently, a further forward and backward citation search and screening of reference lists of the included papers were used to detect any additional relevant studies. If the full text could not be found through searches, the corresponding authors of potentially relevant studies were contacted via email to request access to full-text papers or inquire about ongoing trial protocols.

The search strategy was developed and refined iteratively based on expert consultation with a systematic search librarian at University College London. The search strategy combined three key terms—“mobile digital interventions” AND “physical activity” AND “menopausal women”—including synonyms and components (eg, “mHealth,” “wearables,” “mobile apps,” “Fitbit,” and “smartwatch”). The search strategy was adapted for each database using tailored syntax, Boolean operators, and Medical Subject Heading terms. The full details of the search strategy can be found in Multimedia Appendix 1.

Systematic database searching was supplemented with gray literature searches using the Google Scholar and Google search engines. Search results were sorted by relevance, and the first 20 pages (approximately 200 results) were reviewed. However, no additional papers that met the eligibility criteria were identified through this process, and so gray literature was excluded.

The inclusion criteria were developed based on the Participant, Intervention, Control, and Outcome structure (Textbox 1). Studies of any design comprising quantitative (randomized, nonrandomized, and pretest-posttest studies), qualitative, and mixed methods primary research were all included. Studies that assessed the measurement properties or algorithm performance of digital interventions with no health or behavior change outcomes measured were excluded. Commentaries, conference abstracts, editorials, reviews, registered protocols with no results published, theses, books, and studies not providing an explicit research methodology were excluded.

After removing duplicates using EndNote (version 20; Clarivate Analytics) [63], the first reviewer (GS) screened all titles and abstracts in the first round and then reviewed the full text of potentially relevant or unclear articles against the eligibility criteria using Rayyan (Rayyan Systems, Inc) [64]. A second reviewer (HG) independently reviewed the first 20.83% (215/1032) of the retrieved records, alphabetically sorted by title, and tested them against the eligibility criteria. The percentage of agreement between the reviewers (GS and HG) was 92%, showing substantial interrater reliability (Multimedia Appendix 2). Disagreements between the reviewers were resolved through discussion and, where necessary, consultation with FH and RF.

GS and HG independently extracted data using an adapted data extraction form following the Cochrane Collaboration standardized data extraction templates for quantitative and qualitative studies [65]. The following data were extracted: study characteristics (publication year, authors, and country); study type and aims; participant characteristics and context (sample size, mean age, and menopause stage if available); a description of the interventions as recommended by the Template for Intervention Description and Replication checklist [66], including content, mode of delivery, features, duration, intensity, and theoretical contribution; outcomes measured on the overall effectiveness of mobile PA technology on any menopause-relevant outcomes and PA outcomes as well as the acceptability, user engagement, and adherence to the intervention; and control group treatment (if applicable). In the case of registered or ongoing trials and protocols, we attempted to contact the corresponding authors via email to seek additional unpublished information where applicable (4 were contacted and 2 responded).

Two authors (GS and HG) assessed the methodological quality of each included study independently using the Mixed Methods Appraisal Tool [67] and discussed their assessments to achieve consensus. In this review, we used a star rating system as the Mixed Methods Appraisal Tool has no established quality threshold for inclusion and classification of overall risk of bias [68]. Studies were rated as “low risk of bias” if they obtained stars in up to four domains and as “moderate-to-high risk of bias” when they were awarded stars on ≤3 domains. Studies were not excluded based on critical appraisal given the infancy of research in this area. However, studies with moderate to high risk of bias were reported with caution.

For the included studies with actual mobile PA technology (9/12, 75%), 2 reviewers trained in BCT coding (GS and TR) independently coded all PA interventions in both the intervention and (active) control groups using the BCTTv1 [51]. Published descriptions and supplementary materials, if available, were reviewed in full. All discrepancies between the 2 initial coders were resolved through discussion until agreement was achieved. If necessary, the third and fourth reviewers (FH and RF) were also consulted to mediate an agreement. The average number and type of BCTs used were mapped for each studied intervention.

Narrative synthesis following the guidelines by Popay et al [69] was used for this review because of the heterogeneity of interventions, populations, and outcomes measured. The approach by Popay et al [70] allows for transparency of narrative synthesis by interpreting evidence from different methodologies.

Quantitative data were tabulated, with textual descriptions applied to draw a preliminary synthesis of the findings. Qualitative data were coded inductively in NVivo (version 12; QSR International) using thematic synthesis [71], and analytical themes were generated. In this review, we drew on the technology acceptance model (TAM) [72] to guide the analysis of qualitative data. The TAM suggests that an individual’s intention to use technology is based on two key factors: perceived usefulness, which refers to a user’s beliefs that engaging with the app improves their PA performance, and perceived ease of use, which refers to the perception that using the app requires minimal effort [72]. Although the TAM assumes that acceptability does not change over the life cycle of a digital intervention, it is widely used and has been shown to be robust in several empirical studies [73].

A meta-analysis was conducted for randomized controlled trials (RCTs) only using RevMan (The Cochrane Collaboration) [74] where sufficient studies were available for an outcome. Pooling change scores within and between groups is not recommended [65]; therefore, pre-post studies were not meta-analyzed. Effect sizes were calculated using the absolute mean difference and associated 95% CI between the final values observed for the experimental and control groups. A random-effects model was used to allow for between-study variability. Heterogeneity was quantified using I². Owing to the small number of included studies, tests for asymmetry and publication bias could not be conducted.

The study selection process is summarized in Figure 1 using the PRISMA flow diagram. Of 1627 records identified in addition to 27 potentially relevant records, citation tracking, and reference list screening, 12 studies (0.73%) published in 14 papers were included in the final review synthesis [56,75-85]. Reasons for exclusion are presented in Multimedia Appendix 3, mainly the absence of mobile PA technology, followed by irrelevant age groups.

See Table 1 for characteristics of the included studies (N=12). The studies reflected cross-disciplinary research and different stages of intervention development and evaluation. The studies were conducted in the United States (6/12, 50%), Australia (2/12, 17%), South Korea (2/12, 17%), Italy (1/12, 8%), and Iran (1/12, 8%). In total, 75% (9/12) of the studies were published in the last 5 years.

The included studies were a mix of quantitative (7/12, 58%), qualitative (4/12, 33%), and mixed methods (1/12, 8%) studies. The sample sizes ranged from 8 [77] to 83 participants [83]. Of the 12 studies, 4 (33%) were pilot RCTs, of which 1 (25%) had an active control arm [78] and 3 (75%) had waitlist or no-intervention control groups [76,81,83]. In total, 25% (3/12) of the studies were pretest-posttest studies [56,79,84]. The quantitative study duration varied from 1 [76] to 6 months [83]. Qualitative studies (4/12, 33%) included a semistructured focus group (1/4, 25%), semistructured interviews (1/4, 25%), and participatory design (2/4, 50%).

The participants were midlife women with an average age of 57.6 (SD 4.026) years. Most of the included studies (10/12, 83%) recruited women based on age range, followed by menopause stage, with only 17% (2/12) of the studies [75,77] identifying participants based on the experience of menopausal symptoms. The studied women were culturally diverse; 17% (2/12) of the studies [56,81] targeted African American women, and 8% (1/12) targeted [85] Korean-Chinese migrants. The included participants were heterogeneous concerning health conditions and the presence of chronic diseases. A total of 42% (5/12) of the studies limited recruitment to inactive (ie, ≤60 minutes per week of MVPA) and overweight (mean BMI 29.2, SD 3.5 kg/m²) or obese (mean BMI 33.9, SD 5.9 kg/m²) women [56,78,79,82,83]. In total, 25% (3/12) of the studies were based on midlife women diagnosed with breast cancer [80,81,83], and 8% (1/12) recruited postmenopausal women from cardiology clinics [84].

The overall risk of bias was “Moderate to high” in 58% (7/12) of the included studies and “low” in 42% (5/12) of the studies. All RCT groups (4/12, 33%) were comparable at baseline, whereas randomization was adequately performed and sufficiently reported in 75% (3/4) of these studies. Owing to the nature of mobile PA interventions, participant and assessor blinding could not be achieved in any study. There was poor reporting of a WhatsApp-based intervention and PA outcomes [76].

None of the 25% (3/12) of pre-post studies [56,79,84] accounted for confounders in the design and analysis, reducing the confidence in the observed effects (poor quality overall with high risk of bias). The included pre-post studies (3/12, 25%) had very small sample sizes and low recruitment rates; for instance, Sengupta et al [84] recruited 10 midlife women, and only 8 completed the 12-week follow-up. Similarly, Joseph et al [56] reported a low recruitment rate of 22% with a small sample size of 20.

There was better reporting across the qualitative studies except for reflexivity and the authors’ positions. A lack of data reporting and integration was observed in the mixed methods study design by Kim et al [85]. The risk of bias scoring system is presented in Tables 2-5.

All 12 studies included at least one form of mobile-enabled PA intervention, either as solo mobile apps or web-based applications or paired apps with other sensor-based activity trackers (ie, wearables). The intervention components included in-person training and behavior modification sessions [79,81,83], traditional SMS text messaging or follow-up calls [78,81,83,85], and educational pamphlets [76,83].

See Table 6 for the characteristics of the intervention types, embedded BCTs, and outcomes measured for the interventional studies (9/12, 75%). In total, 78% (7/9) used wearable devices to track activity paired with an app, including Fitbit (4/9, 44%) [56,78-80] and Garmin (2/9, 22%) trackers [80,83] and a tailored smartwatch paired with the HerBeat app (1/9, 11%) [84]. All studies were based on apps designed to promote PA except for 11% (1/9) of the studies [76], which used a WhatsApp-based PA intervention. Control groups included a basic step-counting pedometer without feedback [78] and a waitlist [81,83].

The measures included were weight loss, sleep disturbance, mental health (perceived stress and depressive symptoms), and menopause-specific QoL.

The usability and acceptance of the interventions were examined quantitatively in 42% (5/12) of the studies using surveys [56,78,79,81,84] (Table 7). Acceptability was most frequently assessed in terms of satisfaction and the users’ experience of using mobile apps, WATs, or the overall program. There were high levels of satisfaction and acceptability, favoring the use of mobile apps and Fitbit activity trackers. For instance, Cadmus-Bertram et al [78] compared a Fitbit-based intervention with a traditional pedometer without feedback received in a control group and found significantly higher satisfaction levels in the Fitbit group—96% (24/25) rated Fitbit as “somewhat or very helpful” compared with only 32% (8/26) in the pedometer control group. Similarly, 67% (2/3) of the pre-post studies [56,79] found that midlife women favored Fitbit (combined use of the Fitbit app and activity tracker) and reported that using Fitbit was “motivational to PA.”

The perceived usefulness and ease of use, where measured, were often limited to whether users experienced technical issues associated with the use of mobile apps and activity trackers. Cadmus-Bertram et al [78] found that 80% (20/25) of midlife women had no technical difficulties with the Fitbit trackers and reported technical issues that were easy to resolve. Furthermore, participants reported that more hands-on training could improve their satisfaction and engagement with the app-based website functions. Sengupta et al [84] reported on the acceptability and usability of the HerBeat smartwatch and paired app. In this pilot study, midlife women with cardiovascular diseases found the app features to be easy to use but complained about the short battery life of the HerBeat smartwatch [84].

Across the included studies, no adverse events were reported by the participants themselves or by the research team to be related to the use of mobile apps or trackers.

We identified three main themes from 33% (4/12) of high-quality studies related to perceived usefulness, readiness to use, and ease of using mobile PA technologies [75,77,79,82]. A summary description of the themes with some corresponding excerpts can be found in Multimedia Appendix 4 [75,77,79,82].

The average number of BCTs per mobile PA intervention was 8.8 (range 4-13; Figure 3). BCTs were mapped out for all actual interventions included (9/12, 75%) and the one active control (1/12, 8%) in the study by Cadmus-Bertram et al [78] (Table 6).

Collectively, 22 different BCTs from 9 different clusters of the BCTTv1 were identified across the 75% (9/12) of coded interventions. A total of 8% (7/93) of the BCTs were used in more than half of the interventions—“Self-monitoring of behaviour” (8/9, 89%), “Biofeedback” (8/9, 89%), and “Goal setting (behaviour)” (7/9, 78%) were the most frequently described BCTs. “Problem solving,” “Social support (unspecified),” “Prompts and cues,” and “Information on health consequences” were identified in 67% (6/9) of the interventions each. All the included interventions (9/9, 100%) used at least one BCT from cluster 1 “Goals and planning” or cluster 2 “Feedback and monitoring” of the BCTTv1.

There were no clear patterns between the type and total number of BCTs used and the effectiveness of the included interventions. In total, 50% (2/4) of the RCTs [76,83] reported significant evidence on PA, sedentary time, sleep disturbance, and QoL outcomes and used a total of 10 and 8 BCTs. By contrast, the RCTs by Cadmus-Bertram et al [78] and Valle et al [81] used 6 and 12 BCTs, respectively, yet found no evidence of effectiveness. In the RCT by Cadmus-Bertram et al [78], the distinction in BCTs between the intervention (Fitbit-based) and active control (pedometer-based) groups was feedback on behavior, biofeedback, and review behavior goals, which were only present in the intervention group. Both groups involved the common BCTs “Goal setting (behaviour),” “Action planning,” and “Self-monitoring of behaviour” [78].

In total, 67% (2/3) of the pre-post studies [56,79] found some supporting evidence of significant changes in PA, sedentary time, and weight loss from baseline and used 10 and 8 BCTs. However, Sengupta et al [84] used 9 BCTs and reported no evidence of the effectiveness of mobile apps on PA, perceived stress, and depressive symptoms (see Multimedia Appendix 5 [56,76,78-81,83-85] for the full coding process of BCTs of each intervention studied).

This review of 12 studies found that mobile PA interventions in the form of stand-alone apps or WATs resulted in a small to moderate increase in objectively measured MVPA of approximately 61.36 minutes per week among midlife women, at least in the short term (≤16 weeks). However, precision decreases with a reduced sample size and, thus, the pooled effect size should be interpreted with caution. Although a meta-analysis was not possible for other menopause-related outcomes, moderate- to high-risk evidence suggests significant, positive effects on weight reduction, managing anxiety and sleep disturbance, and enhancing menopause-specific QoL domains in midlife women. Quantitative studies were mostly uncontrolled with small sample sizes. We also found from high-quality qualitative exploratory research that most midlife women perceived mobile technologies as acceptable and potentially helpful in motivating them to increase PA levels. Daily step count was seen as an acceptable and clear outcome to monitor.

To our knowledge, this is the first systematic review to synthesize the current evidence with regard to the effectiveness and acceptability of mobile apps and WATs targeting PA in midlife women. An increase of 61.36 minutes per week, at least in the short term, among both healthy and clinical midlife women is promising. PA apps and WATs tend to be effective in comparison with no intervention or traditional pedometer-based interventions with no mobile technology component. The estimated increase in weekly MVPA represents 40.9% (61/150) of the recommended weekly MVPA for adults aged 19 to 64 years [89]. However, certainty in the pooled effect estimate was downgraded because of small study bias. Although all other individual effect estimates favored mobile PA interventions, not all studies reached statistical significance.

This review was consistent with the recently published meta-analysis of 63 studies (N=8250 participants) of digital and mechanical wearables providing PA feedback showing a small pooled effect for MVPA equivalent to 48.5 minutes per week (95% CI 33.8-63.3) among adult populations [90]. Similarly, the meta-analysis by Laranjo et al [91] found that mobile apps and WATs caused small to moderate increases in PA (equivalent to 1850 steps per day) among healthy adult populations. The meta-analysis by Yerrakalva et al [92] also found a modest increase of 753 steps per day among older adults after using app-based interventions for ≤3 months. Owing to substantial physical inactivity among midlife women compared with the general adult population [5,7], even small increases in MVPA are likely to be beneficial. The effects of mobile PA interventions on sedentary behavior in this population as sedentary time were inconclusive, highlighting a need for more research to assess the impact of mobile apps on sedentary behavior outcomes in midlife women [79,83].

Few studies (6/12, 50%) evaluated menopause-related outcomes. We found mixed evidence of the effect on weight loss [78,79,81]. Single studies found positive effects on sleep disturbance [83], menopause-specific QoL [76], and depressive symptoms but not on perceived stress [84]. None of the included studies assessed the effect of mobile PA interventions on vasomotor symptoms such as hot flushes and night sweats.

In alignment with the key constructs of the TAM [72], findings from the qualitative synthesis suggest that perceived usefulness was grounded in women’s beliefs about the extent to which mobile apps or WATs could increase self-awareness of PA and improve the overall menopause experience by exchanging reliable health information and promoting behavior change. There was a tendency to favor a holistic approach when designing apps for midlife women by focusing on menopause as a whole experience and on lifestyle behaviors rather than on the limited functions of menopausal symptom trackers [77].

In this review, midlife women showed a desire for PA mobile technology that required only limited technical abilities and met their needs and preferences without imposing further burdens on their busy lives [77,80,82]. Echoing previous studies in older populations [93-96], women aged 40 to 64 years showed some hesitancy toward new technology and tended to be reluctant to take up mobile apps and WATs, possibly because of low confidence and SE in using new technology [82]. This may, in turn, reduce the effect of trackers on PA behavior change [97]. However, unlike a common reluctance to learn new technology among older populations [98,99], our review findings suggest that midlife women were willing to learn how to use apps to increase PA and make better lifestyle changes [77].

One of the key findings is related to which specific features of mobile apps and WATs were preferred by midlife women. Previous research suggests that effective PA apps for the general population might need some adaptations to meet the needs and requirements of each subgroup of that population [100]. The qualitative synthesis suggests that midlife women liked simple features, namely, step goal setting, activity monitoring, real-time feedback, easy-to-read content, and a user-friendly interface, with most midlife women considering step counting as the most favored feature [82,83]. We noticed that midlife women shared similar preferences related to functionality with older populations [95,99,101,102]. For instance, using large visual screens and readable text was perceived as helpful [91,93] and, thus, may facilitate PA mobile technology use among midlife women.

Furthermore, the application of the TAM suggests that providing technical support may facilitate uptake and engagement with new technologies [72]. Access to additional telephone or face-to-face technical support may increase midlife women’s confidence in technology, especially among those who were initially reluctant [78,80]. Hands-on training and easy-to-read manuals to guide the installation, synchronization of PA apps, and use of WATs were perceived as essential.

“Self-monitoring,” “biofeedback,” and “goal setting of PA behaviour” were the most frequently used BCTs across the included studies. These findings concur with previous literature on digital behavior change interventions targeting PA, which also highlighted the role of “social support” in adults [103-105] and older adults [106]. In this review, “social support” was used in 67% (6/9) of the studies targeting both healthy [56,76,79,85] and clinical (ie, survivors of breast cancer) [81,83] midlife women. Most recently, research suggests that midlife women are ready to make positive behavior changes, yet they need social support and connectivity [107,108]. Similarly, our qualitative synthesis found that midlife women would prefer an app that offers a safe space to share common experiences and receive social support [75].

However, because of the scarcity of existing evidence and heterogeneity in intervention type (eg, smartphone apps and WATs), multicomponent interventions (eg, in-person sessions, SMS text messaging, and follow-up calls), mode of delivery, and outcomes measured, it was not possible to ascertain which intervention components or BCTs were most effective in increasing PA or improving menopause-related outcomes. Reporting of interventions and mode of delivery in the included studies was insufficient; accordingly, we could not comment on the link between the described BCTs and mechanisms of action, a problem highlighted in similar reviews. Sediva et al [103] highlighted the relevant real concern of low treatment fidelity on the delivery of content as planned across the 13 included complex interventions, including PA apps. By contrast, without adequate information reported on measurements of fidelity or ensuring that the underpinning theory is reflected in the design and implementation process, implementation failure can potentially occur and, thus, the real-world effectiveness of such interventions must be considered with caution [109,110].

Interestingly, the top identified BCTs—“self-monitoring,” “goal setting,” and “biofeedback”—were in parallel with the most preferred app features perceived by midlife women according to the qualitative data synthesis. Hence, to optimize the effectiveness of mobile PA interventions in midlife women, it might be beneficial for future mobile interventions to take advantage of the simple features of step counts, goal setting, and real-time feedback and pair them with a sufficient number of BCTs. Evidence suggests that incorporating more BCTs is more effective than using limited or single BCTs to obtain significant effects on PA [106,111]. Further research is needed to determine which mobile PA components or active BCTs are the most effective in increasing PA in midlife women.

To our knowledge, this is the first systematic review evaluating the use of mobile PA interventions among midlife menopausal women. The main strengths of this review are the rigorous and inclusive methodological approach and the comprehensive and extensive literature search. The screening, data extraction, and risk-of-bias assessment processes were independently reviewed by a second researcher. BCT coding was also independently conducted by 2 trained researchers, with high agreement.

However, this review has certain limitations. It should be noted that the findings of this review were based on healthy and clinical midlife women with potentially chronic conditions as well as on survivors of breast cancer, which may reflect special needs and, thus, different perspectives toward using mobile apps and WATs. For example, Nguyen et al [80] highlighted that survivors of breast cancer may have higher motivation to exercise to prevent recurrence of cancer and, thus, to sustain the use of WATs than healthy women. In this review, we only identified 25% (1/4) of RCTs [78] that targeted middle-aged women from the general population. However, the results were relatively consistent across populations. Future research should focus on targeting midlife women from the general community to ensure the generalizability of the findings.

In total, 58% (7/12) of the studies were at moderate to high risk of bias, and the inclusion of lower-quality study designs (eg, pre-post studies) substantially increased the risk of bias. Moreover, the data extracted from most of the included quantitative studies did not adjust for covariates and, thus, the meta-analysis reflects an unadjusted effect. Owing to the inherent variability in mobile interventions, control groups (active or no intervention), and the methodological quality of the included studies, the effect sizes of the pooled estimate and from individual studies had large CIs and low precision. Hence, further studies with larger sample sizes are needed to measure outcomes in a consistent way as the current lack of significant results of individual studies could be attributed to either a lack of effect or underpowered studies.

Given the novelty of this research area and the scarcity of existing RCTs, the inclusion of different study designs provides valuable insights. A total of 25% (3/12) of the studies included in this review were exploratory, participatory design qualitative studies that involved midlife women to obtain their insights before developing innovative mobile solutions, which may enhance the relevance and uptake of such interventions by this population. However, further rigorous studies informed by relevant theoretical frameworks and best practices are essential to explore how midlife women and their subgroups can best participate in the design of mobile PA interventions. Furthermore, qualitative process evaluations of interventional studies might also be equipped to fill in the gaps regarding the experiences and engagement with mobile PA interventions among midlife women.

Few empirical studies looked at the effect of PA mobile technologies [19,20,23] on menopause-related outcomes. Menopause is a life transition affecting half of the population and, thus, an area requiring further research and innovative applications. Rigorous studies on menopause-related outcomes can then serve as a solid base for policies and interventions to support more inclusive workplaces for women. For instance, promoting mobile apps for PA may be a way that workplaces and health care settings could use as a scalable, cost-effective strategy to deal with menopausal symptoms.

The findings from this review suggest that mobile PA interventions in the form of apps and WATs are likely to be acceptable to midlife women and may potentially increase PA. Evidence was mixed for sedentary time and weight loss, with single studies suggesting positive improvements in sleep disturbance and menopause-specific QoL domains. The most frequently reported BCTs across the included studies were biofeedback, self-monitoring of behavior, and goal setting (behavior). The most acceptable components of PA apps were manual goal setting and step trackers, whereas activity trackers needed to be comfortable and attractive. Although the approach of using mobile PA apps in midlife women appears promising, larger, high-quality studies should address the lack of evidence on effectiveness, acceptability, and feasibility of using mobile PA apps to address menopause-related outcomes and, thus, encourage midlife women to seek support to manage menopausal symptoms.