The scoping review was conducted in accordance with the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) [29,30]. To develop clearly defined research questions, the process created by the Joanna Briggs Institute was followed [31].

The following electronic bibliographic databases were searched: MEDLINE, CINAHL, Scopus, Web of Science, and PEDro, along with CENTRAL. Studies were also identified from reference searching of systematic reviews. The search dates ranged from 2006 to January 2023. Studies before the year 2006 were excluded because mobile technology was unable to support mobile apps and the concept of self-management had not been widely used before this date [32].

Search terms followed the “patient or population, intervention, comparison, and outcomes” model and included the following terms: “women,” “mobile apps,” “mHealth,” “pelvic floor muscle,” “kegel,” “compliance,” and “adherence.” Multimedia Appendix 1 contains a list of all the search terms. A title and abstract search was conducted in January 2023. The full search strategy is presented in Multimedia Appendix 2. The a priori protocol was developed and published on PROSPERO (International Prospective Register of Systematic Reviews) [33]. The “patient or population, intervention, comparison, and outcomes” eligibility criteria for this study are presented in Textbox 1.

The exclusion criteria were as follows: nonexperimental quantitative studies, qualitative studies, studies involving men, case studies, web-based interventions, and adolescent women aged 18 years or younger. Web-based interventions were excluded because of the technological advances in phones to keep up with user activity. Only studies written in English were included. Unpublished studies were not sought. For the review, adherence outcomes were self-reported exercise adherence, in-app adherence measures, and self-efficacy. There was no restriction on what measurement tools could be used to measure adherence. Studies were excluded if they did not measure these concepts.

The integrity of the review was demonstrated using the PRISMA-ScR (Figure 1). Studies were retrieved using a search strategy, and those from additional sources were screened based on their titles and abstracts by the first author (RCH) to determine if they met the inclusion criteria mentioned above. Studies that fulfilled the inclusion criteria were retrieved and independently assessed by the authors (RCH and CJC). The researchers were not blinded to each other’s decisions.

A standardized data extraction form was used by 2 authors to independently extract data on the name and location of the study, study participants, study design, PFMT program, description of PFMT mobile app features, outcome measures, and results in the eligible studies.

Discrepancies between authors RCH and CJC were identified and resolved through discussion. The data were recorded in an Excel (Microsoft Corp) spreadsheet (Multimedia Appendix 3 [34-39]).

Adherence data were analyzed descriptively. A meta-analysis was not possible because of the clinical and methodological diversity in the included studies [40]. The aim of the analysis was to explore differences in adherence levels between the intervention and comparator groups.

BCTs identified in PFMT mobile apps were mapped using the COM-B model [22]. To determine which BCTs used in the intervention group target which core behavior of the COM-B model, the step-by-step mapping guidance to design behavioral interventions using the COM-BCW was used in reverse [22,24]. Rather than designing an intervention, the reverse allowed the identification of each core behavior from each BCT using the theoretical domains framework linked with the COM-B [22]. The 5-step process went as follows:

The characteristics of the intervention group were extracted into an Excel spreadsheet (Multimedia Appendix 3 [34-39]).

The description of the PFMT mobile app features and their purpose in the included papers allowed the authors (RCH and CJC) to identify which BCTs they used. The individual PFMT mobile apps were not downloaded. It was assumed that the authors of the included studies followed the correct way to report complex digital interventions and that their description was accurate. The descriptions of the digital interventions were included in the data extraction table (Table 1). All recognized BCTs were included. The BCT identification was completed by authors RCH and CJC. Any discrepancies between the identified BCTs were resolved through discussion. All PFMT mobile apps include one or more BCTs. No BCTs in the taxonomy were excluded based on perceived appropriateness in health and care settings, for example, future punishment.

Using the description of the PFMT mobile app features and their purpose in the included papers, the associated BCT was identified; for example, “daily reminder notification” is associated with “prompts and cues.” Descriptions of PFMT mobile app features with no associated BCT were excluded. The identified BCTs were then grouped and mapped in a Word (Microsoft Corp) table (Table 2).

The BCTs in the PFMT mobile apps and usual care groups were then directly linked to their associated theoretical domains [22]. For example, “prompts and cues” are associated with the “environmental context and resources.” The final step involved linking the theoretical domains with the COM-B components; for example, “environmental context and resources” is linked with “physical opportunity.” This process allowed the core behavior (the BCT was targeting) to be identified according to the process described in the BCW, a guide to designing interventions [22], which is displayed in Table 3.

Identified BCTs were narratively synthesized following the guidance for conducting systematic scoping reviews [41].

Figure 1 details the process of selecting studies according to PRISMA-ScR standards. A total of 114 studies were retrieved and screened against the inclusion and exclusion criteria. A total of 6 studies met the inclusion criteria, including 4 randomized controlled trials (RCTs) [34-37], 1 pilot RCT [38], and 1 historical controlled trial [39].

The studies included were published between 2017 and 2022. A total of 2 studies were published in Sweden [34,37], and the rest were published in Brazil [35], China [36], Japan [39], and Malaysia [38]. Sample sizes ranged from 26 to 129, for a total of 471 participants. All studies were treatment studies [34-39], and no prevention studies were found. A detailed summary of all the characteristics of the studies selected can be found in Table 1.

Women aged between 21 and 72 years were recruited in the included studies, with half of the studies including women aged 40 years or younger [36,38,39]. Women included in all 6 studies had access to a smartphone [37-39], and 3 studies required internet access [34,37,38]. A total of 2 studies involved pregnant women between 26 and 32 weeks of gestation [36,38], and 2 studies recruited women following recent vaginal delivery [36,39]. A total of 4 studies recruited women who had symptoms of SUI [34-36,38]. A total of 3 studies recruited women with mixed urinary incontinence [35,37,38], and 1 study recruited women with urge urinary incontinence [38].

The study length ranged from 2 to 6 months. In 4 studies [34,37-39] PFMT was recommended 3 times per day, and in 1 study, PFMT was recommended twice a day [35]. Alternatively, in 1 study [36], participants were given the option of completing PFMT twice a day or completing 150 pelvic floor muscle contractions a day [36]. The number of fast contractions varied between 0 and 10, the number of slow contractions between 1 and 6, or training time was recommended up to 15 minutes [36]. A total of 2 studies [34,37], stated their intervention group had different “combinations” of PFMT programs participants could use, but this was not expanded. The variety of PFMT training programs and modes of delivery used in the included studies highlights the lack of standardization of PFMT programs and could potentially influence the outcomes. In addition, a total of 3 studies [36,38,39] used a higher PFMT frequency and duration than recommended [9].

A total of 2 studies [35,39] used health care professionals to provide PFMT information to both the intervention and control groups. A specialist physiotherapist taught PFMT using surface electromyography and pelvic floor examination in 1 study [35]. Midwives taught participants through verbal instruction and written information in the other [39]. In comparison, 1 study used trained members of the research team to deliver PFMT [36]. In this study, all participants were given 2 sessions of a 45-minute presentation on pelvic floor education and one-on-one PFMT practice at the start and just before the end of the study [36]. The first presentation focused on the anatomy of the pelvic floor and the PFMT technique. The second focused on the importance of sustaining PFMT [36].

A total of 1 study offered 3 follow-ups with a continuous health care professional that included a pelvic floor examination for all participants [35]. One study offered follow-up telephone contact from a member of the research team once a month to encourage all participants to complete PFMT [36]. In 1 study [38], participants in the controlled group received usual antenatal care from a midwife.

A total of 2 studies provided written PFMT information to the control group [34,35], 1 study [38] postponed the provision of the intervention app until the study was complete, and 2 studies [36,39] provided no additional information. In 1 study [37], the control group received a restricted version of the PFMT mobile app used in the intervention group. A total of 4 studies [34,36,37,39] provided digital support to the intervention group using a PFMT mobile app. In summary, usual care varied between the studies.

It is evident that the variation in PFMT programs, modes of delivery, and levels of professional involvement underscores the lack of uniformity in PFMT across the studies. This makes it difficult to compare the results between studies.

Adherence was the primary outcome measure for 4 studies [35,36,38,39], with the remaining studies reporting it as a secondary outcome [34,37]. The primary outcomes of these studies were the International Consultation on Incontinence Questionnaire for Urinary Incontinence Short Form [42] and the Lower Urinary Tract Symptoms Short Form [42], used in 1 study [34].

The outcome measures used to measure PFMT adherence included the Exercise Adherence Rating Scale [38,43], the Broome Self-efficacy Scale for Practicing Pelvic Floor Exercises Questionnaire [36,44], self-reported adherence using questionnaires [39], or the Visual Analog Scale scores [35]. In one study [35], in-app adherence measures were also used that recorded the number of PFMT protocol repetitions the participants in the intervention groups completed. The findings suggest there was no agreement on how to measure adherence.

The BCTs used in the intervention groups were “self-monitoring of behavior,” “self-monitoring of outcomes of behavior,” “instruction on how to perform behavior,” and “prompts and cues” (Table 2). The findings suggest that a combination of BCTs were used in the intervention group.

The BCT “prompts and cues” were used in all intervention groups [34-39]. In 5 of 6 intervention groups, the BCT “instruction on how to perform behavior” was used [34-38]. A total of 2 intervention groups used all the BCTs mentioned above [35,37], whereas 1 used only “prompts and cues” [39]. In summary, “prompts and cues” were the most frequently used BCT in the intervention groups.

The COM-B core behaviors targeted by the BCTs used in the studies were opportunity and capability (Table 3). Opportunity was targeted using “prompts and cues” in all the intervention groups [34-39]. Capability was the core behavior targeted using the BCTs “self-monitoring of behavior,” “self-monitoring of outcomes of behavior,” and “instruction on how to perform behavior” in the intervention group [34-38]. The findings suggest that the core behavior opportunity was most frequently targeted using “prompts and cues” in the intervention groups.

The levels of adherence to PFMT were greater in the intervention group compared with the control group, according to 2 studies [34,39]. In 1 study [34], daily PFMT adherence in the intervention group was 40% compared with 3% in the control group, with 60% of the control group reporting “sporadic” PFMT. Conversely, in 1 study [39], daily PFMT adherence was reported at 15 repetitions a day over 7 days in the intervention group compared with once a day over 3 days in the control group. In summary, higher levels of daily adherence to PFMT were observed in the intervention group.

In the control group, PFMT adherence decreased steadily over time [35,36]. In comparison, PFMT adherence increased over time in the control group of another study [38]. Participants in the control group, where adherence increased over time, received a PFMT mobile app at the end of the study period, which may have influenced the results.

To achieve objective 1, we conducted an analysis of the BCTs used within studies with PFMT mobile app features. Our analysis identified a combination of BCTs with a range of strategies to facilitate adherence to PFMT, including “self-monitoring of behavior,” “self-monitoring of outcomes of behavior,” “instruction on how to perform behavior,” and “prompts and cues.” “Prompts and cues” were the most widely used BCT. To achieve objective 2, we mapped the BCTs used in the studies onto the core behaviors of the COM-B model. This highlighted the central emphasis on the core behavior of opportunity, evident through the prevalence of “prompts and cues” across all intervention groups. Additionally, “self-monitoring of outcomes of behavior” and “instruction on how to perform behavior” were used to address the core behavior capability. The widespread use of “prompts and cues” highlights their role in fostering PFMT engagement by making the most of opportune moments. Furthermore, the incorporation of “self-monitoring of behavior” and “instruction on performing behavior” empowers participants with the necessary knowledge and confidence for effectively executing PFMT. To address objective 3, data on PFMT adherence were extracted from the included studies. This included adherence data from self-efficacy measures and patients’ self-reported adherence measures. The findings suggest higher levels of PFMT adherence were observed in women using PFMT mobile apps; however, there was heterogeneity in the outcome measures used in the included studies, and meta-analysis was not possible.

Digital “prompts and cues” were used by all PFMT mobile apps. Digital “prompts and cues” are a BCT that target the core behavior opportunity according to the COM-B model. Digital reminders are a “persuasive” behavioral intervention that may be valuable when it comes to changing women’s behavior around PFMT, since women commonly associate poor adherence with forgetting to complete PFMT [14,28]. Persuasive behavioral interventions can help women overcome known barriers to PFMT by creating more positive attitudes toward the exercises [28]. These findings are supported by a systematic review that found daily PFMT reminders “stimulate” daily adherence to PFMT in women [21]. Similarly, a systematic review that aimed to determine the persuasive features used in PFMT mobile apps recognized that reminder features were a way of assisting users to remember to perform PFMT, which the authors mapped under the core behavior motivation rather than opportunity [28], which demonstrates how elements of BCT are disputed. In the wider literature, digital reminders in populations with long-term conditions, including diabetes, suggest that digital reminders increase adherence to medication and physical activity [45-47], although none of these papers discuss how behaviors might be influenced. In this review, the results suggest digital PFMT prompts increased daily PFMT adherence in women, and it might be argued that leveraging either or both of the core behaviors of opportunity or motivation. However, further rigorous research is required to explore the effects of digital PFMT prompts on women’s PFMT adherence while controlling other behavioral variables.

Some studies have observed higher levels of PFMT adherence among women using PFMT mobile apps [34,39]. Although the intervention groups reported higher levels of daily adherence, women did not meet the daily PFMT recommendations of 3 times a day [2]. In 1 study [35], women completed PFMT 1.6 times a day on average, and in another [37], only 10% of women completed PFMT 3 times a day. One reason for the lower levels of PFMT adherence may have been the heterogeneity of the outcome measures of adherence that were inconsistently monitored and poorly described [12]. Alternatively, the heterogeneity in PFMT program prescription may have influenced adherence to PFMT, with 3 studies prescribing a higher dosage of PFMT [36,38,39] than recommended [9]. The high frequency and intensity of PFMT recommendations are at odds with the “easy, accessible, social, timely” principles of behavior change [48]. If the exercises and the intensity of the regime appear difficult, it is likely that women are less likely to engage in and adhere to PFMT. This is particularly important to consider given the fact that no prevention studies were identified in the scoping review, and women may find adhering to PFMT even harder if they are asymptomatic.

Women’s adherence to PFMT appeared to decrease around 3 months [35,36,38]. These findings are supported by other studies that have explored adherence rates in people with chronic disease and national guidelines [2,48], suggesting digital prompts may only be valuable in the short term. It is recognized that short-term PFMT adherence is greater than long-term adherence [7]. Previous research indicated that behavior change can take up to a year [49]. In the included studies of this review, there were none that followed up a year postintervention, and therefore we are unable to make a comment on the long-term nature of behavioral change. The decrease in PFMT adherence after 3 months may reflect a normal drop in adherence levels [49] because women notice an improvement in their symptoms and no longer consider the exercises to be beneficial. It has been suggested that “social opportunity” rather than “physical opportunity” (referenced in the COM-B model) is important when it comes to behavioral maintenance, since actions are more likely to be sustained when in line with group norms [49]. Further behavioral research on PFMT adherence in women should increase the length of study to 1 year and explore the role of “social opportunity” on adherence.

Conversely, in the control group of 1 study, adherence increased at 2 months [38]. A potential reason for this is that the control group received “usual antenatal care” that involved contact with health care professionals. The influence of health care experts on exercise adherence is “pivotal,” according to other studies [50]. Arguably, contact and follow-up with a health care professional as a “credible source” may be a valuable BCT in relation to PFMT adherence in women because it holds women accountable to complete PFMT [22]. The use of health care professionals to teach PFMT and follow-up participants in some of the included studies may have led to confounding behavioral factors that influenced individuals’ adherence to PFMT. Future research should consider the influence health care professionals have on participants PFMT adherence when conducting future research on PFMT mobile apps.

Using a scoping review methodology, we have been able to identify a breadth of experimental studies investigating the effect of PFMT apps on adherence to PFMT in women as an intervention. To the authors’ knowledge, this study was the first scoping review to identify and discuss specific BCTs embedded within existing features of PFMT mobile apps and to map the core behaviors targeted using the COM-BCW framework. By doing so, this research contributes to a deeper comprehension of the factors influencing PFMT adherence among women and highlights the potential use of PFMT mobile apps, particularly those with digital prompts, to increase adherence to PFMT. Additionally, the study uses the COM-B model to provide a holistic understanding of how capability, opportunity, and motivation are used in the context of PFMT adherence, bridging the gap between theoretical concepts and application that can inform the development and enhancement of interventions targeting PFMT adherence.

A limitation of the study was the way BCTs were assigned to the features of PFMT mobile apps based on descriptions of the apps and app features in the included papers. One BCT was assigned per app feature, and features that did not specifically relate to any of the 93 BCTs were not included. Inaccurate descriptions of mobile app features by the authors of the included studies may have resulted from misattribution or underestimation of the BCTs and therefore misrepresentation of the core behaviors targeted by PFMT mobile apps. The misrepresentation of core behaviors may have made the core behaviors that encourage women to adhere to PFMT unclear. Additional BCTs may have been identified if the PFMT mobile apps were downloaded or original papers around the app development were sought and appraised.

Another limitation of the study was the exclusion of studies without a control group. The exclusion of additional studies, such as observational studies, meant that studies with potentially fuller descriptions of PFMT mobile apps and their intended behavioral mechanisms were not included in the analysis. Arguably, this limits the depth of the analysis in the study by limiting the comparison and corroboration between a greater number of PFMT mobile apps that would have better addressed the first 2 objectives of this study. Alternatively, the inclusion of studies with a control group increased our confidence in the findings that suggest levels of adherence were higher in women using a PFMT mobile app compared to an alternative variable and addressed objective 3.

This review identified the use of mobile apps for the treatment of PFMT and the behaviors being leveraged; however, due to the heterogeneity of the PFMT regimes across the studies, it was not possible to compare. In the future, it will be important to identify the minimum effective training dosage women engage with and why.

The findings suggest there was a trend in higher levels of PFMT adherence in women using PFMT mobile apps over 3 months. However, there was no consistency across the studies in the PFMT regimes and outcome measures. Digital “prompts and cues” were used to encourage adherence and may be the BCT that leveraged improved adherence. Opportunity was the most targeted core behavior in the intervention groups of the included studies. Further research is required to practically assess whether a future RCT that investigates the effectiveness of digital “prompts and cues” on PFMT adherence in women can be conducted.