Physical activity during pregnancy is associated with a variety of health benefits, including reduced risk of excessive gestational weight gain, gestational diabetes, gestational hypertension, preeclampsia, the severity of pelvic girdle pain, macrosomia, instrumental delivery, postpartum weight retention, urinary incontinence, and depressive disorders [1,2]. Despite the many health benefits of physical activity during pregnancy, few women participate in regular physical activity during pregnancy [3]. In addition, women tend to reduce or cease their participation in physical activity once they become pregnant and throughout the course of their pregnancy [3,4]. This may be as a result of various barriers, including mother-child safety concerns, fatigue, change in body shape, and associated pain. Further, a lack of provision of adequate information, knowledge, social support, and self-efficacy for behavior change are other issues that may exacerbate the decline in activity levels throughout pregnancy [2,5].

There are many avenues for women to access information and support related to maintaining a healthy pregnancy, including information about physical activity behaviors. Historically, pregnant women have accessed information from doctors, midwives, family, and friends to guide and inform their physical activity behaviors. However, eHealth platforms such as the internet and mobile apps are now altering the way women access this information [6] and they have become an important information source for pregnant women [7,8]. In fact, a recent Australian study among 410 pregnant women investigated the use of pregnancy and parenting apps and found that almost three-quarters of the studied women used at least one of these types of pregnancy apps [9]. In addition, more than half of the participants reported using 2-4 apps throughout their pregnancy. The frequency of app use was also significant, with almost a quarter of pregnant women reporting daily use of apps [9]. While the use of pregnancy-related apps has significantly increased among pregnant women [8], very little is known about their theoretical underpinnings, including their utilization of behavior change techniques (BCTs). This is despite research suggesting that inclusion of BCTs in eHealth interventions can play an important role in improving, supporting, and maintaining healthy behaviors, including physical activity [10,11].

In 2013, Currie et al [12] systematically evaluated the content of physical activity interventions designed to reduce the decline of physical activity in pregnant women with a specific emphasis on BCTs [13] employed to elicit this change. Six common BCTs shown to have some efficacy in improving physical activity behaviors were identified: prompt intention formation, prompt specific goal setting, prompt review of behavioral goals, prompt self-monitoring of behavior, provide feedback on performance, and provide opportunities for social comparison [12]. Since this review, many behavior change interventions have used these BCTs to promote positive physical activity behaviors among pregnant women [14].

Previous reviews of physical activity apps for other population groups suggest that commercial apps often lack evidence-based BCTs that have demonstrated efficacy for encouraging physical activity behavior change [15-17]. However, no such review of commercial apps designed to promote physical activity among pregnant women has been conducted. Thus, the appropriateness of these apps to promote physical activity during pregnancy is unknown. This review aimed to systematically evaluate the appropriateness of the apps designed to promote physical activity among pregnant women by using a systematic search and content analysis. Apps available through the Australian App Store and Google Play stores were accessed using the MARS tool for app quality and features. A taxonomy of BCT was also used to assess the presence of BCTs utilized within the apps, including BCTs that have demonstrated efficacy for promoting physical activity behavior change among pregnant women.

The primary aim of this app review was to assess app quality and features and presence of BCTs applied within commercial apps promoting physical activity in women during pregnancy. The secondary aim of this app review was to test the relationships between the 6 MARS domains of app quality and the number of MARS technical aspects, theoretical strategies, focus points, and BCTs. In summary, our findings demonstrate moderate MARS overall quality scores, with MARS domains of functionality and aesthetics scoring the highest and the domains of subjective app quality and likelihood for behavioral impact scoring the lowest. An average of 4.74 BCTs per app were used, with the most common BCTs being provide information on consequences and provide instructions. Most apps had none or one of the BCTs that demonstrated efficacy in promoting physical activity behaviors during pregnancy, with the most common BCTs being opportunities for social comparison and prompt self-monitoring of behavior.

Many commercially available physical activity apps for pregnant women were identified—with 69 relevant apps and 19 apps remaining after apps targeting postnatal physical activity (n=9) and apps with user ratings lower than 4.5 (n=39) were excluded. This finding demonstrates a market for apps targeting physical activity during pregnancy, which is not surprising, considering the findings of Lupton and Pedersen [9] who have previously reported that pregnant women frequently use pregnancy-related apps. The 69 apps identified as targeting exercise in pregnancy is comparable to the 51 apps Brown et al [18] identified as targeting nutrition in pregnancy. Pregnancy seems to be a time in women’s lives in which they are actively seeking and using apps to support their behaviors, including physical activity.

The high MARS functionality scores and moderate-to-high MARS aesthetics scores may be a consequence of developers focusing on user experience for visually pleasing and user-friendly apps [26]. Although such features are important for attracting users, they do not translate to behavior change [27,28]. There is also room for improvement in app design in relation to the low MARS scores for quality of information and engagement. Given that pregnant women believe the information in apps has high credibility [9], it is important that apps ensure that evidence-based information is provided. It is also important that apps improve ratings of engagement, as engagement in app content is associated with behavior change [22].

Ratings of engagement could be improved by increasing the use of BCTs. Our study found an association between the number of BCTs and improved MARS engagement and aesthetics scores, which is consistent with the findings from a review investigating apps for weight management in adults [22]. The association between the number of BCTs and the engagement score could be due to the provision of additional content to engage in, increased usefulness, and perceived efficacy. It may be that BCTs are considered as enticements to potential users, thereby increasing engagement. If this is the case, this provides additional incentives for app developers to include BCTs. In line with that reported by Schoeppe et al [15] and Bardus et al [22], this study also found technical aspects of health behavior change apps to be associated with the MARS engagement score. Further, the number of theoretical strategies and the number of focus points were associated with MARS engagement, aesthetics, total quality, and subjective quality scores. Improving the number of technical aspects, theoretical strategies, and focus points may therefore help to further improve engagement of apps designed to promote physical activity among pregnant women.

An average of 4.74 BCTs were used in the reviewed apps. This is in line with the findings from previous research that found apps for physical activity promotion in adults to have an average of 4.2 [16] and 5.0 BCTs [24]. The number of BCTs identified in this review is, however, higher than those identified in a recent review exploring the number of BCTs among apps targeting nutrition behaviors in pregnancy [18], which reported an average of 3 BCTs per app [18]. We know that health behavior interventions using alternative modes of delivery (eg, websites) have improved outcomes when more BCTs are used [29]. Webb et al [29] suggest that this improvement may be a result of using a combination of BCTs that together target several stages and aspects of behavior change. However, we do not know the optimal number or combination of BCTs necessary to increase physical activity.

The most common BCTs identified in this study were provide information on consequences and provide instructions. Schoeppe et al [15] also found that providing instructions was a commonly used BCT in physical activity and diet apps for children and adolescents. This finding differs from the most common BCTs identified in apps targeting weight and physical activity in adults, which are goal setting, self-monitoring, and performance feedbacks [15,29]. Unfortunately, provide information on consequences and provide instructions have not been shown to be effective at improving physical activity behaviors in pregnant women [12] or even in the general adult population [27].

Most apps in this review had none or 1 BCT, which previously demonstrated efficacy in behavior change during pregnancy. Including more of these evidence-based BCTs in the context of pregnancy may improve the ability of the apps to support pregnant women in increasing their physical activity [12]. The evidence-based BCTs of intention formation, goal setting, review of goals, and feedback on performance have been successfully implemented within physical activity apps targeting other population groups [15,16]. This demonstrates that it is feasible to deliver these evidence-based BCTs through an app. In particular, feedback on performance was identified as one of the most commonly used BCTs in apps targeting physical activity, diet, and sedentary behavior in children and adults [15]. Feedback on performance requires the measurement of behavior (eg, frequency, intensity of exercise), and apps that incorporate wearable devices (eg, Fitbit, Garmin) to measure physical activity behavior may be more likely to provide feedback on performance [30]. Therefore, the exclusion of apps that required the use of a wearable device may partially explain the finding of low use of feedback on performance in this study. The most common technical features were sends reminders (9/19, 47%) and allows sharing (8/19, 42%). There is evidence to suggest that reminders improve the effectiveness of health behavior change interventions [31]; however, other evidence suggests that reminders can hinder habit formation, which may have an impact on long-term behavior change [32]. The high use of sharing is contradictory to the findings that adults find social media features unnecessary and off-putting in health behavior change apps [33]. The removal of sharing features may improve engagement ratings.

The strengths of this study include the systematic search for apps from both App Store and Google Play, the use of an established taxonomy for identifying BCTs, the use of the MARS instrument to assess the quality of apps, and the inclusion of apps rated above 4.5/5. Further, app ratings were performed by a minimum of two reviewers following best practices for conducting systematic reviews [19]. Good interrater reliability was found for the scoring of BCTs.

The limitations of this study include the exclusion of apps that required use of a wearable device and low interrater reliability for the scoring of app quality through the MARS scale. This may be due to the subjective nature of some sections of the scale. What one reviewer may find aesthetically pleasing, functional, or engaging, another reviewer may not. To account for these differences, the average of the 2 scores was used as the final score for each domain. The temporal relevancy of the results from this study may also be considered a limitation, despite the search being conducted less than 12 months ago because unlike traditional literature where the content remains consistent and unchanged once published, apps are extremely fluid, resulting in frequent updates and modifications to their contents and features. Despite these limitations, this study is the first of its kind and provides valuable real-world findings and implications that are highly relevant to this field as well as the greater audience. Future research should test the overall effectiveness of commercial apps designed to promote physical activity among pregnant women. Finally, research examining the accuracy of app content and the expertise of developers should also be of high priority.

In conclusion, the use of apps for physical activity advice and support in pregnancy is rising. Therefore, an understanding of their quality and inclusion of effective BCTs is required. This is the first study to investigate the quality of popular commercially available apps for promoting physical activity in women during pregnancy. The findings of moderate app quality, with the highest ratings for functionality and aesthetics, indicate that the apps are user-friendly. However, the low use of evidence-based BCTs for changing physical activity behavior in women during pregnancy indicates that the popular commercial apps currently available may not be effective at promoting physical activity behavior in the pregnant population. More effort needs to be placed on incorporating components most likely to influence behavior change, which is ultimately what they are developed for, while maintaining good functionality. Developers should continue to provide self-monitoring and social comparison and ensure that these components are engaging and effective. In addition, intention formation, goal setting, review of goals, and feedback on performance should be incorporated in new apps and new versions of existing apps.