We examined the content, features, usability, and engagement of mHealth apps based on DBT. There were several notable findings. First, there are several apps available based on DBT, yet they varied greatly in scope, features, and function. Most apps included aspects of skills training, while only a fraction included other crucial aspects to DBT delivery such as a diary card or chain analysis. Approximately half of the apps were solely based on DBT, while the others integrated DBT components to complement other therapy skills, and most apps were designed to be used without a therapist. A few highly popular apps attracted tens of thousands of users, with most apps having fewer than 50 monthly active users. The majority of the apps scored in the usable range. User ratings, number of features, and number of reviews were positively correlated with MARS scores, indicating overall agreement in app quality. Finally, the quality and feature selection available within DBT apps differed based on both the cost of the app and whether the apps were solely focused on DBT or merely included some components of DBT adjunctively with other psychotherapy content.
Our systematic review of mobile apps based on DBT revealed 21 apps that were downloadable, free, and functional. Though we identified 33 apps based on DBT, we were unable to code them all because many were unable to be downloaded or were not free. In an effort to streamline the delivery of DBT, a certification process was adapted, which involves testing and formal adherence ratings. There does not exist a similar process for the development and distribution of mobile apps based on DBT. As a result, the apps that we identified varied greatly in quality, content, and function. Integral to the delivery of DBT is daily tracking of behaviors via a diary card as well as skills use. DBT skills training was the most commonly represented component of DBT across the apps reviewed. This finding was not surprising, given that the skills mode of DBT is the most portable, adaptable, and widely disseminated part of the treatment [31, 32]. Moreover, DBT skills are most easily translatable into a mobile format. DBT skills are discrete, modular, follow a protocol, and some even include a flow chart [33], which make the skills particularly amenable to computerization. Stylistic features of DBT such as irreverence and other essential treatment protocols like chain analysis, were less frequently integrated, likely due to the complexity of translating such highly idiographic and interpersonally interactive elements. Nonetheless, only 23% of apps included a diary card, and of those, only two offered the ability to customize it and share it directly with a therapist through the app. Apps designed to track behavior, thoughts, and emotions are common [34], which is why an app based on DBT that does not include this feature is surprising. Slightly less than half of the apps included features for crisis management and more than a third of apps provided users access to suicide hotlines. These features are particularly important as DBT is considered one of the most effective interventions of choice for those at high risk for suicide [1]. Having access to crisis planning and hotlines may be particularly important with apps that are designed to be delivered in the absence of a trained clinician.
We found that there were large differences in usage between the apps. The three most popular apps accounted for 89% of the total monthly active users. While some apps had thousands of users, half of the apps had 36 or fewer active users. This immense divide between the popular apps and unpopular apps is striking. One plausible explanation is that the highly popular apps (Wysa, Youper, Woebot, and Calm Harm) are transdiagnostic and offer a variety of content, whereas other apps (e.g., DBT Coach, DBT Travel Guide, DBT Trivia and Quiz) are narrower in their focus. Apps like Wysa and Youper may attract users with a broader variety of concerns than apps that exclusively or primarily offer DBT. Another plausible explanation is that users generally tend to gravitate toward a small number of highly engaging and well-advertised apps. Our findings are not unique to DBT apps: previous research has found similar distributions of active users in apps for depression and anxiety [35], eating disorders [36], and other health conditions [27]. Thus, the large differences between popular and unpopular apps may not be as surprising as it may appear at first glance. Future research will be needed to understand specifically why and how the popular mHealth apps successfully attract and retain users.
Our review identified several differences between apps that were advertised as being solely based on DBT vs. apps that included some aspects of DBT to complement a suite of other therapeutic techniques. Apps that were advertised to be solely focused on DBT seemed to be designed with a specific function in mind. For example, the app “DBT: The Dime Game,” is designed to support practice and implementation of one specific DBT skill while “Impulse DBT” is designed to walk users through a chain analysis--a crucial component in individual DBT. These apps suffered in the amount of breadth of “DBT specific” features that they contained, yet were clearly designed to support more in depth features around essential DBT tasks. Apps like Youper and Wysa integrated specific DBT skills within a toolbox of evidence-based interventions designed to improve a broad class of symptoms. While the apps that integrated components of DBT tended to be highly usable and engaging, pulling from a mix of treatments for an app may lead to a confusing or disorienting user experience. For example, someone searching for DBT based support may find it difficult to understand how much DBT is contained in an app that advertises other treatments, or whether any particular skill or content in the app is DBT based or not. On the other hand, apps that draw on a range of evidence-based approaches are philosophically and theoretically consistent with DBT, in the sense that DBT utilizes whatever is proven to work to achieve clinical progress. Our usability analysis also showed that mixed-DBT apps showed just as high, if not higher, usability ratings, suggesting that adding other evidence-based treatments to a well-designed DBT app may not compromise the user experience.
Apps that included integration with a therapist (such as sending a diary card to a therapist) were coded as “adjunctive;” most apps were designed specifically for skills practice in the absence of therapy while less than a quarter were designed to augment in person treatment. DBT along with other evidence-based treatments were originally designed to be delivered in person; however, the COVID-19 pandemic has likely permanently shifted how outpatient therapy is delivered. While many clinicians report difficulties delivering treatment via telehealth, the convenience for both patients and providers indicate that at least some proportion of telehealth delivered mental health care is here to stay [37, 38]. As a result, effective technologies and mobile applications that can support teletherapy are needed now more than ever.
Our results revealed an overall lack of attention to age and developmental factors with regard to DBT apps in the sample. Only two apps were explicitly designed to be appropriate for younger people, while most apps were generally described as being appropriate for “everyone”. Adolescents are particularly amenable to mobile mental health, with 95% of adolescents reporting that they own a smartphone [39], and approximately 64% of adolescents reported using apps to manage mental health symptoms [40]. Practitioners and researchers of DBT for adolescents and youth have long advocated for the importance of developmentally sensitive modifications to DBT when applied for this population [30, 41]. Although some apps in the current sample may be effective for adult users, they may be less appropriate for youth for a variety of reasons such as containing content written at too complex of a reading level, not being engaging or “youth friendly” enough, or not addressing or acknowledging the importance of the family context of youth. Future work in this area should remain sensitive to how DBT based apps can be developed to be efficacious and well adapted to the full age range of DBT mHealth users. Otherwise, DBT mHealth app developers may consider developing different versions of effective DBT apps that are specifically tailored to be developmentally appropriate (e.g., DBT adult apps, DBT adolescent apps, and DBT apps for children).
The significant positive associations between MARS scores and other indices of app quality suggests that both DBT clinicians and patients can use app ratings and reviews as a proxy for app quality; however, it should be noted that the correlations were not particularly strong with expert ratings. Moreover, the large positive correlation between MARS scores with user star ratings, number of ratings, and features indicates that apps that are downloaded frequently and have several features tend to win over users and raters. Engagement in mHealth is a critical, but often overlooked aspect in app development. In general, most users of mobile mental health apps tend to stop using an app after 10 uses [42], and given mHealth use is associated with clinical outcomes [13], how the content is designed and displayed is as important as the content itself. To note, the majority of the apps were developed by a commercial entity and some of the highest rated apps included a “pay for more” feature, indicating a business model that could sustain the app development, updating, and maintenance. The median cost of developing a smartphone app is $171,000, and yearly maintenance costs is estimated to be around 20% of development costs [43]. As such, it may be too expensive and infeasible for researchers to develop and maintain their own apps. Additionally, it appears to be the case that only a small number of apps has attracted and retained active users. However, the consequence of the costly app development is the lack of research supported apps on app stores, relegating usable and engaging apps to be developed by business. To note, neither the Behavioral Tech developed DBT coach (e.g. 11) nor Pocket Skills (e.g. 12) were identified in this app review because they are not available to download, highlighting a “lab to marketplace” gap for apps developed in the context of research.
Limitations
While this is the first review that has systematically reviewed mobile apps based on DBT for content quality and user experience, there are several limitations that need to be addressed. One of the most glaring omissions in this review is the lack of concurrent systematic review to evaluate the apps’ clinical efficacy, as such a review would be outside the scope of this study. As a result, we are only able to comment on the apps usability and engagement, rather than potential efficacy. In addition, we did not thoroughly review mobile apps that required payment to download, limiting the breadth of our review. The inclusion of paid apps may have shifted the pattern of findings in the current study, that is, apps that required payment may have more features and may have been more engaging. However, clinicians report that the move to teletherapy has disproportionately affected low-income patients [44], and non-free mobile apps exclude these patients from the marketplace. Another limitation is that our app selection process only included a systematic search of iOS and Android app stores, and no systematic literature search was performed. We relied primarily on descriptive statistics with this small sample size, which limited our ability to draw statistical inferences from the observed patterns in the current sample. Thus, trends and patterns described in the current study should be interpreted cautiously. Although our content coding included some essential DBT components such as skills training and the diary card, there are many features of DBT that we did not code or discuss in the context of mHealth apps. Future studies on DBT mHealth apps should aim to expand coding and analyses to characterize a broader range of DBT features in order to further assess how well this complex treatment can be translated into apps and technology. Finally, the coders in this study were researchers rather than individuals with lived experience with DBT. Future studies should analyze DBT apps through the lens of patients and individuals with lived DBT experience as well as consumers of mental health apps.