Although thousands of smartphone mental health apps exist, research suggests they often provide overlapping features, give insufficient privacy information, and do not target patients who may need them most.
Among more than 500 apps studied, most offered basic functions such as psychoeducation, goal tracking, and mindfulness and fewer included innovative features such as biofeedback or specialized therapies.
Measures of app popularity, such as consumer star ratings or the number of downloads, did not correlate well with privacy and efficacy according to the study published in JAMA Network Open.
“The findings of this cross-sectional study suggest that the current app marketplaces lack diversity in their offerings and fail to implement potentially high-impact features,” report Erica Camacho and colleagues from Beth Israel Deaconess Medical Center in Boston, Massachusetts.
“Another challenge to the app space is that easily accessible metrics like star ratings fail to consider privacy capabilities.
“Thus, clinicians and patients must discern apps beyond such measures to ensure the discovery of apps that both fit their unique needs and protect their privacy.”
The COVID-19 pandemic has led to increased reliance on digital tools for mental health care delivery, say the researchers.
With more than 10,000 mental health–related apps available today, they stress the importance of being able to select a safe and effective app.
The team therefore developed a large public database of mental health apps, the M-Health Index and Navigation Database (MIND).
MIND is derived from the principles behind the American Psychiatric Association’s app evaluation framework and evaluates apps across 105 unique dimensions.
Between January and July 2022, Camacho et al asked raters trained using MIND to assess 578 mental health apps available on the Apple App or Google Play stores.
Of these, 27.7% were available only on iOS, 26.6% were available only on Android, and 45.7% were available on both.
Although 88% were free to download, 44% involved in-app purchases and 34% needed subscriptions to unlock the entire app functionality.
Overall, 77% of apps had a privacy policy but nearly half (44%) shared personal health information with third parties.
The apps offered 22 unique features relating to therapeutic areas. Of these, the most common were psychoeducation, in 41%, followed by goal setting/habit and mindfulness, each in 38%.
Just 1% of apps provided biofeedback with sensor data, 2% Acceptance and Commitment therapy, and another 2% Dialectical Behavioral therapy.
The three most common types of conditions the apps claimed to help treat were substance abuse related to smoking or tobacco (33%), stress and anxiety (28%), and nonserious mood disorders (20%). Only 13 apps—equivalent to just 2%—were built to address schizophrenia.
Apps that collected biodata, geolocation data, and accessed cameras and microphones were more likely than others to implement privacy and security measures.
Privacy scores did not correlate with consumer star ratings and only weakly correlated with download scores, which meant that popularity metrics provided little help in identifying those with greater privacy.
“The lack of correlation between privacy scores and consumer ratings on both the iOS App Store and the Google Play Store also suggests consumers may not be aware of or seeking apps based on privacy features—underscoring an opportunity for further education or even regulation,” the researchers report.
Nonetheless, they add that it is reassuring that the majority of apps possessed a privacy policy.
“With so many apps on the commercial marketplaces, clinicians and patients can now be more demanding in what they seek from any given app,” the investigators maintain.
They add that publicly available app libraries and validated app evaluation frameworks like MIND are innovative tools to support users in their app selection.
