Mobile Health Apps Targeting High-Need Populations Fall Short

Researchers evaluated 137 mobile health apps designed to help patients manage their health and found inconsistencies in the apps’ functionalities, privacy policies, and reactions to indications of health danger.

Researchers evaluated 137 mobile health apps designed to help patients manage their health and found inconsistencies in the apps’ functionalities, privacy policies, and reactions to indications of health danger. They also determined that app store ratings correlated poorly with clinical utility and usability.

The use of health apps to manage chronic disease is on the rise, and previous studies show they can help boost patient engagement. However, there is little research on apps intended for high-need, high-cost populations, and no clear guidelines exist on how healthcare professionals should evaluate and recommend apps for these patients. A recent study published in Health Affairs examined whether the available apps can address the needs of patients with varying levels of health engagement, along with several other questions.

Researchers identified 166 apps that targeted high-need, high-cost patients by searching the Apple and Google app stores for 26 chronic condition keywords, as well as gathering recommendations from the relevant professional societies. After excluding 29 apps that could not be reviewed for logistical or technical reasons, each of the remaining 137 apps was evaluated by 2 of the study authors (1 clinician and 1 nonclinician). For each app, 9 variables were analyzed: target population, functionality type as related to patient engagement, app store rating, clinical utility, usability, reactivity to user input that could indicate health danger, privacy policy, mechanism of data sharing, and cost model.

Some target populations had a wider selection of available apps than others. For instance, 25 apps targeted people with diabetes, while only 4 apps targeted people with arthritis. However, researchers noted that “many apps were designed to serve multiple high-need, high-cost populations, and 11 of them were deemed general enough to be useful across all such populations.”

Nearly all of the apps included functionalities more likely to be useful to patients with relatively low health engagement levels, like health data tracking or educational information and reminders. Very few apps provided guidance based on user-entered information, rewarded behavior change, or offered support through social networks, which would be more helpful for patients with high levels of engagement.

One of the major study findings was that an app’s rating on the app store correlated poorly with clinical utility and with reviewer-assessed usability, even among apps with high ratings. For instance, an app with an almost 5-star rating was assigned a usability score of 30 (on a scale of 0 to 100) by a reviewer evaluating how easy it would be for a consumer to use.

Alarmingly, only 28 of the 121 apps that allowed patients to enter health information responded appropriately to indications that the user could be in health danger. Some examples of appropriate reactions would be providing a suicide hotline number if the user reported suicidal ideation or suggesting that a user reporting prolonged chest pain call 911.

The researchers also raised concerns about privacy and data sharing. Most apps allowed users to share their health information with other people, but 60% did so through insecure data-sharing methods like e-mail or text messaging. Just 64% of the apps reviewed had a privacy policy.

In light of these findings, the study authors recommended that “medical professional societies and patient advocacy groups should give serious thought to how apps may be used to benefit specific patients.” One proposed solution was the use of clinician-developed health app labels that would help patients find an app that fits their needs while including warnings for any of the app’s shortcomings.