How to Make Digital Health Apps and Devices Actually Useful in Healthcare

PHILADELPHIA— Digital health apps and devices hold great promise in helping patients change their behaviors and make healthier choices, but there are several steps clinicians and health systems need to take in order to motivate change effectively, according to a new piece by a pair of Penn Medicine researchers.

Published in JMIR mHealth and uHealth, a perspective piece by Anish Agarwal, MD, an assistant professor of Emergency Medicine and Mitesh Patel, MD, the director of the Nudge Unit and an associate professor of Medicine and Health Care Management, argues that, as things stand now, there is not enough research or testing around the trustworthiness of most health apps and devices for clinicians to confidently use them to guide prescriptions, nor is there an effective and easy-to-interpret system for the data they bring in. But it’s possible that this could soon change.

First, the digital health modalities patients use are disparate – ranging from smart watches to smartphone apps and many combinations thereof. Often, these devices and apps are not clinically vetted.

“Clinicians and patients must be able to trust the data collected through apps and devices to inform health decisions,” Agarwal and Patel wrote. “Many consumer-grade health technologies have not been extensively evaluated for large-scale use, and their accuracy, validity, and reliability remain unknown.”

Once a digital health device or app is tested or researched enough to be considered “trusted” – as other health tools are vetted -- it’s important to have the proper repository for that information. The electronic health record (EHR) has become a critical tool for the delivery of care at every level, and Agarwal and Patel believe that if digital device data is to be used effectively, there needs to be a way for it to be ported to a dedicated part of the EHR.

Part of standardizing this data in the EHR, though, must be to make it possible to see the forest through the trees—to gather the right kind and amount of data, rather than aggregating anything and everything.

“Granular data reflecting daily step counts are likely to become overwhelming,” Agarwal and Patel note. “Instead, longitudinal trends and inflection points may reveal opportunities for intervention and longer-term monitoring.”

By looking for and collecting this kind of trend data, clinicians will have a better ability to create a personal plan for their patients and update it in real-time.

To that end, Agarwal and Patel also argue that patient care plans instituted for patients will need to follow evidence-based methodologies. That includes using time-tested approaches and being able to try new things – and have the ability to research them. Among the approaches that research does exist to support are the use of incentives and reimbursement. So in addition to having a repository in the EHR that can store and interpret data from digital devices, there will need to be the ability to set benchmarks and pull information about them.

“Similar to ‘interactive’ automobile or life insurance policies, which use smartphone data to inform incentives offers based on safe driving practices or physical activity, health insurers could stand to benefit from using enrollees’ data from behavior change programs,” the authors wrote.

Overall, Agarwal and Patel are optimistic about the use of digital devices and how they may affect behavioral interventions, allowing clinicians to prescribe them in a manner similar to prescribing a medication or a visit to a specialist. But there is work to be done before it can be considered standard.

“While several challenges exist, technology will continue to evolve rapidly,” they wrote. “Developing behavioral strategies which are evidence-based and practically scalable will help clinicians focus on implementation in their local environments. The critical next step is to better integrate and test behavioral strategies to support clinicians’ efforts to improve their patients’ long-term health.