Health IT, Patient Engagement,

A closer look at TEXT ME clinical trial reveals 5 ways to improve clinical validation of apps

Among the observations: it’s not good enough to look at apps as a standalone intervention strategy.

A study using text messages to improve management of cardiovascular disease published in JAMA this week offered yet another example of how apps can affect behavior change, in this case cardiovascular disease. But an editorial citing the study noted that the process for clinical validation could use some improvement and offered some recommendations.

For those unfamiliar with it, the TEXT ME randomized clinical trial (short for Tobacco, Exercise and Diet Messages) sent text messages with advice, motivation and information on diet, physical activity and smoking cessation to roughly half of the 710 participants with cardiovascular disease over a six-month period. The rest did not receive them. It then compared LDL levels, blood pressure and other relevant data. Those who received the text messages showed statistically significant, albeit modest, improvement compared with those who did not.

Drs. Zubin Eapen and Eric Peterson of Duke Clinical Research
Institute, who wrote the editorial, liked the way the investigators worked with patients (as well as academics and clinicians) both in structuring the messages sent to participants. They were part of the collaboration process for developing the messages used in the study. The study also tracked patients’ use of text messages and satisfaction with them. Their insights could easily apply to the broader clinical validation programs underway at companies and institutions.

Longer clinical trials The author quibbles that the effectiveness of the intervention was only evaluated for six months, which raises questions on how much longer users would have engaged with it. “User retention or ‘stickiness’ eludes many mobile applications, particularly those intended to effect behavior change. Understanding if the effects on cardiovascular risk factors are nullified over time because of user dropout and regression to their prior behavior is important for mobile health interventions.” From my point of view, people who are still using mobile health app after 30 days is a strong sign they will continue to use it.

For more diversity, need to break language barrier The article makes a good point here. It notes that some were excluded because they didn’t speak English. A more diverse set of users would highlight potential problems with particular users and may point to a better way to iron them out to boost  engagement. The article also takes issue that some were excluded from the Australia study because they did not own a mobile phone. It would be interesting to know how many.

Don’t rely on self-reported data: use wearables If participants used wearables, the article argues, secondary outcomes, such as physical activity, wouldn’t depend on users reporting that information themselves. The use of activity trackers (providing everyone has the same one) has its own set of limitations such as accuracy and ability to be manipulated. Using wearables would mean secondary data like physical activity could be passively reported and that activity could objectively quantify outcomes like physical activity.

Need to factor in “dose effects”  The study misses an opportunity by not looking at whether more or less texting would result in better outcomes.

Evaluation as a standalone strategy isn’t good enough One of the things that has been drummed into my head as a reporter on healthcare innovations is that chronic conditions are complex to treat and manage because frequently people have more than one and the complications are endless. So to evaluate an app or digital health tool without factoring in other interventions that are frequently required is unrealistic. As an example, it notes that:

“..even though close to half of the study participants reported attending structured cardiac rehabilitation before or during the trial, it was not clear whether the intervention was more or less beneficial to those in traditional programs. As the authors suggest, mobile health interventions should be evaluated when integrated into larger prevention efforts and used in tandem with in-person counseling.”

This may be one of the biggest challenges of designing these kind of trials, but it is also one of the most important.

Most serious healthcare tech developers know at this point that if apps are to be prescribed in a big way, they need to show a clinical benefit to persuade payers to reimburse for them. The validation process needs to be challenged to be more nimble and efficient.

Photo: Bigstock Photos