番茄社区

Extra bells and whistles on a blood pressure tracking smartphone app didn't help reduce hypertension in a randomized trial.

People assigned to an artificial intelligence (AI)-based coaching app did not have better systolic blood pressure readings than those who got a blood pressure tracking app with fewer capabilities, according to researchers led by Stephen Persell, MD, MPH, of Northwestern University Feinberg School of Medicine in Chicago.

The AI app group's average fell from 140.6 mm Hg at baseline to 132.3 mm Hg at 6 months, whereas controls similarly went from 141.8 mm Hg to 135.0 mm Hg, Persell's group reported in a study published .

No prespecified subgroup stood out for a larger effect with the intervention.

Based on AI, the app tested in the trial provided conversational support and coaching to promote healthy behaviors through cognitive behavioral therapy techniques. The app also used encouragement, reminders, and tracking to promote home blood pressure monitoring.

"The degree of blood pressure elevation observed here was lower than in many prior studies," Persell and colleagues noted. "It is possible that many participants thought that the degree of elevation was not high enough to warrant taking action."

"Future research should examine whether more directive messages from a coaching app could lead to more rapid treatment intensification," the authors suggested.

The study included 333 patients with uncontrolled hypertension (regardless of antihypertensive treatment) who were recruited in 2016 and 2017 in Chicago. Of those, 297 were randomized and completed follow-up.

Mean age of study participants was 58.9 years, and women accounted for 61.3% of the group. Just over one-third were black.

People in the intervention group received the smartphone coaching app plus a home blood pressure monitor.

Controls also got a home blood pressure monitor, but received a blood pressure tracking app instead. They were taught how to interpret home blood pressure results and to respond to extreme values.

The sole outcome favoring the AI app was a modest increase in users' self-confidence in controlling blood pressure. No differences between groups were observed in self-reported antihypertensive medication adherence, health behaviors, or other secondary outcomes.

One limitation of the study was the fact that research staff were not blinded. Also, the sample size was not large enough to detect systolic blood pressure differences smaller than 5 mm Hg or to detect clinical differences between subgroups.

Persell's team suggested that changes to the technology and more users contributing data to the machine learning algorithm could turn the story around for the AI-based app.

Comment