Research points to lack of evidence that wearable biosensors improve patient outcomes

Wearable biosensors have grown increasingly popular, with many people wearing wristbands or watches to count steps or track sleep. But there is not enough proof that these devices are improving patient outcomes, such as weight or blood pressure, according to a study published in npj Digital Medicine.

“As of now, we don’t have enough evidence that they consistently change clinical outcomes in a meaningful way,” said senior author Brennan Spiegel, director of Health Services Research at Cedars-Sinai Medical Center, Los Angeles. “But that doesn’t mean they can’t.”

Wearable biosensors that automatically transmit data to a web portal or mobile app for patient self-monitoring or health provider assessment, have been touted as a means to reduce demand for healthcare, decrease costs and generate research data.

In their literature analysis, Spiegel and co-authors found remote patient monitoring with these sensors had no statistically significant impact on any of six clinical outcomes studied: body mass index, weight, waist circumference, body fat percentage, and systolic and diastolic blood pressure.

However, there is some evidence these devices show early promise in improving outcomes for certain conditions, including obstructive pulmonary disease, Parkinson’s disease, hypertension and low back pain.

“There is a big difference between using these sensors to track sleep for self-betterment and using them make medical decisions,” said co-author Michelle Keller, a clinical research specialist at the Cedars-Sinai Center for Outcomes Research and Education.

The researchers analysed 27 studies from 13 countries published between January 2000 and October 2016. Each study examined the effects of remote patient monitoring using wearable biosensors.

The interventions targeted patients who were overweight or suffering from heart disease, lung disease, chronic pain, stroke or Parkinson's.

The devices studied included physical activity trackers, blood pressure monitors, electrocardiograms, electronic weight scales, accelerometers (devices measuring acceleration) and pulse oximeters (oxygen saturation monitors), among others.

These devices were embedded in everything from watches and belts to skin patches and textiles.

Overall, remote patient monitoring resulted in no significant impact on any of the reported clinical outcomes. Certain types of interventions worked best, including efforts grounded in social science models and established care guidelines, and those that used personalised coaching.

The researchers narrowed down the 27 pieces of research they analysed from more than 4,000 studies initially reviewed, with fewer than 1 percent eligible to be included in the study, and only 16 were considered high-quality research.

There were very few randomised controlled trials for each of the clinical outcomes analysed, and studies varied significantly in terms of the types of devices used, the populations studied and the interventions tested.

“Many of the studies we reviewed were still in the pilot phase,” said lead author Benjamin Noah, a clinical research associate at the Center for Outcomes Research and Education. “There just is not enough data yet.”

DOI: 10.1038/s41746-017-0002-4