No evidence that wearables improve patient outcomes
Evidence is still lacking on whether remote patient monitoring using wearable biosensors can improve patient clinical outcomes such as blood pressure (BP), body mass index (BMI), or weight, suggests a meta-analysis, despite the rising popularity of wearables among the general public.
“As of now, we don’t have enough evidence that they consistently change clinical outcomes in a meaningful way,” said study principal investigator Dr Brennan Spiegel, director of Health Services Research at Cedars-Sinai Medical Center, Los Angeles, California, US. “But that doesn’t mean they can’t.”
The meta-analysis included 27 randomized controlled trials (mean duration, 7.8 months [range, 7 days to 29 months]) involving 6,453 participants and focusing on a range of clinical outcomes, which were broadly categorized into BMI, weight, waist circumference, body fat percentage, and systolic as well as diastolic BP. Treatment arms received noninvasive wearable biosensors while the control arms mostly offered education together with standard care but without wearable devices. Studies were weighted by sample size and treatment effects between groups were analysed using difference-in-differences point estimation. [NPJ Digital Med 2017;doi:10.1038/s41746-017-0002-4]
Compared with control, the researchers found no significant impact of wearable biosensors on any of the six reported outcomes: BMI (-0.96, 95 percent confidence interval [CI], -2.30 to 0.37), weight (-1.29, 95 percent CI, -3.06 to 0.48), waist circumference (-2.41, 95 percent CI, -5.16 to 0.34), body fat percentage (0.19, 95 percent CI, -1.2 to 1.57), systolic BP (-2.62, 95 percent CI, -5.31 to 0.06), and diastolic BP (-0.74, 95 percent CI, -2.34 to 0.86).
Nonetheless, based on systematic review of high-quality studies, these wearables did demonstrate early promise in improving patient outcomes for select conditions, such as hypertension, obstructive pulmonary disease, low back pain, and Parkinson's disease.
Also, interventions based on personalized coaching, validated care pathways, and targeting health behavioural change were more likely to be successful, while cash incentives and text messages notification were ineffective.
“There is a big difference between using these sensors to track sleep for self-betterment and using them make medical decisions,” said study co-author Dr Michelle Keller also from Cedars-Sinai Medical Center.
As behaviour change is complex and that most devices resulted in only short-term behavioural changes based in previous studies (for eg, 3–6 months for activity trackers), [iConf Proc 2015; Shih et al] Spiegel and co-authors noted that “understanding how and if specific devices and device related interventions and incentives motivate health behaviour change is an important area that is still not well understood.”
The researchers also acknowledged that the heterogeneity in study design, patient outcomes, and device type as well as paucity of high-quality studies were main limitations of the meta-analysis. Sixteen of the studies included were considered “high quality” while the other 11 were “low quality”.
“[O]ur results indicate that while some RPM interventions may prove to be promising in changing clinical outcomes in the future, there are still large gaps in the evidence base … that should be considered before implementation of remote patient monitoring in the clinical setting,” said Spiegel and co-authors.