Wearable studies lean on activity and heart rate, review of 147 finds

The gap matters for anyone who reads a wearable study as proof of a health benefit. The review maps what researchers measured. It does not show whether that measurement changed outcomes.

Hospitals, insurers and device makers already sell wearable data as clinical infrastructure. The BMJ Open review screened 13,744 records and landed on 147 original articles. It describes a field that is broad but methodologically thin. Most studies rely on device-derived measures and conventional analysis, not composite digital biomarkers. It does not establish whether wearable-guided care improves outcomes.

A map of 147 wearable studies

Cardiovascular diseases were the most studied indication, followed closely by neurological diseases. Median sample size was 85. The range ran from 1 to 455,699.

Physical activity was measured in 106 of the 147 studies, or 72 percent. Heart rate appeared in 48 (33 percent) and sleep in 34 (22 percent). Ninety-eight studies used wearable data to measure outcomes: 52 primary and 46 secondary. A total of 49 used it for disease monitoring. The authors searched Scopus, Web of Science, the Cochrane Library, PubMed and ClinicalTrials.gov. They looked for studies published or registered between January 2020 and July 2026.

The authors ran no formal quality appraisal or risk-of-bias assessment, consistent with scoping review method. That limits what the review can say about the strength of that evidence.

The analytics gap

Only 27 of the 147 studies applied advanced computational or machine learning methods. Nine integrated multiple wearable-derived variables into composite or disease-specific digital biomarkers.

The authors' conclusion is blunt: most studies relied on device-derived measures and conventional analytical methods. That is narrower than saying wearables do not work. It sits alongside a second review with a similar tone. A MASCC systematic review of digital health in community end-of-life cancer care found 24 publications covering 23 unique studies. The studies spanned mobile apps, video conferencing, wearables, electronic health record tools and web platforms. It called the interventions generally feasible and acceptable to patients, caregivers and clinicians. But it said the evidence remains heterogeneous and in its early stages.

Hospitals and vendors are already pitching wearable data

Garmin announced its 2026 Health Awards at a summit in Bangkok on 29 October. The awards drew 113 applications from 31 countries.

Garmin named six finalists in June: Asahi Quality & Innovations, HeraMED, Montefiore Einstein, Phyathai Paolo Hospital, Thaivivat Insurance and UPTIVO. Montefiore Einstein, a US cancer center, took the Expert's Choice award. It used Garmin wearables in prospective trials of patients receiving radiotherapy, chemotherapy and immunotherapy. Garmin said its researchers were the first to show that wearable data collected before treatment can identify patients likely to struggle with therapy. HeraMED, an Australian company, won the People's Choice award for AI-enabled women's health care across pregnancy, postpartum, menopause and chronic conditions. Each winner received Garmin smartwatches and fitness trackers worth up to $10,000. The second day of the summit drew more than 200 executives and industry leaders.

Garmin's own announcement describes the Montefiore work as demonstrating feasibility, not as a controlled outcome result.

What a more intensive study looks like

The TIME Study at the Buck Institute for Research on Aging will enroll up to 150 adults aged 55 and older in stable health.

Its protocol, published in Frontiers in Digital Health, combines weekly blood sampling and four-day bursts of near-continuous molecular profiling. It also combines standardised physiological challenges and continuously recording wearables. Participants complete an 11-week main phase with nine weekly visits: seven standard and two challenge visits. Four high-frequency sampling days fall on consecutive Thursdays, Fridays, Mondays and Tuesdays. Follow-up visits come at 6 and 12 months after enrollment.

What comes next

The study is registered as NCT07107386. It began enrollment on 15 June 2025. The paper reports planned methods, not results. No findings about aging biology have come from the cohort yet.

Digital health data will be collected throughout the main and follow-up periods. The first paired comparisons between wearable signals and molecular measurements are still ahead. The BMJ Open review's search closed in July 2026. That leaves newer trials outside its count of 147.