Your doctor’s office now wants your Apple Watch data
Health systems are pulling wearable data into patient records. The evidence for what that improves is thinner than the pitch.
By The Weekend · · 5 min read
Google's health API now lets wellness apps push step counts, heart rate and sleep data straight into a user's medical record, joining a similar pipeline Apple built years ago. The pitch is continuity of care. The evidence that continuous wearable data changes a clinical outcome, rather than just a chart, covers a much shorter list of conditions than the marketing suggests.
What actually happened
A wellness platform announced an integration with Google's Health Connect API, the plumbing that lets apps on Android phones share fitness and health data with each other and, increasingly, with clinical systems. This is not a new idea. Apple's HealthKit has offered comparable data-sharing since 2014, and Health Connect, which Google folded into Android as a system-level service in 2023, was built to do the same job for the wider Android ecosystem, which runs on roughly 2.5 billion active devices worldwide according to Google's own developer figures.
What's changed is who's on the receiving end. Early wearable integrations mostly shuttled data between consumer apps — a run tracked in Strava showing up in a sleep app, say. The current wave connects that same data to hospital systems and primary-care platforms, so a patient's resting heart rate trend or step count can sit inside the electronic health record a physician actually opens. Several large U.S. health systems have piloted this over the past two years, feeding data from Fitbit, Apple Watch and Oura into patient portals built on Epic's MyChart, the record system used by more than 250 million patients in the U.S.
Who pays, who gains
The wearable companies gain a stickiness argument: a device that talks to your doctor is harder to abandon than one that just counts steps. Apple and Google gain platform lock-in, since the health data lives inside their operating systems' frameworks even when a third-party app displays it. Health systems gain a marketing line — "connected care" — and, in some pilot programs, a genuine reduction in phone-tag: one 2023 Kaiser Permanente pilot using remote blood-pressure and weight monitoring for heart-failure patients reported 30% fewer readmissions over 90 days compared with usual care, a result consistent with a 2021 Cochrane review of telemonitoring in heart failure that found a similar reduction across pooled trials.
Patients pay in a currency that's easy to miss: data-sharing consent that's broader than most people read closely. Health Connect requests permission per data type, but the default flow nudges toward "allow all," and a 2022 Mozilla Foundation review of fitness-app privacy policies found that most reserved the right to share aggregated or de-identified data with third parties, including advertisers, regardless of the medical integration layered on top.
The mechanism
The clinical case for wearable integration rests on a narrow set of measurements that are both accurate and actionable. Heart rate and step count, tracked over weeks, are decent proxies for cardiovascular fitness and are cheap to collect continuously — something a clinic visit every six months cannot replicate. That's the strongest use case: continuous data catching a trend a single office reading misses. The 2021 Apple Heart Study, run with Stanford and covering over 400,000 participants, found the Apple Watch's irregular-rhythm notification correctly flagged atrial fibrillation in 84% of cases where a follow-up ECG patch confirmed it — useful, but also a reminder that 34% of people who got a notification never had an issue the patch could confirm, meaning a meaningful share of alerts sent healthy people into unnecessary follow-up testing.
Sleep staging, the data type wellness apps push hardest, is the weakest link in this chain. Consumer wearables estimate sleep stages from movement and heart-rate variability, not from the brainwave monitoring that clinical sleep labs use, and validation studies comparing wearables against polysomnography — the lab-grade standard — consistently find the devices overestimate total sleep time and misclassify REM versus deep sleep in a meaningful minority of nights. Feeding that data into a medical record doesn't make it more accurate; it just gives an inaccurate number a more official home. A physician looking at a "sleep score" inside a patient chart is looking at the same movement-based estimate a smartwatch always produced, now wearing a lab coat it hasn't earned.
The actual mechanism that makes any of this useful clinically is narrower than the interoperability pitch: it's not that more data flows into the record, it's that specific, validated data types — blood pressure, glucose for diabetics, heart rhythm for people with known arrhythmia risk — get monitored continuously instead of episodically, and a clinician has a protocol for what reading triggers what action. Without that protocol, the data is noise with a nice interface.
What happens next
Expect health systems to keep integrations narrow even as the platforms make broad sharing technically trivial. The FDA has cleared wearable-based ECG and blood-oxygen features under specific 510(k) pathways for defined conditions — atrial fibrillation detection being the main one — and clinicians are more likely to act on data with that clearance than on a wellness app's proprietary "readiness score." Watch for insurers, not hospitals, to be the more aggressive adopters: several U.S. health plans already offer premium discounts for hitting step or activity targets tracked through these same APIs, which raises a separate, unresolved question about whether continuous personal health monitoring becomes a pricing input rather than a care tool. If that shift accelerates, the interoperability story stops being about doctors and starts being about actuaries.
FAQ
Does sharing my wearable data with my doctor actually improve my care?
For a specific list of conditions — heart failure, hypertension, diagnosed arrhythmia, diabetes — continuous monitoring has trial evidence behind it, including the 30% readmission reduction Kaiser reported and the pooled Cochrane findings for heart failure. For general fitness and sleep tracking shared with a primary-care doctor who has no specific protocol for it, there's little evidence it changes what happens in your appointment.
Who actually sees this data besides my doctor?
Depends on your permission settings and the app's privacy policy. Health Connect and HealthKit both allow granular, per-data-type consent, but most fitness apps' policies, per the 2022 Mozilla review, retain rights to use de-identified or aggregated data for purposes beyond your care, including product analytics and, in some cases, advertising partners.
Is my sleep score from my watch something a doctor would trust?
Not as a diagnostic tool. Consumer wearables estimate sleep stages from movement and pulse, not brain activity, and validation studies against clinical polysomnography find real gaps in accuracy, particularly for REM and deep-sleep classification. A doctor concerned about a sleep disorder will still order a lab-grade sleep study rather than read your app's dashboard.