Do Doctors Use Wearable Data? 97% Have Looked at It, 6% or Fewer Have It in Their Workflow

Short answer: nearly every doctor has looked at a patient’s wearable data, and almost none has it in the chart.

In early 2026 the American Medical Association and Medscape surveyed 2,222 physicians in the US, Canada, the UK, France, Germany and Spain. 97% said they review patients’ wearable data in some capacity. The share who have integrated it into clinical workflow, for example through direct import into the electronic health record, did not exceed 6% in any country [1][2].

That gap is the most important fact for anyone who wants wearable data to matter in healthcare. The problem is not that doctors are uninterested, and it is not mainly that patients are unwilling. It is that there is no routine place for the data to land, no clear payment for reviewing it, and no clear answer to who is responsible once it has been seen. This post covers what the survey found country by country, what the GAO added in August, what distinguishes the few physicians who have integrated wearable data, and what that means for companies building products that want to reach care.

What did the AMA survey find?

The survey was run by the AMA Center for Digital Health and AI with Medscape between January and early March 2026 [2].

• Review is near universal. Only 3% of physicians said they never review wearable data [2]. Between 85% and 90% in each country review it at least sometimes [3].
• Integration is rare everywhere. 6% in the US, 5% in Germany, 4% in Canada, 2% in the UK and Spain, 1% in France [3].
• Patients are asking. Fewer than one in four physicians overall get weekly patient requests to review data, but in the US about a third do [2][3]. About 30% of US physicians took clinical action at least weekly after reviewing wearable data [4].
• Physicians see value. Large majorities in and outside the US said wearable data offers at least some clinical advantage, with US figures reported in the high 70s [2][3][4].

What US physicians review most is revealing [3]:

Cardiologists and endocrinologists showed the highest adoption [1]. That is not an accident. Heart rhythm, blood pressure and glucose are where FDA-cleared consumer features and reimbursed remote monitoring already exist, which this blog has covered for sleep apnea and cuffless blood pressure.

What stops doctors from using wearable data?

US physicians named three concerns most often: medical-legal barriers (43%), accuracy (33%), and a belief that regulatory approval is essential for trust (44%) [3]. The AMA’s summary of the barriers is broader: unclear or missing reimbursement, workflow integration, regulatory and liability concerns, difficulty interpreting the data, limited clinical validation, and false positives [1].

The AMA’s chief executive put the core of it plainly in coverage of the survey: the data exists, but without a way into clinical workflow it is not actionable [4].

The more interesting finding is in the detail. When the AMA compared physicians who have integrated wearable data with those who have not, the integrated group reported higher perceived clinical value, stronger patient demand, more confidence interpreting the data, more trust in its accuracy, and fewer worries about false positives and liability. Regulatory trust and reimbursement, on their own, did not separate the two groups [2].

At the same time, the AMA found that differences between countries tracked reimbursement, workflow feasibility and regulation more closely than physician interest [1]. Put together, the picture is consistent: the system decides how many physicians can integrate, and confidence decides which ones do.

Why Germany expects to move faster

Germany is the outlier in expectations. 40% of German physicians expect to integrate wearable data within 12 months, against 18% in the US, with Spain at 38% and France at 33% [3].

The likely reason is payment. Germany’s DiGA pathway reimburses approved digital health applications, and 35% of German physicians had used it. About 10% of US physicians had used remote monitoring codes [4]. Only four of the six countries surveyed had any payment pathway for this work: Canada, France, Germany and the US [3]. The link between payment and adoption is the same one described in how governments are making wearable data a regulated input, where CMS, the NHS and the EU each attach data to a reimbursed service.

What did the GAO add?

In August 2026 the US Government Accountability Office published a technology assessment, Wearable Technologies: Potential Benefits and Challenges in Clinical Decision-Making (GAO-26-107847), based on literature, agency guidance, stakeholder interviews and site visits [5].

Benefits it found: faster diagnoses from more information, more personalized care plans and better adherence, and wider remote monitoring, particularly for rural and underserved patients [5].

Challenges it found [5][6]:

• Devices vary in accuracy and reliability, and wellness devices are not required to disclose how they were tested, which leaves clinicians unable to judge them.
• Wearables are not integrated into most facilities’ technology infrastructure or workflows, and poorly designed alerting can overwhelm clinicians.
• Coverage uncertainty is significant for manufacturers, and a slow, costly Medicare path pushes some toward wellness positioning instead.
• Liability uncertainty deters some products, and AI models can drift as real-world patterns change.

Three policy options: keep the status quo with voluntary testing by non-government groups; improve workflow integration and clarify clinicians’ responsibilities; or create a public database of wearables certified through independent testing, with incentives for manufacturers to participate [5]. The GAO notes that a certification database could improve testing but might confuse users about the difference from FDA review and disadvantage smaller manufacturers [6].

What does “integrated” actually require?

Integration is not one feature. It is four things that all have to exist at once.

1. A route into the chart. The most established one runs through the patient portal. In a Mount Sinai case report, a pulmonologist ordered peak-flow and inhaler data in Epic, patients shared it from Apple’s HealthKit through the MyChart app, and a reading below a preset threshold triggered a notification that led to a phone call and a new prescription [7]. The same report shows the other half of the problem: of four patients who agreed to share, only two sent more than one reading. The technical link has existed for years. What most practices have not done is decide which data types to accept, at what thresholds, who reviews the alerts, and how to keep patients sending.

2. A way to be paid. In the US that mostly means remote physiologic monitoring. From 1 January 2026 Medicare added 99445, device supply for 2 to 15 days of data in a 30-day period, and 99470, 10 to 19 minutes of management in a month, alongside the existing 16-day and 20-minute codes [8]. The device must meet the legal definition of a medical device, which rules out most consumer wellness features as the billed device [9]. There is no general code for “reviewed the patient’s smartwatch”.

3. A clear answer on responsibility. A clinician who accepts a data feed wants to know what they are expected to do with it and when. The GAO’s second policy option is explicitly about clarifying that [5].

4. Data a clinician can trust and read quickly. Which device, how accurate, compared with what, and does anything need attention. This is the part vendors control.

What product teams should build

The survey is, in effect, a specification for any product that wants its data to reach a clinician.

• Send summaries and flags, not streams. A clinician has seconds per patient. A weekly summary against a stated threshold is usable; a minute-level heart rate export is not. The same logic applies to how wearable data should flow through a pipeline.
• Label every value with its source and its accuracy. Say which device measured it and whether that measurement is FDA-cleared, a wellness estimate, or a notification. Physicians who integrate trust the data more; give them a reason to.
• Lead with measures that already have evidence. Rhythm notifications, cuff blood pressure, CGM glucose, sleep apnea notifications, activity. The data types physicians already review most are a good place to start.
• Keep false positives rare and actionable. Alert fatigue was named by both the AMA and the GAO [1][5]. Tune thresholds for specificity and say what to do next.
• Arrive through a pathway that already exists. An RPM program, an EHR integration, or a care team inside a payer or employer program. A new dashboard for a physician to log into is the least likely route to use.
• Expect time lags. Data that arrives hours late is not a real-time signal, which matters when thresholds route to an in-basket; see why wearable data arrives late.

The short version

97% of physicians in six countries review wearable data in some capacity, and no country has more than 6% with it integrated into clinical workflow [1][2][3]. Activity, heart data, alerts and sleep are what US physicians look at most [3]. Medical-legal risk, accuracy and trust are what they worry about [3]. The physicians who have integrated it trust it more and know what to do with it; the countries that expect to move fastest pay for it [2][3][4]. The GAO’s options, from better workflow to independent certification, point the same way [5]. For builders, the lesson is to make data a clinician can trust in thirty seconds, delivered through a route that already exists.