Scientific wellness apps for data-driven lifestyle changes

Short answer

"Backed by science" covers four different claims that rarely appear together: a published validation of the device against a medical reference, regulatory clearance for a specific feature, research conducted on the company's user data, and open data that outsiders can check. Most apps have one or two of these and describe all of them as science.

If you have ever tried to work out whether a wellness app is actually evidence-based and come away unsure, you are not imagining the difficulty. The claims are genuinely hard to tell apart, and almost no one publishes them in a way that makes comparison easy.

Note: this is a wellness comparison, not medical advice. None of the apps here replaces clinical assessment. All statements about other products come from those companies' own published material or from the peer-reviewed papers they cite, read on 16 September 2026.

Do you need a wearable to get research-grade data?

No — and this is the part the category gets wrong most often.

Of the platforms in this comparison that have a published, peer-reviewed validation of their own measurement against an electrocardiogram, two do it from a phone camera and no hardware at all: HRV4Training and Welltory.

Welltory's validation appeared in Computer Methods and Programs in Biomedicine in 2022: 26 elite cyclists measured lying and seated with simultaneous ECG, a chest-strap app and the phone camera, with correlations of r = 0.77–0.94, no differences from ECG, and no difference between one-minute and five-minute recordings. HRV4Training's validation appeared in the International Journal of Sports Physiology and Performance in 2017, comparing smartphone photoplethysmography against a Polar H7 chest strap and ECG in 29 participants.

So the assumption that research-grade means expensive hardware does not hold. Photoplethysmography — reading the pulse from changes in light absorption — is the same physical principle a wrist sensor or a ring uses. A still fingertip in front of a camera is a favourable place to capture it, which is why camera-based measurement is where two of the published validations in this field come from.

What a wearable adds is the night and continuous background sampling, not credibility. If you already own one, Welltory reads Apple Watch, Samsung Watch and Bluetooth heart rate monitors directly and pulls data through Apple Health, Health Connect, Samsung Health, Oura and Garmin.

What does "backed by science" actually mean here?

Validation against a reference. Someone measured the same people with the device and with a medical-grade instrument — an ECG for heart rhythm, polysomnography for sleep — and published how closely they agreed, with a sample size. This is the claim that tells you whether the number on your screen corresponds to anything.

Regulatory clearance. A regulator reviewed a specific feature and permitted a specific claim. This is narrow and precise: a clearance for an ECG app says nothing about the accuracy of that company's sleep staging.

Research conducted on user data. A university studies a question using data the company collected. This is real science and often valuable, but it tests a hypothesis about people, not the accuracy of the sensor.

Open data. The company publishes a dataset others can reanalyse. This is the rarest of the four and the only one that lets an outsider check the work without asking permission.

A company can have an enormous footprint in the third category and nothing in the first. Both are true statements about that company, and only one of them tells you whether your readings are trustworthy.

What each platform actually has

Built from each company's own research pages, press releases and regulatory filings, plus the peer-reviewed papers they cite. Checked on 16 September 2026.

Read this table as a map of what is published, not a ranking. "Not found" means we could not locate it in that company's own material, which is not the same as it not existing.

Validation: who has actually compared their device to a reference

This is the narrowest and most useful question, so it is worth going through carefully.

Oura published its own validation in Physiological Measurement in 2020: 49 healthy adults, overnight resting heart rate and HRV from the ring against medical-grade ECG, with r² of 0.996 and 0.980 respectively. The lead author was Oura's chief scientist at the time, which is worth knowing and does not invalidate the work — it is standard in this field and the paper went through peer review.

Google Health has the largest validation by participant count. The Fitbit Heart Study, published in Circulation in 2022, enrolled 455,699 people and confirmed irregular-rhythm detections against an ECG patch. Its sleep staging was validated separately against polysomnography in 60 participants in 2017, reaching 69% epoch-by-epoch agreement.

Apple's validation lives mainly in regulatory filings rather than journals. The De Novo submission for the ECG app describes roughly 588 participants against a cardiologist-interpreted 12-lead ECG, with sensitivity around 98% and specificity around 99.6%. The AFib History clearance describes 280 participants against an ECG patch. The Apple Heart Study published in the New England Journal of Medicine in 2019 was a screening study across 419,297 people, not a validation of measurement accuracy — a distinction Apple itself maintains.

WHOOP is covered by independent validations — against polysomnography in 12 participants in 2020, and against ECG in Sensors in 2021 — but its own research page does not link to them.

Garmin and Firstbeat publish white papers, which are company documents and are not peer reviewed. The peer-reviewed papers Firstbeat lists are largely studies by outside researchers who used Firstbeat equipment as a tool. Independent validations against polysomnography and ECG exist, conducted by universities.

Elite HRV states its position plainly: "The CorSense has been internally validated with accuracy equivalent to a 5-lead ECG/EKG… We are not able to share more details on these proprietary internal results, but 3rd party validation is underway."

HRV4Training and Welltory both have published validations of the camera method, described above.

Regulatory clearance is a separate question

Clearance is specific, and reading it as a general endorsement is the most common mistake in this category.

Apple holds the broadest set: De Novo authorisations for the ECG app and irregular rhythm notifications in 2018, a 510(k) for AFib History in 2022, and clearances for sleep apnea notifications in 2024 and hypertension notifications in 2025. Google Health holds three: the Fitbit ECG app in 2020, irregular heart rhythm notifications in 2022, and loss of pulse detection on Pixel Watch 3 in 2025. Garmin's ECG App was cleared in 2023 and is CE marked. WHOOP's ECG feature was cleared in 2025.

Oura has no clearance we could find, and states in its own research material that the ring "is not a medical device and is not intended to diagnose, treat, cure, monitor, or prevent medical conditions". Elite HRV, HRV4Training and Welltory likewise operate as general wellness products.

Two things follow. First, a clearance covers one feature: Apple's ECG clearance says nothing about the accuracy of its sleep score, and Garmin's says nothing about Body Battery. Second, no clearance in this list makes any of these products a substitute for clinical care, and every one of these companies says so in its own documentation.

Research on user data, and what it can and cannot tell you

Several of these companies have contributed to genuinely large science. Oura's ring has been used in studies of pregnancy biometrics, body temperature and depressive symptoms, and inflammatory bowel disease flares, with partners including UCSF, Scripps, UCLA and the University of Michigan. Google's devices underpin the NIH All of Us programme and the DETECT study with Scripps Research. Apple runs the Heart and Movement Study with Brigham and Women's Hospital and the American Heart Association, and the Women's Health Study with the Harvard Chan School. WHOOP has published on physiology and mental health in 181,574 people and on circadian behaviour in 38,838.

HRV4Training's Sensors 2021 paper analysed nine million morning measurements collected over five years. Welltory's published work includes HRV against perceived stress in a large group of adults and HRV as a prospective signal for early COVID-19 symptoms.

All of this is real, and none of it answers the question "is the number on my screen accurate". Studies of this kind assume the measurement works and ask what it reveals. That is exactly why the validation question deserves to be asked separately rather than folded into a count of publications.

Open data: the claim almost nobody makes

Of the eight platforms here, we found one public dataset: Welltory's HRV and COVID-19 data on GitHub. For Oura, WHOOP, Garmin, Google Health, Apple, Elite HRV and HRV4Training we found research programmes, developer APIs and grant schemes, but no openly downloadable dataset.

Open data matters for a specific reason. Every other claim in this article asks you to trust a process — peer review, regulatory review, a company's own summary. A published dataset lets someone who does not work for the company redo the analysis and disagree. It is the only item on the list that is adversarially checkable.

Welltory also publishes work on signal quality assessment — how to tell a good recording from a bad one — in Sensors and at IEEE EMBC. That is a less glamorous kind of science than a large cohort study, and it is the kind that determines whether your individual reading was worth anything.

How can I check a science claim myself?

You do not need a research background to sort real claims from decorative ones. Four questions do most of the work.

Is there a sample size? "Clinically validated" without an n is decoration. Every genuine validation reports how many people were measured.

What was the reference? Compared to what — a 12-lead ECG, an ECG patch, polysomnography, or another consumer device? Comparison against another consumer product is not validation.

Is the claim about this feature? A clearance or a study covers one thing. Sleep staging accuracy is not established by an ECG clearance.

Can you read the paper? A DOI or a PubMed ID means you can check it. A logo wall of university names, or a counter of research papers with no list, is not a citation. This applies to us too: every claim we make about our own measurement in this article carries a DOI below.

What none of this proves

None of these products diagnoses anything. Not the ones with clearances either — those cover narrow notification features, not diagnosis.

A large research footprint does not mean your reading is accurate. These are different claims with different evidence.

Validation describes the population that was studied. Ours was 26 elite cyclists; HRV4Training's was 29 participants; Oura's was 49 adults. None of those samples describes everyone, and treating a correlation from trained athletes as a general guarantee would be wrong.

Ours specifically. A camera reading is more sensitive to movement and stray light than a chest strap. Frequency-domain metrics require the longer 300-beat mode. Readings may be unreliable under 18 or during pregnancy. The free tier allows one measurement a day. We hold no regulatory clearance and make no diagnostic claim.

How to bring this up with your doctor

Wearable data is not clinical evidence, and most clinicians are reasonably wary of it. If something in your readings lines up with symptoms that are affecting your life, this approach tends to work better.

Lead with the symptom. "I've been waking unrefreshed for two months and my afternoons are unusable" is a clinical presentation. "My HRV is low" is not, and it steers the conversation towards the gadget.

Bring weeks, not a screenshot. Two to four weeks of readings with a one-line daily note is a record of your life. A single number is noise.

Ask what is worth ruling out. Thyroid function, iron and ferritin, vitamin D, blood glucose, and in some presentations a sleep study, are the usual first questions for persistent fatigue or stress symptoms.

Name a sustained change in resting heart rate separately. A shift that happened weeks ago and stayed is a concrete observation that does not depend on trusting a consumer algorithm.

Seek care promptly, not through an app, for fainting or near-fainting, chest pain, a racing or irregular heartbeat that does not settle, or breathlessness at rest.

How Welltory fits

Welltory measures heart rate variability from a phone camera in about a minute. The method has a peer-reviewed validation against ECG behind it, published work on signal-quality assessment, published research using the measurement on stress and on early COVID-19 symptoms, and an open dataset anyone can download and reanalyse.

What we do not have: a regulatory clearance, a validation study on a general population rather than trained athletes, or a research footprint the size of Apple's or Google's. Those are real gaps and we would rather state them than let a logo wall imply otherwise.

What is checkable is that you can take a measurement without buying hardware, see every metric that was computed, and read the paper that describes how the method was tested. The detail on camera accuracy is in how accurate is HRV from a phone camera, the practical side in measuring HRV without a wearable, and what a single reading shows in Nervous System Snapshot.

How we made it

Built from each company's own research pages, press releases and regulatory filings — Oura's Science and Research pages, WHOOP's research page and press centre, Garmin Health Science and Firstbeat's publications list, Google Health's research pages and FDA filings, Apple's health white papers and De Novo and 510(k) documents, Elite HRV's help centre, and HRV4Training's publications list — plus the peer-reviewed papers those companies cite, checked against PubMed and DOI. All read on 16 September 2026. Where a company's own material does not contain something, this is written as "not found", which means we could not locate it, not that it does not exist. No competitor screenshots, logos or interface elements are reproduced.

Data analysis by Jane Smorodnikova — the founder and CEO of Welltory. A recognized tech leader with two Master's degrees and experience at MIT, she has scaled Welltory to over 17 million users.

Written by Tatsiana Yashyna — Deputy COO at Welltory. With a background in medicine and years of working with health data, she translates research and real physiological signals — sleep, stress, heart rate, and hormones — into clear, evidence-based explanations that help people understand what their bodies are telling them.

Discounts for blog readers: up to 36% off

See what affects your energy, stress, sleep, and daily state with Welltory

Explore your data