The World Health Organization (2019) formally included burnout in the ICD-11 as an occupational phenomenon. The core problem hasn’t changed, though; emotional exhaustion doesn’t show up on a blood test. That gap between felt experience and measurable data is exactly where a new generation of health software is trying to operate. Some of it works. A lot of it doesn’t. This article breaks down what’s real.
What’s Actually Being Built Right Now
The mental health tech space isn’t just meditation apps anymore. Clinical-grade software, sensor-equipped wearables, and AI platforms that flag biometric patterns weeks before someone hits a wall are all either shipping or in late-stage trials.
Philips, Garmin Health, and Google’s Fitbit have been expanding their wellness APIs to include stress and recovery scoring. On the clinical infrastructure side, companies working at the intersection of healthcare IT and digital health, like DXC Technology (https://dxc.com/industries/healthcare-solutions), are increasingly involved in how health data pipelines are structured.
That infrastructure question matters more than it sounds. Raw data from a sleep tracker is useless without a clinical framework behind it. And building that framework properly is harder than most startup pitch decks suggest.
What Burnout Looks Like in Data
Burnout isn’t one signal. It’s a cluster of physiological and behavioral shifts that accumulate over weeks. The good news is that some of those shifts are measurable.
HRV — Overhyped and Still Useful
Heart rate variability measures the time variation between heartbeats. Higher variation generally means better recovery. Chronic stress flattens it. Whoop, Oura, and Apple Watch all track HRV in some form, and the underlying research, published in journals like Psychophysiology, is solid (Berntson et al., 1997). The catch? HRV drops for a dozen reasons: bad sleep, alcohol, overtraining, illness. One low reading means nothing. A downward trend over three or four weeks starts to tell a story.
Sleep as an Early Warning System
Most people don’t know that fragmented REM sleep is one of the earliest measurable signs of emotional overload. Oura’s Readiness score and Garmin’s Body Battery both weigh sleep stage data heavily. In practice, you’ll often see someone’s Readiness score sitting in the low 30s for a week before they consciously register that something’s wrong. Dreem, acquired by Beacon Biosignals, built a clinical EEG headband that measures sleep stages far more accurately than a wrist accelerometer ever could. It works. It also costs $400 and requires sleeping with electrodes on your head, which limits the market somewhat.
Cortisol: Still Mostly Prototype Territory
In healthy people, cortisol peaks in the morning and drops throughout the day. In burnout, that curve flattens or inverts. Measuring it continuously and non-invasively is the hard part. Biolinq in San Diego is developing a continuous sweat sensor that reads metabolite data alongside glucose. UC Berkeley researchers published work on a skin patch for cortisol tracking (Gao et al., 2016). Neither is commercially available yet, but both represent where the field is heading.
SaMD, ISO, and Why Regulatory Standards Actually Matter
Many apps that claim to help with burnout exist in a genuine regulatory gray area. The FDA category that clarifies this is SaMD (Software as a Medical Device), which covers software that performs medical functions such as detecting atrial fibrillation or assessing depression severity, even without being part of a physical device.
That triggers a specific set of medical device software standards:
- IEC 62304 – Governs the entire software development lifecycle for medical devices. Requirements, testing, risk management, and maintenance. If your software influences patient health outcomes, this is the baseline.
- ISO 13485 – Quality management for medical device organizations. Required to sell in the EU, US, Canada, and Japan.
- ISO 14971 – Risk management framework. Directly relevant when software can influence clinical decisions.
- IEC 82304-1 – The newer standard, written specifically with standalone health software and SaMD in mind.
Why Software Development for Medical Devices Is a Different Animal
When a regular app has a bug, you push an update, and some users are annoyed. But when medical device software has a bug, someone potentially receives the wrong clinical guidance. That’s why software development for medical devices involves formal risk analysis, traceability matrices, and validation documentation that feels foreign to most agile teams.
But many mental health apps want to be medical-grade but haven’t built their development processes to match. A proper SaMD development workflow ties every feature back to a clinical requirement, documents the associated risk, and validates the outcome before deployment.
What’s Actually on the Market
An honest breakdown of what exists:
Platforms with some clinical backing:
- Woebot – CBT-based conversational AI, published outcomes data in peer-reviewed journals, genuinely more rigorous than most competitors
- Koa Health’s Foundations – built with NHS guidance, closer to a validated digital therapeutic than a wellness app
- Mindstrong – used smartphone interaction patterns (typing speed, scrolling) to infer mood states before being acquired by Spring Health in 2022
Wearables relevant to burnout tracking:
- Oura Ring Gen 3 – sleep staging, HRV trends, temperature deviation, Readiness score
- Whoop 4.0 – recovery and strain modeling; started in sports performance, increasingly used in clinical pilots
- Garmin HRV Status -5-night rolling average displayed as a directional indicator; built into most mid-range Garmin devices
- Empatica Embrace Plus – research-grade, tracks electrodermal activity and HRV; used in clinical studies, not a consumer product
What none of these do well: connect physiological data to psychological context in a validated way, or explain why scores are dropping rather than just showing that they are.
The Passive Sensing Question
Some companies are trying to infer emotional state without any wearable at all. Ellipsis Health built voice biomarker technology that screens for depression and anxiety from a short voice recording; their clinical trial accuracy approached that of standard questionnaires. Sonde Health is doing similar work.
The acoustic features of speech change measurably under sustained stress: pitch variability, speech rate, and pause patterns. This isn’t new science; researchers have been publishing on voice analysis for mental health since at least 2014. What’s new is the computing power to run it on a phone in real time. It sounds like a privacy nightmare. And honestly, it can be, which leads to the next problem.
The Data Privacy Problem
Continuous physiological data is fine when it stays on your device. It becomes a different question entirely when it flows to a server, gets processed by a third-party algorithm, and potentially informs insurance assessments or employer wellness dashboards.
Amazon’s Halo tracker recorded ambient audio to analyze “vocal tone.” The backlash was fast and deserved. Several US states have since passed mental health data protection laws precisely because general health privacy legislation had left so many gaps (Klobuchar, 2020).
Here’s the odd part: the same app tracking your HRV to prevent burnout may have weaker legal protections around that data than your dentist has around your dental X-rays. HIPAA applies to covered healthcare entities; many wellness apps don’t qualify. GDPR in Europe is stricter, requiring explicit consent for health data processing (IAPP, 2023).
Where This Is Heading
The next two to three years will likely see the regulatory framework catch up to the technology. The FDA’s 2022 guidance on clinical decision support software drew clearer lines around when software becomes a regulated device (FDA, 2022). The EU’s MDR (Medical Device Regulation) substantially expanded and codified software regulation, and the IVDR covers digital diagnostic tools (European Parliament & Council of the European Union, 2017a).
That tightening of standards isn’t bad news for innovation — it’s actually what separates tools that genuinely help people from those that just feel like they do.
For anyone building in this space, the path forward looks like this: clinical validation first, not as an
afterthought. SaMD classification early in development. IEC 62304-compliant software lifecycle
processes baked in from the start, not retrofitted. And genuine transparency about what the data can
and cannot tell you.
For people dealing with burnout, the practical takeaway is simpler. A Garmin or an Oura Ring tracking your HRV for a few months probably won’t hurt, and might give you useful leading indicators before things get bad. But the app isn’t a replacement for an actual conversation with a therapist or a doctor. Digital tools are good at seeing patterns. They’re not good at meaning-making.
And meaning-making is still, for now, a human job.
Sources
- Berntson, G. G., Bigger, J. T., Eckberg, D. L., Grossman, P., Kaufmann, P. G., Malik, M., Nagaraja, H. N., Porges, S. W., Saul, J. P., Stone, P. H., & van der Molen, M. W. (1997). Heart rate variability: Origins, methods, and interpretive caveats. Psychophysiology, 34(6), 623–648. https://doi.org/10.1111/j.1469-8986.1997.tb02140.x
- European Parliament & Council of the European Union. (2017a). Regulation (EU) 2017/745 of the European Parliament and of the Council on medical devices [MDR]. Official Journal of the European Union. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32017R0745
- Gao, W., Emaminejad, S., Nyein, H. Y. Y., Challa, S., Chen, K., Peck, A., Fahad, H. M., Ota, H., Shiraki, H., Kiriya, D., Lien, D.-H., Brooks, G. A., Davis, R. W., & Javey, A. (2016). Fully integrated wearable sensor arrays for multiplexed in situ perspiration analysis. Nature, 529(7587), 509–514. https://doi.org/10.1038/nature16521
- International Association of Privacy Professionals. (2023, March 28). Filling the void? The 2023 state privacy laws and consumer health data. IAPP. https://iapp.org/news/a/filling-the-void-the-2023-state-privacy-laws-and-consumer-health-data
- Klobuchar, A. (2020, December 11). Following privacy concerns surrounding Amazon Halo, Klobuchar urges administration to take action to protect personal health data [Press release]. U.S. Senate. https://www.klobuchar.senate.gov/public/index.cfm/2020/12/following-privacy-concerns-surrounding-amazon-halo-klobuchar-urges-administration-to-take-action-to-protect-personal-health-data
- Nesse, W., Stuldreher, I. V., Kaneko, D., & Brouwer, A.-M. (2024). Acoustic and prosodic speech features reflect physiological stress but not isolated negative affect: A multi-paradigm study on psychosocial stressors. Scientific Reports, 14, Article 5398. https://doi.org/10.1038/s41598-024-55550-3
- U.S. Food and Drug Administration. (2022, September 28). Clinical decision support software: Guidance for industry and Food and Drug Administration staff. https://www.fda.gov/media/133660/download
- World Health Organization. (2019, May 28). Burn-out an “occupational phenomenon”: International Classification of Diseases. https://www.who.int/news/item/28-05-2019-burn-out-an-occupational-phenomenon-international-classification-of-diseases