> ## Content Index > Fetch the complete content index at: https://compounded.ghost.io/llms.txt > Use this file to discover other available public pages before exploring further. # Your Wearable Wants to Read Your Hormones. Here's What It's Actually Measuring - URL: https://compounded.ghost.io/your-wearable-wants-to-read-your-hormones-heres-what-its-actually-measuring/ - Published: 2026-07-29T13:47:29.000Z - Updated: 2026-07-29T13:47:29.000Z - Author: Connor Hayes Consumer wearables spent a decade counting steps and heart rate. That era is ending. Dexcom's over-the-counter glucose sensor, once limited to adults, has recently been cleared for children as well, engineers at UC San Diego built a ring that reads glucose, ketones, and several other biomarkers directly from finger sweat, and a well-funded startup called Clair Health is building a wrist device that estimates hormone levels without ever drawing blood. The common thread is that wearables are shifting from watching what the body does to estimating what is happening inside it chemically. That shift matters for anyone treating health data as a genuine asset, because the reliability of that data now hinges on a distinction most marketing does not spell out: whether a device is directly measuring a molecule, or inferring it statistically from proxy signals like skin temperature and heart rate variability. Those are two very different claims wearing the same friendly wristband. ### From the Lab to the Ledger Direct chemical biosensing works the way a lab test does, just continuous and miniaturized. A glucose monitor's sensor sits in contact with interstitial fluid under the skin or with sweat at the surface, runs an electrochemical reaction against the target molecule, and reports an actual concentration. The UC San Diego sweat ring works on this same principle, reading several biomarkers, including glucose and ketones, straight off finger sweat in real time. Inferred sensing is a different animal entirely. Clair Health's wearable does not measure estrogen, progesterone, or other reproductive hormones directly. Instead, it feeds skin temperature, heart rate variability, electrodermal activity, sleep, and breathing patterns collected through ten onboard sensors into a machine learning model trained to associate those physiological signals with likely hormonal states. That is a pattern-recognition estimate, not a chemical assay, and the two should not be read with the same confidence. One reports what is physically present. The other reports a statistical best guess about what is probably happening, based on correlations the model has learned. ### Bio-Pipeline Ledger Over-the-counter continuous glucose monitors (Dexcom Stelo, Abbott Lingo, Libre Rio): commercially available and well-validated. Direct electrochemical glucose sensing, FDA-cleared for non-prescription use and recently expanded to include children who do not take insulin, the most clinically mature entry in this category. Multi-biomarker sweat-sensing rings (UC San Diego research prototype): early academic stage. Demonstrated simultaneous sweat-based readings across several chemical markers in a research setting, not yet a commercial product or independently validated device. Machine-learning inferred hormone tracking (Clair Health): early-stage, not yet independently validated. Uses a stack of physiological sensors and a trained model to estimate reproductive hormone patterns rather than measuring them directly, with a consumer launch planned but no published accuracy data against blood-draw testing yet. Wearable microneedle drug-level monitoring: preclinical. Early testing has shown real-time tracking of antibiotic levels in the body, aimed initially at closely monitored clinical settings, still well removed from everyday consumer or outpatient use. Wrist-worn ECG and irregular heart rhythm detection (established smartwatch features): commercially available and clinically validated as a screening aid, though explicitly framed by manufacturers and regulators as a flag for further evaluation, not a diagnostic replacement for a cardiologist. ### The Clinical Reality Check What is genuinely useful today is narrower than the wearable aisle suggests. Continuous glucose monitoring has real, validated clinical grounding behind it, built on years of use in people with diabetes before reaching a general wellness audience. Wrist-based heart rhythm screening sits in a similar category: a legitimate, regulator-reviewed early-warning tool, not a substitute for medical evaluation. Inferred measurements are a different story, and worth treating with real caution regardless of how confidently a product markets its output. A model trained to guess hormone levels from skin temperature and heart rate is not the same as a blood panel, and no company has yet published the kind of independent validation that would let a reader treat its numbers as clinical fact rather than a helpful, imperfect pattern. That does not make the technology worthless. Pattern recognition can be a genuinely useful early signal, prompting someone to look closer or track a trend over time. The realistic takeaway is to ask, of any wearable claim, whether it is reporting a direct chemical measurement or an inferred estimate, and to weight confidence accordingly. Direct sensing, where it exists and is regulator-cleared, deserves real trust. Inferred sensing deserves interest and patience, not the same certainty, and any meaningful health decision built on either kind of data still belongs in conversation with a qualified clinician rather than a wearable's app. ![](https://storage.ghost.io/c/93/20/932004ad-b501-4cef-8a02-28e1473c42cb/content/images/2026/07/wearables-inferred-vs-direct-biosensing-3-editorial.jpg)