Can your wearable detect you're getting sick before you feel it?

This is a different question from what to do once you already feel sick and are deciding when to train again — that's a training-decision guide for during and after illness. This one is about the narrower, more interesting claim: can the data move before you notice anything at all, and how much should you trust it if it does.

What actually moves first, and why

An infection triggers an immune response before it produces symptoms you'd notice. Cytokines released early in that response nudge the autonomic nervous system toward a more sympathetic (stress-like) state, which shows up in wearable-measurable ways:

• Resting heart rate rises — the heart works a little harder even before fever sets in.
• HRV falls — reduced parasympathetic ("rest and digest") activity relative to sympathetic drive.
• Overnight respiratory rate and skin temperature rise — usually a bit later, often more clearly once symptoms are already present.

None of this is unique to infection — the same signature shows up after alcohol, heavy training, heat, poor sleep, or dehydration, which is one of the several reasons your HRV baseline itself can shift. What makes an illness signal identifiable in research isn't any single reading; it's a combination of metrics moving together and persisting without one of those obvious explanations.

How much warning do you actually get — what the research found

Most of the strongest evidence here comes from studies run during the COVID-19 pandemic, simply because it was the first time large numbers of wearable users had a known infection date to measure back from. The findings are still the best available evidence for respiratory illness generally, though direct data for common colds and ordinary flu is thinner.

The honest summary: a lead time of one to four days before symptoms is the more typical finding across individuals who show a signal at all; the widely quoted "up to 7-9 days" figures come from the fastest-detected participants within these studies, not the average person. And a meaningful fraction of infected people in these same studies showed no clear signal before symptoms started.

How good is "good enough" to trust

63% real-time detection (from the best-performing study above) means roughly 4 in 10 actual cases weren't caught before symptoms in that dataset — a real ceiling on how much you should lean on this. It also doesn't distinguish COVID from flu from a common cold from a bad night of drinking; a wearable is detecting a stress response, not a pathogen. The sensor doesn't know why your nervous system shifted, only that it did.

That's why the practical use case isn't "trust the alert like a test result." It's closer to a smoke detector with a real but imperfect hit rate: worth a second look, not worth panicking over on its own, and definitely not a substitute for testing when a real test is available and the stakes matter (travel, an immunocompromised household member, a positive contact).

A simple protocol for reading an early flag without overreacting

1. Wait for a second data point, not just one off morning. A single low-HRV or high-resting-heart-rate night has dozens of ordinary causes.
2. Check whether two or more metrics agree. Resting heart rate up and HRV down and, if visible, respiratory rate creeping up — together is a stronger signal than any one alone.
3. Rule out the obvious explanations first. Hard training session, alcohol, poor sleep, heat, travel, or a stressful day can all produce the same shape of change for one or two nights.
4. If the pattern holds for 2+ mornings with no obvious cause, treat it as a real flag. Test if you have access to a test, scale back demanding plans, and prioritize sleep and hydration.
5. Don't wait on the data if symptoms actually show up. Once you feel sick, the wearable's job is done — switch to a training and recovery decision, not a detection one.

What it genuinely can't do

• It can't identify which illness, or confirm you're contagious, or replace a test.
• It can't reliably separate "getting sick" from "trained too hard, drank, or slept badly" on a single night — only the combination and persistence do that.
• It works best against your own multi-week baseline, not a population norm — someone with a naturally low resting heart rate and someone with a naturally high one need very different absolute thresholds, which is part of why cross-device or cross-person comparisons of raw numbers are unreliable.
• Skin temperature specifically has real value for ovulation-timing and fever-pattern detection, but wrist/finger sensors read skin temperature, not core body temperature — a true fever can be present without the wearable showing a dramatic jump.

When to stop watching your wearable and call a doctor

None of the above should delay medical attention for actual symptoms. See a doctor rather than waiting on data if you have a fever with chest pain, palpitations, unusual shortness of breath, confusion, or symptoms lasting more than about 10 days without improvement. A wearable trend is a nudge toward testing or resting earlier — it is never a reason to defer care once you're genuinely unwell.

What Vita already does with this

Vita's recovery score already tracks HRV, resting heart rate, and — where available — overnight respiratory rate against your own rolling baseline, and can send a proactive alert when respiratory rate runs clearly above your personal norm or HRV stays suppressed across several nights, which is close to the multi-metric, persistence-based read the research above suggests actually works. The AI coach can also walk through a flagged morning with you — is this two nights running, does resting heart rate agree, was there a plausible explanation yesterday — instead of leaving you to eyeball three separate charts. It's built to catch your attention a day or two earlier, not to replace a test when one actually matters.