Temperature Sensors and Fever Detection
What wearable temperature sensors measure, why a skin reading is not a thermometer reading, and what studies of fever detection with patches, rings and infrared devices show.
Content last updated . This page has not been medically reviewed.
Start here: how you feel counts
The NHS says a high temperature (fever) is usually a symptom of infection and often improves on its own, and that you may have one even if a thermometer reads below 38C if you feel hot, cold or shivery. [1] A wearable that shows no fever does not mean you are well. If you are unwell and treating it at home but it is not getting better or is getting worse, the NHS advises seeking urgent medical advice (in the US, contact your clinician or urgent care). [1] For a child or anyone with severe symptoms, seek care rather than checking a device first.What these sensors measure
Wearable temperature sensors sit against the skin; the hospital devices in the review below were worn at the armpit, chest, wrist or upper arm. Skin temperature is a direct reading of the skin surface. Any "body temperature" or "fever alert" is an estimate produced by software from skin readings, often combined with heart rate and sleep data. The NHS notes that normal body temperature differs from person to person and changes during the day, and uses 38C or above as the usual threshold for a high temperature. [1]| Type | Direct or derived | What the evidence covers |
|---|---|---|
| Skin temperature reading | Direct (of the skin) | Not the same as an oral or armpit thermometer reading |
| Estimated body temperature | Derived by an algorithm | Studied mostly in hospital patches |
| Fever alert or illness flag | Derived, proprietary | Studied in surveillance research, not as personal diagnosis |
What studies show
Hospital and clinic patches. A scoping review of wearable temperature monitoring in acute care hospitals included 29 publications from 26 studies, covering 18 devices worn at the axilla, chest, wrist or upper arm. It concluded that clinical integration is at an early stage of maturity. [2] In one research study of 86 people at high risk of fever after chemotherapy and stem cell transplant, an armpit patch estimated temperature with an error of 0.35 ± 0.88°F against oral readings and detected clinic-assessed fevers with 90.2% sensitivity and 87.8% specificity. The algorithms were tested by cross-validation within the same dataset. [3] The population, device and setting are narrow, so these numbers do not apply to a ring or watch worn by a healthy person.Rings and wearable data for fever flags. A 2024 analysis used data from 63,153 participants of one smart-ring product: skin temperature, heart rate and sleep from nights around a day. It trained a model to detect self-reported fevers (724 fever days from 463 people). The model had an area under the curve of 0.85 and an average precision of 0.25; at 50% sensitivity the false positive rate was 0.8%. Fever was self-reported with a self-measured temperature and the model was cross-validated. The authors framed this as potentially useful for public-health surveillance, not for personal diagnosis. [4] Label: promising but limited, for population signals only.
Non-contact infrared thermometers. A meta-analysis of 30 studies (19 pooled) compared forehead infrared thermometers and thermal scanners with conventional thermometers. Pooled sensitivity was 0.808 and specificity 0.920 for forehead thermometers, and 0.818 and 0.923 for thermal scanners. Positive predictive value was low early in an outbreak, and accuracy varied with the setting, subjects and reference method. [5] These are handheld or camera devices, not wearables.
Consumer wrist and finger sensors. Proper evidence not found: the PubMed, Europe PMC and web searches run for this page on 2026-10-22 turned up no independent peer-reviewed study that validated a consumer watch or ring sensor against a reference thermometer for detecting fever in individuals. No regulator statement on the topic was read. A company's own claims are a marketing claim.
Accuracy is not benefit
Even where a device estimates temperature well, no source read for this page shows that home wearable fever alerts lead to better health outcomes in healthy people. Body temperature itself changes during the day [1], which can make a single reading hard to interpret; see baseline variation and normal fluctuation.Who should be careful
- People caring for infants or older relatives, or anyone at high risk of infection, should use a proper thermometer and follow their clinician's advice. The NHS describes using a digital thermometer for a child's armpit reading. [6]
- Do not delay care or change medicine because a ring or watch shows a normal temperature.
How to judge a claim
Ask what the device measures directly, which reference thermometer and site (oral, armpit, rectal) it was compared with, in whom, and whether the claim is "trend" or "fever". See smart rings: what they measure and estimate, algorithm-derived scores and consumer data versus clinical measurements.Frequently asked questions
Can a smart ring or watch detect a fever?
Not reliably for an individual, as far as the sources read for this page show. Research on ring data targets population surveillance. [4] No independent validation of consumer devices was found.
Is wrist skin temperature the same as body temperature?
No. The device reads the skin; body temperature is estimated. The patch study compared armpit estimates with oral readings, a different setup from a wrist sensor. [3]
Are forehead or infrared thermometers accurate?
Pooled sensitivity was about 81% and specificity about 92% for forehead infrared thermometers against conventional thermometers, with variation by setting. [5]
What temperature counts as a fever?
The NHS usually counts 38C or above as a high temperature, while noting that normal temperature varies by person and through the day. [1]
Should I buy a wearable to watch for illness?
That is a personal choice, and no outcome evidence for it was found. The NHS says you do not need a thermometer to know you have a high temperature, but you can use one, and it describes how to use a digital thermometer. [1][6]
More details: evidence, limits and cautions
The details behind this page: cautions, limits of the evidence and where sources disagree. Open any note to read it.
How to read the evidence labels
Labels used on this page: established, promising but limited, mixed, insufficient evidence, Proper evidence not found, marketing claim.
References
- NHS (UK), High temperature (fever) in adults. Read 2026-10-22.
- Continuous Temperature Monitoring via Wearable Devices for Fever or Infection Management in Acute Care Hospitals: Scoping Review of Clinical Implementation and Data Analytics, JMIR Mhealth Uhealth 2026 (PMID 42727075). Read 2026-10-22.
- Verma N et al., A Novel Wearable Device for Continuous Temperature Monitoring and Fever Detection, IEEE J Transl Eng Health Med 2021. Read 2026-10-22.
- Kasl P et al., Utilizing Wearable Device Data for Syndromic Surveillance: A Fever Detection Approach, Sensors (Basel) 2024. Read 2026-10-22.
- Diagnostic accuracy of non-contact infrared thermometers and thermal scanners: a systematic review and meta-analysis, J Travel Med 2020 (PMID 33043363). Read 2026-10-22.
- NHS (UK), High temperature (fever) in children, last reviewed 03 January 2024. Read 2026-10-22.