CoChat
Sign Up
Cited by 133Non-Invasive Vital Sign MonitoringInfrared Thermography in MedicineData-Driven Disease Surveillance

The DOI matches its Crossref record; 31 of the 31 references it cites were found in a registry or an index.

Feasibility of continuous fever monitoring using wearable devices

Open access
Smarr, Benjamin L., Aschbacher, Kirstin, Fisher, Sarah M., Chowdhary, Anoushka, Dilchert, Stephan, Puldon, Karena, Rao, Adam, Hecht, Frederick M., Mason, Ashley E.·2020·Scientific Reports
DOI 10.1038/s41598-020-78355-6PMC PMC7736301

Published version of a preprint

Open this paper in CoChat — free accountOpen PDF

Abstract

Elevated core temperature constitutes an important biomarker for COVID-19 infection; however, no standards currently exist to monitor fever using wearable peripheral temperature sensors. Evidence that sensors could be used to develop fever monitoring capabilities would enable large-scale health-monitoring research and provide high-temporal resolution data on fever responses across heterogeneous populations. We launched the TemPredict study in March of 2020 to capture continuous physiological data, including peripheral temperature, from a commercially available wearable device during the novel coronavirus pandemic. We coupled these data with symptom reports and COVID-19 diagnosis data. Here we report findings from the first 50 subjects who reported COVID-19 infections. These cases provide the first evidence that illness-associated elevations in peripheral temperature are observable using wearable devices and correlate with self-reported fever. Our analyses support the hypothesis that wearable sensors can detect illnesses in the absence of symptom recognition. Finally, these data support the hypothesis that prediction of illness onset is possible using continuously generated physiological data collected by wearable sensors. Our findings should encourage further research into the role of wearable sensors in public health efforts aimed at illness detection, and underscore the importance of integrating temperature sensors into commercially available wearables.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/s41598-020-78355-6

Cites

31 of 31 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1 Wearable distal temperature sensors are suitable for developing digital biomarkers for fever with and without paired symptom reports. 50 individuals (a) display a wide range of temperatures (inter-individual means ± standard deviation) both during baseline (dark blue) and symptom report window (pail lines); there is a small but significant difference between the two sets of means. Having been normalized (norm) by individual range (see “Methods”), the mean 65-day temperature profile for individuals reporting fever (purple) and those not reporting fever (orange) reveal an apparent rise in maximum finger temperature (T) in fever-reporting cases near the beginning of the symptom window (black arrow; symptom window: grey box, mean duration for the whole population). Norm daily maximum (c) and minimum (d) T highlights changes in both that correspond to fever onset report. Such cases informed the construction of digital biomarkers in the form of thresholds for daily max and min that identify fever-like days. An example T record (e) with fever-like days identified by exceeding these thresholds before onset of symptom report (e; black dots represent daily min and max above thresholds) has similar changes in heart rate (HR), HR variability (HRV), and respiration rate (RR), to the reported fever event (f by variable, and g,h with overlay, respectively). (f–h) All lines are smoothed by 360 min radius, displaying the same smoothing used to generate median minimum and maximum values for each day. Faded blue line in (g,h) is the raw T (1 point/min).
Figure 1 Wearable distal temperature sensors are suitable for developing digital biomarkers for fever with and without paired symptom reports. 50 individuals (a) display a wide range of temperatures (inter-individual means ± standard deviation) both during baseline (dark blue) and symptom report window (pail lines); there is a small but significant difference between the two sets of means. Having been normalized (norm) by individual range (see “Methods”), the mean 65-day temperature profile for individuals reporting fever (purple) and those not reporting fever (orange) reveal an apparent rise in maximum finger temperature (T) in fever-reporting cases near the beginning of the symptom window (black arrow; symptom window: grey box, mean duration for the whole population). Norm daily maximum (c) and minimum (d) T highlights changes in both that correspond to fever onset report. Such cases informed the construction of digital biomarkers in the form of thresholds for daily max and min that identify fever-like days. An example T record (e) with fever-like days identified by exceeding these thresholds before onset of symptom report (e; black dots represent daily min and max above thresholds) has similar changes in heart rate (HR), HR variability (HRV), and respiration rate (RR), to the reported fever event (f by variable, and g,h with overlay, respectively). (f–h) All lines are smoothed by 360 min radius, displaying the same smoothing used to generate median minimum and maximum values for each day. Faded blue line in (g,h) is the raw T (1 point/min).CC BY · PubMed Central
Figure 2 Elevated temperature is clearly detectable, but only loosely correlated to changes in HR, HRV, and RR. All time points for all individuals with original (not interpolated) measurements from temperature (T), heart rate (HR), HR variability (HRV), and respiration rate (RR) sorted into those not showing fever-like days during their reported symptom window (a) and those showing fever-like days during their reported symptom windows (b). Baseline points: white. Symptom windows: color (bar in lower left, b, proportional to T, max set to threshold of digital biomarker for nighttime fever detection). Dark background points: symptom window from opposite group for comparison. Dot size proportional to RR. Insets: symptom window (dark) and baseline (white) points for that panel. NB: despite being normalized by the mean of baseline day min and max values as described for temperature (see “Methods”), HR and HRV show wide ranges, reflecting large day-to-day variance within individuals. Larger and redder spots are more apparent in the “fever-associated excursion” but not restricted to this region, reflecting the lack of strong correlation across variables. Both panels use identical axes as well as scales for color and size.
Figure 2 Elevated temperature is clearly detectable, but only loosely correlated to changes in HR, HRV, and RR. All time points for all individuals with original (not interpolated) measurements from temperature (T), heart rate (HR), HR variability (HRV), and respiration rate (RR) sorted into those not showing fever-like days during their reported symptom window (a) and those showing fever-like days during their reported symptom windows (b). Baseline points: white. Symptom windows: color (bar in lower left, b, proportional to T, max set to threshold of digital biomarker for nighttime fever detection). Dark background points: symptom window from opposite group for comparison. Dot size proportional to RR. Insets: symptom window (dark) and baseline (white) points for that panel. NB: despite being normalized by the mean of baseline day min and max values as described for temperature (see “Methods”), HR and HRV show wide ranges, reflecting large day-to-day variance within individuals. Larger and redder spots are more apparent in the “fever-associated excursion” but not restricted to this region, reflecting the lack of strong correlation across variables. Both panels use identical axes as well as scales for color and size.CC BY · PubMed Central
Figure 3 Use of signal-processing allows identification of more complex digital biomarkers in continuous data. Wavelet frequency decomposition (a) provides a frequency-by-time surface usable for feature extraction from continuous data. We found that 226/244 daytime fever-like events as detected by digital biomarker (d, aligned by onset) are preceded by a relative peak in the power of the ~ 22–26 h frequency band (b, all instances aligned by peak; mean: lighter fat line; median: darker fat line); example peak and alignment to fever-like day 2 days later (a, above to below; dots are temperature minimums for days with daytime fever threshold exceeded). The relative height of the wavelet peak correlates loosely but significantly (r = 0.36, p = 1 × 10–7) with time to fever onset (c). Hollow dots: points from individuals without detected fever-like days during reported symptom window; filled dots: points from individuals with detected fever-like days during their reported symptom window. Transparency highlights where points overlap and cluster.
Figure 3 Use of signal-processing allows identification of more complex digital biomarkers in continuous data. Wavelet frequency decomposition (a) provides a frequency-by-time surface usable for feature extraction from continuous data. We found that 226/244 daytime fever-like events as detected by digital biomarker (d, aligned by onset) are preceded by a relative peak in the power of the ~ 22–26 h frequency band (b, all instances aligned by peak; mean: lighter fat line; median: darker fat line); example peak and alignment to fever-like day 2 days later (a, above to below; dots are temperature minimums for days with daytime fever threshold exceeded). The relative height of the wavelet peak correlates loosely but significantly (r = 0.36, p = 1 × 10–7) with time to fever onset (c). Hollow dots: points from individuals without detected fever-like days during reported symptom window; filled dots: points from individuals with detected fever-like days during their reported symptom window. Transparency highlights where points overlap and cluster.CC BY · PubMed Central
Figure 4 CONSORT (consolidated standards of reporting trials) format flow chart for exclusion of cases from analyses. Although neither this effort nor TemPredict are clinical trials, we show exclusion of participants from the larger dataset for ease of interpretation.
Figure 4 CONSORT (consolidated standards of reporting trials) format flow chart for exclusion of cases from analyses. Although neither this effort nor TemPredict are clinical trials, we show exclusion of participants from the larger dataset for ease of interpretation.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 30 references

Cited by

Sorted by citations

Items 101–133 of 133

Related

See all

Metadata

Authors
Smarr, B. L., Aschbacher, K., Fisher, S. M., Chowdhary, A., Dilchert, S., Puldon, K., Rao, A., Hecht, F. M., & Mason, A. E.
Affiliations
  • University of California San Diego (US)
  • University of California, San Francisco (US)
  • Baruch College (US)
Corresponding author
Smarr, Benjamin L.
ORCID
Venue
Scientific Reports
Publisher
Springer Science and Business Media LLC
Published
2020-12-14
Volume
10
Issue
1
Pages
21640-21640
ISSN
2045-2322 (ISSN-L)
DOI
10.1038/s41598-020-78355-6
PMC
PMC7736301
OpenAlex
W3112300562
Cited by
133
References
30
Citations by year
2026: 7 · 2025: 12 · 2024: 19 · 2023: 20 · 2022: 44 · 2021: 31
FWCI
7.46
Type
journal article
Language
en
Update notices
correction (2022-03-15)
Open access
Licence
Locations
  • Scientific Reports (publishedVersion, cc-by)
  • Research Square (acceptedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • CUNY Academic Works (City University of New York) (submittedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • eScholarship (California Digital Library) (submittedVersion)
  • eScholarship (California Digital Library) (submittedVersion, cc-by)
Funders
Oura Health Oy
Topics
  • Physical Sciences › Engineering › Biomedical Engineering › Non-Invasive Vital Sign Monitoring
  • Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Infrared Thermography in Medicine
  • Health Sciences › Medicine › Epidemiology › Data-Driven Disease Surveillance
Keywords
COVID-19 pandemic, fever monitoring, wearable devices, peripheral temperature, core body temperature, symptom reporting, health monitoring
MeSH terms
Wearable Electronic Devices, COVID-19, Adult, Aged, Feasibility Studies, Female, Fever, Humans, Male, Middle Aged, Monitoring, Physiologic, Telemedicine, Young Adult, Self Report, Thermometry

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review