Feasibility and effectiveness of daily temperature screening to detect COVID-19 in a prospective cohort at a large public university

Abstract

Background: Many persons with active SARS-CoV-2 infection experience mild or no symptoms, presenting barriers to COVID-19 prevention. Regular temperature screening is nonetheless used in some settings, including university campuses, to reduce transmission potential. We evaluated the potential impact of this strategy using a prospective university-affiliated cohort.

Methods: Between June and August 2020, 2912 participants were enrolled and tested for SARS-CoV-2 by PCR at least once (median: 3, range: 1-9). Participants reported temperature and symptoms daily via electronic survey using a previously owned or study-provided thermometer. We assessed feasibility and acceptability of daily temperature monitoring, calculated sensitivity and specificity of various fever-based strategies for restricting campus access to reduce transmission, and estimated the association between measured temperature and SARS-CoV-2 test positivity using a longitudinal binomial mixed model.

Results: Most participants (70.2%) did not initially have a thermometer for taking their temperature daily. Across 5481 total person months, the average daily completion rate of temperature values was 61.6% (median: 67.6%, IQR: 41.8-86.2%). Sensitivity for SARS-CoV-2 ranged from 0% (95% CI 0-9.7%) to 40.5% (95% CI 25.6-56.7%) across all strategies for self-report of possible COVID-19 symptoms on day of specimen collection, with corresponding specificity of 99.9% (95% CI 99.8-100%) to 95.3% (95% CI 94.7-95.9%). An increase of 0.1 °F in individual mean body temperature on the same day as specimen collection was associated with 1.11 increased odds of SARS-CoV-2 positivity (95% CI 1.06-1.17).

Conclusions: Our study is the first, to our knowledge, that examines the feasibility, acceptability, and effectiveness of daily temperature screening in a prospective cohort during an infectious disease outbreak, and the only study to assess these strategies in a university population. Daily temperature monitoring was feasible and acceptable; however, the majority of potentially infectious individuals were not detected by temperature monitoring, suggesting that temperature screening is insufficient as a primary means of detection to reduce transmission of SARS-CoV-2.

Keywords: COVID-19; Detection; Fever; SARS-CoV-2; Screening; Temperature monitoring; Thermometer.

Fig. 1

Participant count of the proportion of days quantitative temperature was recorded via the daily survey, per total days of enrollment (A) and total days for which the daily survey was at least partially completed (B)

Fig. 2

Heat map of temperatures measured each day before and after specimen collection at which 56* participants tested PCR positive for SARS-CoV-2. The vertical black line denotes the day that participant provided the specimen which would subsequently test positive for SARS-CoV-2. White boxes are days that the participant did not record a temperature (including days before they had enrolled in the study), and colored boxes range from yellow to red depending on the self-reported participant temperature, as found in the temperature color scale on the bottom left of the figure