Skip to main content

Decreasing median age of COVID-19 cases in the United States-Changing epidemiology or changing surveillance?

Our take —

This study utilizing surveillance data from across the US, predominantly from Utah, compared testing positivity rates and inpatient trends by age in March/April 2020 versus June/July 2020. Over time testing positivity rates increased among those under 50 years and decreased in those over 50 years; the lack of a corresponding decrease in the median age of inpatients suggests that changes in testing are partly driving the observed case increases among young people. However, the increased positivity rates over time among young people suggests that there may be more than surveillance changes occurring and that the number of undiagnosed younger infections remains high despite expanded testing. Further, these inferences rely on a limited amount of inpatient data and depend largely on demographic data from individuals tested for SARS-CoV-2 in the state of Utah. Thus, caution is warranted for any generalizations made to describe incidence trends outside of Utah, and data presented do not strongly support the claim that surveillance and not true epidemiological shifts explain changing trends over time.

Study design


Study population and setting

This retrospective analysis of surveillance data explored the age distributions and positivity rates of 227,601 SARS-CoV-2 molecular testing results from a Utah-based national reference laboratory between March 10 and July 8, 2020. Conducted using three assays of similar sensitivity and detection limits, the national reference laboratory processed SARS-Cov-2 tests from over 1000 hospitals across the US, with nearly half (48.4%) from Utah. Test positivity and age distributions of cases were compared across two time periods, defined as early (March 10 to April 30,2020) and late (June 1 to July 8,2020). The age of COVID-19 inpatients and outpatients were also compared using data available from individuals tested through the University of Utah Healthcare system during these time periods.

Summary of Main Findings

A total of 19,320 out of 277,601 SARS-CoV-2 tests (7%) were positive. In Utah and other represented states, the median age of positive cases decreased from early (March-April) to late (June-July) period, with the overall median age of those testing positive dropping from 40.8 to 35.8. While the positivity rate increased for each age group younger than 50 years between the early and late time periods, with the greatest increase among individuals under 18 years (rising from 3.3% to 10%), positivity rates decreased for each age group over 50 years (e.g., decreasing from 6.1% to 3.6% in those over 70 years). Using a limited number of Utah individuals with known inpatient status, the authors described an increase in the median age of inpatients (increase of 5.8 years) and a decrease in the median age of outpatients (decrease of 3.9 years) from the early to late period. Finally, while no detailed information on changes in testing strategies were provided aside from increased test availability in the late period, the authors state that the observed changes in the median age of inpatients vs. outpatients over time suggest changes in testing rather than changes in underlying COVID-19 epidemiology explain the decreased age among those testing positive for COVID-19 later in the pandemic.

Study Strengths

This study utilizes data from a large national reference laboratory that provides clinical testing capacity for over 1000 US hospitals. With roughly half of the data analyzed derived from Utah, the study was well powered to detect clinically meaningful differences in age among those testing positive for SARS-CoV-2 in Utah.


As noted by the authors, much of the testing data and all the results with inpatient-related information were from Utah. The assumption is made that the national reference laboratory testing data is representative of the true burden of disease in Utah, which may be more likely for the early period considering half of all confirmed Utah cases were captured within this early period. This assumption is problematic for the late period as a smaller proportion (20%) of confirmed Utah cases were captured during the late period. We agree with the authors that it would be inappropriate to make generalizations outside of Utah if Utah’s age distribution and testing strategies were not representative of other states or the whole US. If age or inpatient status were associated with missing inpatient-status data, it is possible that the inferences based on the limited number of inpatient cases from the early (N=39) and late periods (N=47) are biased, as inpatient status was missing for 38.6% of the Utah early period cases compared to 22.4% of the Utah late period cases. Therefore, larger studies with complete inpatient data are needed to confirm inpatient status-related findings. While the three assays utilized to identify positive cases are stated to have similar limits of detection and sensitivity, no data or literature is provided to support these claims. Aside from the increased availability of tests, no descriptions of the various testing strategies employed across the US or in Utah were included for either time period. This is problematic considering such changes in testing strategy were suggested by the authors to explain the decreased age among those testing positive for SARS-CoV-2. Finally, while the median age does decrease across time periods, the increased positivity rates over time among younger individuals and decreased positivity rates among older people suggests that it may be more than expanded testing that is changing case counts among younger populations, calling the conclusions into question.

Value added

This study replicates previous observations of increased diagnoses of COVID-19 among younger individuals, showing that over time test positivity rates for SARS-CoV-2 infection are increasing among younger people. This overall increase in cases is likely due to increased diagnosis of asymptomatic infections and cases with mild infections, given that the inpatient ages did not decrease over time, but also in part may be due to epidemiological changes in infections given the increasing positivity rate amongst younger populations.

This review was posted on: 19 December 2020