Indiana's 2020 Random-Sample Study: 2.8% Infected, 10 Times the Reported COVID-19 Cases

In late April 2020, Indiana did something no state had yet done: it tested a random sample of its residents to find out how many had been infected with SARS-CoV-2, the virus that causes COVID-19. The answer was about 2.8 percent — some 187,802 people. The state had confirmed only 17,792 cases, so reported cases captured roughly one infection in ten.

CDC's summary of the Indiana findings. Credit: CDC, Morbidity and Mortality Weekly Report.

How it worked

Participants were drawn at random from Indiana tax-return filers and dependents aged 12 and older, stratified by the state's 10 public health preparedness districts, with institutionalized and deceased people removed. Of 15,495 people contacted, 3,658 (23.6 percent) took part from April 25 to 29, at 68 sites run with help from state agencies, the Indiana National Guard and private organizations. Each gave a nasal swab to test for current infection and a blood sample to test for antibodies from past infection, and results came back online within three days. The data were weighted for age, race and Hispanic ethnicity to reflect the state.

What it found

• More active than past infections, suggesting Indiana was still early in its epidemic.
• Households mattered: people whose household member had been diagnosed with COVID-19 were 15 times as likely to test positive.
• Hispanic residents had higher prevalence in the random sample.
• Fatality: with 1,099 COVID-19 deaths reported by late April, the estimated infection-fatality rate was 0.58 percent — about six times the 0.1 percent for influenza, lower than on a cruise ship (1.3 percent) and close to an estimate for China (0.66 percent).

Because racial and ethnic minority residents responded at lower rates, civic and faith leaders helped organize two days of additional, nonrandom testing in Indianapolis on May 2 and 3, for 898 people. Minority participants there had higher prevalence — though people worried about being infected may have been more likely to come, so those figures are reported separately.

What it meant

The low overall prevalence, and the strong household link, suggest that stay-at-home orders and other distancing measures in effect from March 24 to May 3 likely limited community spread. With the great majority of Indianans still susceptible, the authors urged continued distancing, masks and hand hygiene; protecting household members when patients isolate at home; and outreach tailored to the languages and cultures of minority communities, many of whose members were essential workers.

Limits

Only 23.6 percent of those invited took part, though the data were adjusted for nonresponse. Tests can miss early infections or give some false positives. The study covered one state at one moment, and excluded people who did not file state taxes, those in institutions, and children under 12. An erratum to the report has been published.

Sources

Based on "Population Point Prevalence of SARS-CoV-2 Infection Based on a Statewide Random Sample — Indiana, April 25–29, 2020," by Nir Menachemi and colleagues (Indiana University Richard M. Fairbanks School of Public Health; Regenstrief Institute; Indiana State Department of Health; and others), Morbidity and Mortality Weekly Report, Centers for Disease Control and Prevention; a work of the United States government in the public domain.