Table 2.

Impact of insurance coverage on four COVID-19 outcome variables based on panel data regressions in the CDC surveillance data set through February 13, 2021. The regression is given in Eq. (1); the proportion change on the insurance variable is from Eq. (2). The change in the outcome variables is calculated as the product of the proportion change and the number of recorded outcomes. Because the sample is not the entire population, the estimate for the effect of insurance on the entire population comes from applying the proportion change to the entire population at risk.

	Proportion Change	Total through Feb 13	Effect if all insured	Range (+/- 1 s.e.)
Person-days between symptom onset and test	-0.017 (s.e. 0.063)	45,591,172	(775,049)	(-3,692,884 - 2,142,785)
Cases	-0.112⁎⁎⁎ (s.e. 0.041)	25,942,037	(2,905,508)	(1,841,885 - 3,969,132)
Hospitalization	-0.185⁎⁎⁎ (s.e. 0.054)	1,207,878	(223,457)	(158,232 - 288,683)
Deaths	-0.264⁎⁎⁎ (s.e. 0.085)	225,210	(59,455)	(40,313 - 78,598)

Note: *p<0.10, **p<0.05, ***p<.01.