Table 3.

Adjusted associations between unhealthy opioid use and outcome severity for hospitalized patient encounters carrying a diagnosis of COVID-19 in Chicago between January 1, 2020, and December 31, 2020 (N=4110).

Predictor			Odds ratio (95% CI)		P value
Unhealthy opioid usea			1.16 (1.04-1.29)		.009
Age			1.01 (1.01-1.02)		<.001
BMI			1.02 (1.01-1.03)		<.001
Sexb			0.75 (0.63-0.90)		.002
Race/ethnicityb					.21
	Black	0.92 (0.73-1.18)		.52
	Hispanic or Latinx	1.12 (0.84-1.45)		.39
	Other	1.22 (0.86-1.72)		.27
Insuranceb					.18
	Medicare	0.87 (0.69-1.07)		.22
	Private	0.79 (0.62-1.00)		.053
	Other	0.78 (0.55-1.11)		.17

^aOpioid misuse classifications were log transformed in this analysis.

^bThese rows report the P value for the omnibus effect for categorical predictors with more than 2 levels, and rows nested with them represent comparisons with the reference categories of male, non-Hispanic White, and Medicaid.