Table 2. Factors Associated With Coronavirus Disease 2019 Positivity.
| Factor | OR (95% CI)a | P value |
|---|---|---|
| Patient race and socioeconomic status | ||
| Black race | 5.37 (3.94-7.29) | .001 |
| Enrolled in Medicaid | 0.66 (0.41-1.08) | .10 |
| Interaction of race × poverty status | 0.90 (0.49-1.62) | .72 |
| Demographic characteristics | ||
| Age, y | ||
| <60 | 1 [Reference] | NA |
| ≥60 | 2.04 (1.53-2.73) | .001 |
| Sex | ||
| Women | 1 [Reference] | NA |
| Men | 1.55 (1.21-2.00) | .001 |
Abbreviations: NA, not applicable; OR, odds ratio.
The ORs and 95% CIs were obtained from a model estimated using a generalized estimating equation logit specification that adjusts for clustering at the 9-digit zip code–level of patients’ residence. All models also include a constant term and indicators for comorbidities (categorized as 0, 1-2, ≥3, based on presence of diabetes, hypertension, heart disease, lung disease, kidney disease, and cancer); symptoms at testing (categorized as fever, cough, shortness of breath, and other symptoms, including gastrointestinal symptoms); body mass index, including an indicator for missing body mass index; and history of smoking. The intracluster correlation coefficient, estimated based on an exchangeable-correlation covariance structure, had a ρ of 0.79 (95% CI, 0.58-0.91).24