Skip to main content
. 2021 Apr 20;82(6):260–268. doi: 10.1016/j.jinf.2021.04.012

Table 4.

Multivariable binomial logistic regression model for a positive SARS-CoV-2 PCR result among patients meeting a clinical and radiological reference standard for COVID-19a

Model 1 Model 2
Characteristic Odds ratio (95% CI) P value Odds ratio (95% CI) P value
Age, per year 1.00 (0.99 – 1.02) 0.54 1.00 (0.99 – 1.02) 0.59
Duration of symptoms at time of PCR testingb 0.21 0.20
 3 days Reference Reference
 5 days 1.09 (0.89 – 1.35) 1.14 (0.92 – 1.41)
 7 days 1.13 (0.88 – 1.45) 1.22 (0.93 – 1.62)
 9 days 1.08 (0.46 – 2.55) 1.25 (0.86 – 1.80)
 14 days 0.76 (0.01 – 56.8) 1.14 (0.63 – 2.06)
Number of comorbidities, per additional comorbidity 1.22 (1.04 – 1.43) 0.02
Current, vs never or ex-smoker 0.23 (0.12 – 0.42) <0.001 0.25 (0.13 – 0.50) <0.001
Received oxygen therapy 1.90 (1.10 – 3.27) 0.02 1.89 (1.09 – 3.29) 0.02
COPD 0.95 (0.48 – 1.88) 0.89
Other respiratory disease 0.77 (0.27 – 2.17) 0.62
Chronic heart disease 2.87 (1.51 – 5.44) 0.001
a

Model 1 adjusted for age, duration of symptoms, number of comorbidities, smoking and receipt of oxygen therapy (as a marker of disease severity). Model 2 adjusted for additional confounders associated with smoking comprising COPD, chronic heart disease and respiratory disease. b A cubic spline for duration of symptoms at time of testing with four knots was included in the model. P value presented for this variable is the Wald likelihood test comparing a model including the cubic spline variable with one without it.