. 2020 Sep 3;3(9):e2019722. doi: 10.1001/jamanetworkopen.2020.19722

Table 4. Multivariable Association of Vitamin D Deficiency and Treatment With Testing Positive for COVID-19 in 489 Patients.

Characteristic	No. (%)	Relative risk (95% CI)	P value
Age (linear spline)^a
<50	260 (53)	1.05 (1.01-1.09)	.02
≥50	229 (47)	1.02 (1.00-1.05)	.06
Sex
Male	123 (25)	1 [Reference]
Female	366 (75)	0.87 (0.52-1.44)	.58
Race
White	158 (32)	1 [Reference]
Other than White	331 (68)	2.54 (1.26-5.12)	.009
Ethnicity
Non-Hispanic	448 (92)	1 [Reference]
Hispanic	41 (8)	0.29 (0.04-2.01)	.21
Employee status, UCM employee
No	328 (67)	1 [Reference]
Yes	161 (33)	0.93 (0.52-1.64)	.79
Most recent vitamin D <20 ng/mL
Likely deficient^b	124 (25)	1.77 (1.12-2.81)	.02
Uncertain deficiency^c	48 (10)	1.10 (0.49-2.43)	.82
Uncertain deficiency^d	30 (5)	1.09 (0.43-2.82)	.85
Likely sufficient^e	287 (59)	1 [Reference]
Comorbidity indicators
Hypertension	261 (53)	1.08 (0.60-1.97)	.79
Diabetes	137 (28)	0.78 (0.49-1.26)	.31
Chronic pulmonary disease	117 (24)	0.91 (0.55-1.52)	.73
Pulmonary circulation disorders	20 (4)	0.64 (0.23-1.79)	.40
Depression	119 (24)	1.22 (0.74-2.02)	.44
Chronic kidney disease	116 (24)	0.80 (0.49-1.32)	.39
Liver disease	56 (11)	0.99 (0.47-2.08)	.98
Comorbidities with immunosuppression	105 (21)	0.39 (0.20-0.76)	.005
BMI, mean	29.8	1.02 (0.996-1.048)	.10
Goodness-of-link test of squared predicted value³¹	NA	NA	.87
Hosmer-Lemeshow goodness-of-fit decile test³²	NA	NA	.89

Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); COVID-19, coronavirus disease 2019; NA, not applicable; UCM, University of Chicago Medicine.

^{^a}

A piecewise linear spline with a single knot at 50 improved model fit over models with unadjusted age or more complex parameterizations.

^{^b}

Answer was yes to most recent vitamin D level within 1 year being deficient (<20 ng/mL); dose was stable or decreased after last visit. Vitamin D dose was rank-ordered as follows: calcitriol > 3000+ IU D3 > 2000 IU D3 > D2 > 1-1000 IU D3/multivitamin > no vitamin D.

^{^c}

Answer was yes to most recent vitamin D level within 1 year being deficient (<20 ng/mL); dose was increased after last visit. Vitamin D dose was rank-ordered as follows: calcitriol > 3000+ IU D3 > 2000 IU D3 > D2 > 1-1000 IU D3/multivitamin > no vitamin D.

^{^d}

Answer was no to most recent vitamin D level within 1 year being deficient (<20 ng/mL); dose was decreased after last visit. Vitamin D dose was rank-ordered as follows: calcitriol > 3000+ IU D3 > 2000 IU D3 > D2 > 1-1000 IU D3/multivitamin > no vitamin D.

^{^e}

Answer was no to most recent vitamin D level within 1 year being deficient (<20 ng/mL); dose was stable or increased after last visit. Vitamin D dose was rank-ordered as follows: calcitriol > 3000+ IU D3 > 2000 IU D3 > D2 > 1-1000 IU D3/multivitamin > no vitamin D.