Pooled relative risk of mortality in patients with COVID-19 comparing vitamin D supplementation with controls (placebo or no vitamin D supplementation) in (a) tertiary prevention studies and (b) both secondary and tertiary prevention studies. Studies are stratified by study design (randomized controlled trials (RCTs) [3,12,13,25] vs. non-randomized intervention studies (NRIS) [4,5,6,15,28,29,30,33,34,35]. Subgroup and pooled summary estimates are reported with 95% confidence interval, analyzed with the Mantel–Haenszel fixed-effects model method. Heterogeneity was quantified by I2. The chi-square test served to examine group difference between study designs.