Skip to main content
. 2020 Dec 15;32(5):1476–1503. doi: 10.1057/s41287-020-00333-5

Table 2.

Regression result of the models for counts of COVID-19 deaths

Variables Model 1 parameter estimates Model 2 parameter estimates
mean SD 2.5% 97.5% mean SD 2.5% 97.5%
Constant − 12.87 0.27 − 13.45 − 12.36 − 12.46 0.29 − 13.01 − 11.91
Individual-level variables
Female − 3.89 0.45 − 4.79 − 3.1 − 4.05 0.45 − 4.94 − 3.24
Urban 2.88 0.11 2.69 3.11 2.11 0.09 1.95 2.28
SC (Dalit) 3.8 0.13 3.56 4.09 4.16 0.16 3.89 4.49
ST (Adivasi) 1.79 0.18 1.46 2.14 2.16 0.2 1.78 2.58
Smoking 4.08 0.25 3.63 4.61 4.14 0.24 3.68 4.63
Ill-health − 0.26 0.03 − 0.33 − 0.2 − 0.29 0.04 − 0.36 − 0.22
LowAssets − 3.08 0.4 − 4 − 2.42
District-level variables
Age 65+* 28.37 0.62 27.15 29.52 26.77 0.46 25.88 27.69
Migration* 2.23 0.26 1.72 2.77 1.85 0.27 1.29 2.38
Obesity% 26.77 0.48 25.82 27.67 26.68 0.64 25.41 27.93
Underweight% 9.49 0.36 8.8 10.21 9.71 0.34 9.07 10.37

Sources Dependent variable (COVID-19 death) is from How India Lives (2020). Individual variables and obesity % and underweight % by district averaging are from NFHS 2015/16; and *age and migration data were obtained from Census data, 2011

Notes All the coefficients are significant. Rows in bold indicate increased risk of death. The dependent variable’s, Dj, units are the counts of COVID-19 deaths. The population variable Pj acts as offset, leading to the predictions made in terms of rates μj. Mean, SD, 2.5% and 97.5% refer to the mean, standard deviation and percentiles of the posterior distributions for the model parameters. The last four row variables are measured at district level, and the rest at individual level, reflecting the hierarchical model. Variables: At Level 1, Ill-health is an index of comorbidities. See Table 1 for details of SC, ST, smoker status and urban residence. At Level 2 (districts), migration indicates the proportion of in-migrants (0–9 years) living in the district. Age 65+ is the % (scaled 0 to 1) of population in each district who are over age 65. Obesity% is the percentage (scaled 0–1) of district population who are obese (BMI ≥ 30); and underweight% is the percentage (scaled 0–1) of district population who are underweight (BMI < 18). The reference category is coded “0” (as in Table 1). Model 1 vs 2: In Model 2 only, LowAssets indicates the lowest asset index quintile, coded as 1, with other quintiles set to 0 (see Table 1). This is an indicator of household poverty