Table 3. Sensitivity Analysis of Factors Associated With COVID-19 Incidence in Massachusetts Prisons, April 21, 2020, to January 11, 2021.
Variable | Incidence rate ratio (95% CI)a | ||
---|---|---|---|
Model 1 | Model 2 | Model 3 | |
1-wk Lag | |||
County-level COVID-19 incidence rate (per 10 cases per 100 000 person-weeks) | 1.06 (1.05-1.08) (P < .001) | 1.07 (1.05-1.08) (P < .001) | 1.06 (1.05-1.07) (P < .001) |
Prison population density (per 10-percentage-point difference) | 1.16 (1.05-1.29) (P = .004) | NA | NA |
Prison population density | |||
<70% | NA | 1.00 [Reference] | NA |
70%-100% | NA | 5.08 (1.51-17.02) (P = .008) | NA |
>100% | NA | 6.97 (2.50-19.41) (P < .001) | NA |
Single-cell occupancy (per 10-percentage-point difference) | NA | NA | 0.80 (0.71-0.90) (P < .001) |
2-wk Lag | |||
County-level COVID-19 incidence rate (per 10 cases per 100 000 person-weeks) | 1.06 (1.05-1.08) (P < .001) | 1.07 (1.05-1.08) (P < .001) | 1.06 (1.05-1.07) (P < .001) |
Prison population density (per 10-percentage-point difference) | 1.17 (1.05-1.31) (P = .004) | NA | NA |
Prison population density | |||
<70% | NA | 1.00 [Reference] | NA |
70%-100% | NA | 4.39 (1.09-17.73) (P = .04) | NA |
>100% | NA | 7.04 (2.58-19.20) (P < .001) | NA |
Single-cell occupancy (per 10-percentage-point difference) | NA | NA | 0.80 (0.70-0.91) (P = .001) |
Abbreviation: NA, not applicable.
All estimates derived from generalized estimating equations Poisson regression models. Incarcerated population density is reported as percent of facility design capacity. Model 1: incarcerated population density as a percentage of design capacity modeled as a continuous variable; model 2: incarcerated population density as a percentage of design capacity modeled as a categorical variable; model 3: percentage of incarcerated persons in single-cell units.