Table 3 Adjusteda hazard ratios and odds ratios for each outcome measure per decrease of 10 percentiles in rainfall versus historical values in an observational cohort study of the effect of rainfall on HIV outcomes in Southern Africa
| Â | Â | Per 10 percentile decrease in rainfall |
|---|---|---|
Mortality | N PWH (N with outcome) | Adjusted hazard ratios (95% confidence interval) |
 Overall | 104,138 (2951) | 1.18 (1.07–1.32) |
 Urban | 75,061 (1899) | NAb |
 Rural | 29,077 (1052) | 1.18 (1.06–1.32) |
CD4 counts < 200 cells/mm3 | N PWH (N with outcome) | Adjusted odds ratios (95% confidence interval) |
 Overall | 27,580 (3990) | 0.94 (0.89–1.00) |
 Urban | 23,154 (3167) | 0.94 (0.88–1.01) |
 Rural | 4426 (823) | 0.99 (0.89–1.10) |
HIV-1 viral loads > 400 copies/mL | N PWH (N with outcome) | Adjusted odds ratios (95% confidence interval) |
 Overall | 82,860 (13,788) | 1.05 (1.01–1.09) |
 Urban | 77,519 (13,341) | 1.05 (1.01–1.09) |
 Rural | 4741 (447) | Did not converge |
 ≥ 12-month gaps in care | N PWH (N with outcome) | Adjusted hazard ratios (95% confidence interval) |
 Overall | 270,708 (9426) | 0.98 (0.94–1.01) |
 Urban | 222,865 (8015) | 0.96 (0.92–1.00) |
 Rural | 47,843 (1411) | 0.86 (0.80–0.93) |