Table 3.
Segmented regression models of impact of drought on retention in care 6 months after ART initiation (regression coefficients and 95% confidence intervalsa).
| Pre-drought | January–December 2013 | January–December 2014 | January–December 2015 | January–December 2016 | January 2017–January 2019 | |
|---|---|---|---|---|---|---|
| All individuals | ||||||
| Mean retentionb | 91.4% | 86.7% | 84.3% | 92.2% | 89.5% | 86.6% |
| Trendc | 1.00 (1.00,1.01) | 0.99 (0.98,1.00) | 1.00 (1.00,1.01) | 1.01 (1.00,1.02) | 1.00 (1.00,1.01) | 1.00 (0.99,1.00) |
| P = 0.024 | P = 0.092 | P = 0.390 | P = 0.002 | P = 0.621 | P = 0.002 | |
| Level changed | N/A | 0.88 (0.68,1.14) | 0.70 (0.53,0.92) | 1.14 (0.85,1.54) | 0.67 (0.52, 0.85) | 0.90 (0.72,1.12) |
| P = 0.331 | P = 0.010 | P = 0.382 | P = 0.001 | P = 0.347 | ||
| Men | ||||||
| Mean retentionb | 91.6% | 90.4% | 90.5% | 91.1% | 89.6% | 86.1% |
| Trendc | 1.01 (1.00,1.01) | 1.00 (0.99,1.02) | 1.01 (1.00,1.02) | 1.00 (0.99,1.02) | 1.00 (0.99,1.01) | 1.00 (0.99,1.00) |
| P = 0.002 | P = 0.474 | P = 0.060 | P = 0.653 | P = 0.705 | P = 0.227 | |
| Level changed | N/A | 0.69 (0.49,0.98) | 0.55 (0.34,0.88) | 0.87 (0.51,1.49) | 0.76 (0.48, 1.19) | 0.85 (0.57,1.26) |
| P = 0.038 | P = 0.013 | P = 0.615 | P = 0.227 | P = 0.408 | ||
| Women | ||||||
| Mean retentionb | 91.5% | 85.0% | 82.3% | 92.4% | 89.3% | 86.3% |
| Trendc | 1.00 (1.00,1.01) | 0.99 (0.98,1.00) | 1.00 (0.99,1.01) | 1.01 (1.00,1.03) | 1.00 (0.99,1.01) | 1.00 (0.99,1.00) |
| P = 0.421 | P = 0.039 | P = 0.716 | P = 0.007 | P = 0.744 | P = 0.004 | |
| Level changed | N/A | 1.00 (0.70,1.43) | 0.76 (0.54,1.05) | 1.22 (0.85,1.75) | 0.65 (0.47, 0.90) | 0.93 (0.72,1.20) |
| P = 0.987 | P = 0.099 | P = 0.281 | P = 0.010 | P = 0.563 | ||
| Age 15–24 years | ||||||
| Mean retentionb | 90.8% | 82.4% | 76.6% | 85.9% | 82.3% | 80.0% |
| Trendc | 1.00 (1.00,1.01) | 0.99 (0.98,1.01) | 1.00 (0.99,1.01) | 1.01 (1.00,1.02) | 1.00 (0.99,1.01) | 1.00 (0.99,1.00) |
| P = 0.396 | P = 0.337 | P = 0.665 | P = 0.011 | P = 0.907 | P = 0.041 | |
| Level changed | N/A | 0.71 (0.47,1.08) | 0.61 (0.42,0.90) | 0.91 (0.57,1.46) | 0.74 (0.51, 1.08) | 1.09 (0.78,1.54) |
| P = 0.108 | P = 0.012 | P = 0.703 | P = 0.120 | P = 0.604 | ||
| Age 25–59 years | ||||||
| Mean retentionb | 91.4% | 88.0% | 87.0% | 94.1% | 91.9% | 89.0% |
| Trendc | 1.00 (1.00,1.01) | 0.99 (0.98,1.00) | 1.00 (1.00,1.01) | 1.01 (1.00,1.02) | 1.00 (0.99,1.01) | 1.00 (0.99,1.00) |
| P = 0.022 | P = 0.279 | P = 0.306 | P = 0.034 | P = 0.786 | P = 0.035 | |
| Level changec | N/A | 0.92 (0.67,1.27) | 0.73 (0.53,1.02) | 1.32 (0.95,1.83) | 0.67 (0.49, 0.92) | 0.82 (0.61,1.10) |
| P = 0.611 | P = 0.063 | P = 0.094 | P = 0.014 | P = 0.180 | ||
aEstimates from GEE logistic regression, adjusted for median CD4 count at ART initiation. bMean % retained in care at end of each period (December 2012; December 2013; December 2014; December 2015; December 2016; December 2018). cOdds ratio for the linear trend in being retained in care from one month to the next. dImmediate change in odds of being retained at start of interval.