Skip to main content
. 2019 Feb 7;22(2):e25239. doi: 10.1002/jia2.25239

Table 3.

Cox regression analyses of predictors of attrition to carea

Variables n (%) Crude HR (95% CI) p‐value Adjustedb HR (95% CI) p‐value
Cohort (n = 1381)
UTT 155 (17.0) Ref=1 Ref=1
Pre‐UTT 111 (23.8) 1.45 (1.14 to 1.85) 0.003 1.29 (1.09 to 1.53) 0.003
Age (n = 1381)
10 to 19 23 (17.6) 0.94 (0.61 to 1.45) 0.769 0.97 (0.57 to 1.67) 0.926
20 to 24 46 (30.5) 1.70 (1.23 to 2.35) 0.001 1.53 (1.01 to 2.32) 0.045
25 to 49 173 (18.6) Ref=1 Ref=1
≥50 24 (14.1) 0.77 (0.50 to 1.18) 0.235 0.84 (0.54 to 1.31) 0.437
Gender category (n = 1381)
Femalec 118 (16.5) Ref=1 Ref=1
Male 95 (18.9) 1.14 (0.87 to 1.49) 0.350 1.17 (0.85 to 1.61) 0.333
Pregnant/breastfeeding 53 (33.1) 2.17 (1.57 to 3.00) <0.001 1.87 (1.30 to 2.38) 0.001
Facility (n = 1381)
Rural 179 (18.6) Ref=1 Ref=1
Urban 87 (20.9) 1.11 (0.86 to 1.43) 0.430 1.21 (0.71 to 2.07) 0.481

CI, confidence interval; UTT, Universal Test and Treat.

aAttrition is the inverse of retention on ART and is the sum of Dead, LTFU and Stop ART. bEstimates are based on the Cox proportional hazards model adjusted for cohort, age, gender and facility type, and additionally adjusted for clustering by health facility. cNon‐pregnant and non‐breastfeeding.