Table 1. Correlates of willingness to circumcise male children for reduction of future risk for HIV.
| Male partners (N = 316) | Female partners (N = 317) | |||||
| Characteristics | Willing to circumcise male childn/total (%)* | A PR† (95% CI) | P-value | Willing to circumcise male childn/total (%)* | A PR† (95% CI) | P-value |
| Age (years) | ||||||
| <30 | 56/61 (91.8) | - | - | 135/142 (95.7) | - | - |
| 30–39 | 122/133 (91.7) | - | - | 114/121 (94.2) | - | - |
| ≥40 | 107/122 (87.7) | - | - | 51/54 (94.4) | - | - |
| HIV Status | ||||||
| Positive | 148/161 (91.9) | 1.09 (1.01, 1.18) | 0.03 | 148/155 (95.5) | - | - |
| Negative | 137/155 (88.4) | Reference | 152/162 (93.8) | - | - | |
| Male partner's circumcision status | ||||||
| Circumcised | 93/95 (97.9) | 1.09 (1.02, 1.16) | 0.004 | 110/113 (97.4) | - | - |
| Uncircumcised | 192/221 (86.9) | Reference | 190/204 (93.1) | - | - | |
| Residence | ||||||
| urban | 203/220 (92.3) | - | - | 208/218 (95.4) | - | - |
| Rural | 82/96 (85.4) | - | - | 92/99 (92.9) | - | - |
| Education level attained | ||||||
| ≤Primary | 128/145 (88.3) | - | - | 190/201 (94.5) | - | - |
| ≥Secondary | 157/171 (91.8) | - | - | 110/116 (94.8) | - | - |
| Prior discussion about circumcision with partner | ||||||
| Yes | 75/76 (98.7) | 1.08 (1.02, 1.14) | 0.009 | 84/84 (100) | 1.07 (1.03, 1.10) | <0.001 |
| No | 210/240 (87.5) | Reference | 216/233 (92.7) | Reference | ||
| Aware that circumcision reduces risk of HIV infection in adult men | ||||||
| Yes | 236/243 (97.1) | 1.34 (1.20, 1.49) | <0.001 | 273/284 (96.1) | 1.14 (1.00, 1.30) | 0.05 |
| No | 49/73 (67.1) | Reference | 27/33 (81.8) | Reference | ||
| Male child<18 years with partner | ||||||
| Yes | 164/186 (88.2) | - | - | 177/186 (95.2) | - | - |
| No | 121/130 (93.1) | - | - | 123/131 (93.9) | - | - |
| Religion | ||||||
| Christian | 230/260 (88.5) | - | - | 252/267 (94.4) | - | - |
| Muslim | 55/56 (98.2) | - | - | 48/50 (96.0) | - | - |
*Row percentages displayed; APR adjusted prevalence ratios.
†
Addition of age, religion, education level, and place of residence did not substantially affect estimates from the multivariable models.