Research Article

Factors Determining Survival and Retention among HIV-Infected Children and Adolescents in a Community Home-Based Care and a Facility-Based Family-Centred Approach in Kampala, Uganda: A Cohort Study

Table 2

Factors associated with attrition among HIV-infected children and adolescents on ART in a community home-based care model and a facility-based family-centred approach, Kampala, Uganda, (2003–2010) using Cox regression.

CharacteristicsUnadjusted (univariate analysis)Adjusted* (multivariate analysis)
HR (95% CI) HR (95% CI)

Model of care
  FBFCA10.00210.006
  CHBC0.40 (0.23–0.70)0.29 (0.12–0.70)
Age (months)
 <12 11
 12–35 0.73 (0.31–0.70)0.4651.95 (0.39–9.89)0.419
 36–59 0.29 (0.09–0.87)0.0281.63 (0.18–15.17)0.664
 60+0.43 (0.22–0.85)0.0153.19 (0.27–37.03)0.354
Gender
  Female10.675
  Male1.13 (0.64–1.99)
Cd4 count and percent
 CD40.86 (0.69–1.06)0.162
 CD4%1.00 (0.57–1.77)0.992
Level of immunosuppression by CD4
  Not significant11
  Mild4.17 (1.31–13.31)0.016 4.66 (1.21–17.98)0.026
  Advanced0.87 (0.16–4.73)0.8680.89 (0.09–9.12)0.921
  Severe3.14 (1.10–8.94)0.0323.16 (0.81–12.39)0.099
Age at ART initiation
 Months0.74 (0.61–0.90)0.0030.67 (0.31–1.44)0.309
WHO clinical staging
 I-II10.309
 III-IV1.52 (0.68–3.38)
Growth response
Weight-for-age
 ≤−2SD10.00610.002
 >−2SD0.40 (0.21–0.77)0.31 (0.15–0.65)
Height-for-age
 ≤−2SD10.060
 >−2SD0.54 (0.28–1.03)

Hazard of attrition; log transformed due to skewed data;
*only factors significant at univariate or borderline were considered into multivariate.