The role of family factors in antiretroviral therapy (ART) adherence self-efficacy among HIV-infected adolescents in southern Uganda

Table 3 Multilevel analysis model with participants’ sociodemographic and household characteristics

				95% Confidence Interval
Variable	Coef.	Std. Err.	P > \|z\|	Lower	Upper
Study group (Treatment)	−0.877	1.108	0.428	−3.049	1.294
Gender (female child)	0.892	0.930	0.338	−0.931	2.716
Age	0.835	0.207	0.000	0.429	1.241
Primary caregiver (ref: Biological Parent)
Grandparents	0.680	1.177	0.563	−1.627	2.987
Other relatives	2.153	1.160	0.063	−0.120	4.427
Orphanhood status (ref: Both parents are alive)
Single orphan	0.133	1.100	0.904	−2.023	2.289
Double orphan	0.278	1.427	0.845	−2.518	3.074
Household composition
Number of adults	−0.132	0.236	0.577	− 0.594	0.330
Number of children	0.187	0.323	0.56	−0.445	0.820
Enrolled in school (yes)	0.122	1.083	0.91	−2.000	2.244
HIV status disclosure (ref: never)
Sometimes	3.162	0.984	0.001	1.233	5.091
About half the time	1.095	2.026	0.589	−2.876	5.065
Most of the time	1.336	1.131	0.238	−0.881	3.552
Always	5.224	1.046	0.000	3.175	7.273
HIV medication regimen	−0.502	0.606	0.408	−1.690	0.687
Constant	78.82	3.89	0.000	71.192	86.447
Wald test: χ²(16, 702) = 84.2, p = 0.000
				95% Confidence Interval
Random-effects Parameters	Estimate	Std. Err		Lower	Upper
Variance (Clinic)	2.261	2.523		0.254	20.144
Variance (Child ID)	62.742	8.365		48.314	81.478
Variance (Residuals)	389.816	10.741		369.322	411.447
LR test vs. linear models: χ² (2, 702) =100.9, p = 0000

ISSN: 1471-2458