Table 3. Logistic regression predicting child HIV testing results both without (model 1) and with (model 2) an interaction term for orphans by program participation.

	Model 1				Model 2
		95% CI		Log odds		95% CI		Log odds
	OR	Lower	Upper	p-Value	OR	Lower	Upper	p-Value
Age in years	0.94	0.91	0.96	<0.0001	0.94	0.91	0.96	< 0.0001
Male gender	1.07	0.83	1.38	0.6118	1.07	0.82	1.38	0.6278
Orphan (single or double)	1.16	0.79	1.72	0.4534	0.77	0.43	1.39	0.3869
Male guardian	0.32	0.14	0.72	0.0058	0.32	0.14	0.72	0.0063
Youth guardian (< 25 years)	0.18	0.04	0.69	0.0127	0.17	0.04	0.67	0.0115
Married guardian	1.05	0.71	1.55	0.8128	1.03	0.70	1.53	0.8816
Guardian without secondary education	0.96	0.61	1.50	0.8396	0.93	0.59	1.45	0.7364
Guardian is not biological parent	1.15	0.77	1.74	0.4934	1.15	0.77	1.73	0.5016
Guardian is HIV+ biological parent	3.14	1.77	5.56	<0.0001	3.15	1.79	5.55	< 0.0001
Guardian has correct HIV knowledge	1.70	1.13	2.54	0.0104	1.70	1.14	2.55	0.0100
Household asset index	1.06	0.93	1.20	0.4019	1.06	0.93	1.21	0.3615
Household in informal settlement	2.22	1.32	3.72	0.0026	2.22	1.33	3.73	0.0024
Enrolled previously (intervention group)	1.97	1.34	2.92	0.0007	1.69	1.09	2.62	0.0190
Enrolled previously* orphan	–	–	–	–	2.12	1.00	4.47	0.0494