Protective factors for adolescent sexual risk behaviours and experiences linked to HIV infection in South Africa: a three-wave longitudinal analysis of caregiving, education, food security, and social protection

Rudgard, William E.; Saminathen, Maria Granvik; Orkin, Mark; Banougnin, Boladé Hamed; Shenderovich, Yulia; Toska, Elona

doi:10.1186/s12889-023-16373-5

Table 2 Multivariable associations between hypothesised protective factors and HIV risk practices in girls and boys. Average effects are modelled when there is no evidence that between- and within-individual effects are different. N = 1563, Observations = 4402

From: Protective factors for adolescent sexual risk behaviours and experiences linked to HIV infection in South Africa: a three-wave longitudinal analysis of caregiving, education, food security, and social protection

	Multiple sexual partners			Transactional sex			Age-disparate sex			Condomless sex			Sex on substances^a
	aOR (95%CI)	naïve p-value^b	sharpened q-value	aOR (95%CI)	naïve p-value^b	sharpened q-value	aOR (95%CI)	naïve p-value^b	sharpened q-value	aOR (95%CI)	naïve p-value^b	sharpened q-value	aOR (95%CI)	naïve p-value^b	sharpened q-value
Girls
Number of social grants
Between
Within
Average	0.99 (0.88–1.12)	0.864	0.976	0.99 (0.86–1.14)	0.851	0.253	1.06 (0.92–1.23)	0.416	0.263	1.08 (0.99–1.19)	0.088	0.258	0.86 (0.70–1.07)	0.181	0.293
Positive caregiving
Between
Within
Average	0.98 (0.95–1.01)	0.283	0.403	0.94 (0.91–0.98)	0.002	0.004	0.96 (0.93–1.00)	0.043	0.061	1.00 (0.98–1.03)	0.921	0.698	1.04 (0.98–1.10)	0.228	0.296
Caregiver supervision
Between	0.90 (0.86–0.93)	< .001	0.001							0.95 (0.92–0.97)	< .001	0.001	0.86 (0.80–0.92)	< .001	0.001
Within	1.02 (0.99–1.04)	0.123	0.258							1.01 (0.99–1.03)	0.440	0.459	0.96 (0.92–1.00)	0.056	0.202
Average				0.94 (0.92–0.97)	< .001	0.001	0.97 (0.94–0.99)	0.006	0.016
Adolescent-caregiver communication
Between
Within
Average	0.99 (0.93–1.04)	0.637	0.806	1.10 (1.02–1.18)	0.008	0.007	0.97 (0.91–1.04)	0.357	0.263	0.98 (0.94–1.02)	0.274	0.358	0.93 (0.84–1.03)	0.139	0.293
Education enrolment
Between										0.42 (0.24–0.72)	0.001	0.005
Within										0.92 (0.56–1.51)	0.753	0.698
Average	1.10 (0.70–1.73)	0.669	0.806	1.45 (0.85–2.46)	0.168	0.073	0.44 (0.26–0.74)	0.002	0.013				0.83 (0.39–1.74)	0.618	0.464
Days with enough food
Between										0.86 (0.71–1.04)	0.117	0.258
Within										1.08 (0.94–1.25)	0.250	0.358
Average	0.83 (0.73–0.95)	0.008	0.025	0.78 (0.68–0.90)	< .001	0.003	1.15 (0.95–1.38)	0.145	0.123				0.92 (0.71–1.20)	0.545	0.464
Variance components
Level 2: In level-1 intercept	2.52			1.15			2.13			1.51			3.43
Goodness of fit
AUC	0.96			0.94			0.96			0.93			0.98
Boys
Number of social grants
Between
Within
Average	1.05 (0.92–1.20)	0.484	1.000	1.02 (0.86–1.21)	0.783	0.354	0.95 (0.77–1.16)	0.613	1.000	1.08 (0.97–1.21)	0.143	0.306	1.05 (0.87–1.27)	0.618	0.981
Positive caregiving
Between
Within
Average	1.00 (0.96–1.05)	0.924	1.000	0.94 (0.90–0.99)	0.030	0.031	1.01 (0.94–1.07)	0.855	1.000	1.06 (1.02–1.10)	0.002	0.015	1.05 (0.98–1.12)	0.155	0.633
Caregiver supervision
Between
Within
Average	0.93 (0.90–0.95)	< .001	0.001	0.93 (0.90–0.96)	< .001	0.001	0.94 (0.91–0.98)	0.002	0.013	0.99 (0.96–1.01)	0.242	0.320	0.92 (0.88–0.96)	< .001	0.001
Adolescent-caregiver communication
Between
Within
Average	1.04 (0.96–1.12)	0.319	1.000	1.03 (0.93–1.12)	0.593	0.311	0.97 (0.87–1.08)	0.622	1.000	0.96 (0.90–1.02)	0.187	0.306	0.95 (0.85–1.06)	0.349	0.871
Education enrolment
Between										0.69 (0.28–1.75)	0.438	0.487
Within										2.78 (1.18–6.58)	0.020	0.064
Average	1.68 (0.80–3.53)	0.168	0.725	5.56 (1.91–16.24)	0.002	0.006	1.52 (0.56–4.12)	0.410	1.000				1.27 (0.50–3.22)	0.619	0.981
Days with enough food
Between
Within
Average	1.03 (0.82–1.29)	0.823	1.000	0.77 (0.62–0.96)	0.019	0.026	1.06 (0.75–1.51)	0.738	1.000	0.99 (0.82–1.18)	0.897	0.513	1.05 (0.76–1.46)	0.766	1.000
Variance components
Level 2: In level-1 intercept	2.69			0.00			1.70			0.66			3.18
Goodness of fit
AUC	0.97			0.86			0.96			0.88			0.98

Coefficients for additional covariates are available in Additional file 1: Table 10
Abbreviations: aOR Adjusted odds ratio, CI Confidence interval, AUC Area under curve
^aSex on substances was only measured at wave two and wave three
^bWald p-value for significance. A significant p-value indicates that a coefficient is significantly different from 1.00. The AUC statistic ranges from 0 to 1 and gives an idea of how well a model is able to distinguish between positive and negative outcomes. The higher the AUC, the better the model is at correctly classifying outcomes

Back to article page

ISSN: 1471-2458

Contact us

Submission enquiries: bmcpublichealth@biomedcentral.com
General enquiries: ORSupport@springernature.com

BMC Public Health

Contact us