Table 6.

Multivariable evaluations of the association between explanatory variables and the presence of LSIL+ lesions

Type of analysis	Study groups included	Number of observations*	Explanatory variables included in the model	Adjusted OR	95 % CI
Logistic regression	1 (HIV screening centres HIV-positive women)	124	Age of first sexual intercoursea	0.86	0.72 – 1.02
			Number of pregnancies
			0 – 2	1
			3 – 5	0.95	0.31 – 2.86
			6 or more	0.83	0.28 – 2.44
			Abortion	1.98	0.83 – 4.75
Logistic regression	2 (HIV screening centres HIV-negative women)	123	Agea	1.11	0.95 – 1.31
			Marital status
			married	1
			single	5.26	0.27 – 102.32
			widowed	4.34	0.18 – 105.99
			divorced	§
			Number of pregnancies
			0 – 2	1
			3 – 5	1.15	0.07 – 18.34
			6 or more	4.27	0.22 – 81.82
Logistic regression	3 (Hospitals HIV status unknown)	502	Formal employment	2.08	0.74 – 5.88
			Abortion	1.65	0.53 – 5.14
			≥3 lifetime sexual partners	1.75	0.69 – 4.46
			Plants for vaginal care	2.85	0.75 – 10.82
Logistic regression	4 (Health centres HIV status unknown)	203	Age of first sexual intercoursea	0.86	0.68 – 1.07
			Formal employment	0.30	0.06 – 1.50
GEE	1, 2 3 and 4	886	Abortion	1.60	0.97 – 2.63
			≥3 lifetime sexual partners	1.29	0.83 – 1.99
			Chemicals for vaginal care	0.65	0.37 – 1.14
Logistic regression	3 and 4	643	Alcohol consumption	1.76	0.80 – 3.86
			≥3 lifetime sexual partners	1.58	0.72 – 3.46
			Plants for vaginal care	2.70	1.04 – 7.01**

LSIL+: cytology findings compatible with (pre)cancerous lesions (includes low- and high-grade squamous intraepithelial lesions and carcinoma)

OR odds ratio; 95 % CI 95 % confidence interval of the odds ratio

GEE: generalized estimating equations (population-averaged model; group variable: study group)

^aAge was treated as a continuous variable

§ Odds ratio could not be calculated because there were cells without observations

**P-value of Wald test < 0.05