Adherence to Diaphragm Use for Infection Prevention: A Prospective Study of Female Sex Workers in Kenya

Table 3

Associations between consistent diaphragm use with “helping” partner during previous 2 weeks and selected demographic characteristics and attitudes toward diaphragm use, Nairobi, Kenya, 2004-2005∗.

	No. of intervals with consistent use	No. of intervals without consistent use	Bivariable model		Multivariable model†
			OR	(95% CI)	OR	(95% CI)
Time-independent factors
Age
≤27 years	63	49	Referent
28–34 years	75	45	1.4	(0.7, 2.7)
≥35 years	46	35	1.1	(0.5, 2.4)
Marital status
Never married and cohabiting	57	34	1.2	(0.6, 2.2)
Cohabiting, divorced or widowed	127	95	Referent
Education completed
0–8 years	141	104	Referent
9–12 years	43	25	1.3	(0.6, 2.8)
Parity
0-1 children	45	35	Referent
≥2 children	139	94	1.2	(0.6, 2.3)
Time-dependent factors
Study follow-up visit
2-month	55	53	Referent
4-month	63	40	1.4	(0.9, 2.2)
6-month	66	36	1.5	(1.0, 2.4)
All sex partners in past 2 weeks
1–5	106	61	1.2	(0.8, 1.9)
6–42	78	68	Reference
Coital acts with all partners in past 2 weeks
0–5 acts	31	17	1.5	(0.7, 2.8)
6–15 acts	109	73	1.3	(0.8, 2.2)
≥16 acts	44	39	Referent
Consistent condom use with “helping” partner in past 2 weeks
Yes	57	26	1.5	(0.9, 2.6)
No	127	103	Referent
Component 1 (Perceived partner support of diaphragm use)			1.3	(1.1, 1.7)	1.4	(1.1, 1.7)
Component 2 (Attitudes toward study product attributes)			1.2	(1.0, 1.5)
Diaphragm  +  gel use is easier than condom use
Yes	140	81	1.9	(1.2, 3.1)	2.0	(1.2, 3.1)
No	44	48
Important to participant that diaphragm use does not interrupt sex
Yes	176	119	2.7	(1.1, 6.6)	2.8	(1.1, 7.1)
No	8	10	Referent

OR  =  odds ratio; CI  =  confidence interval
∗Analyzed with logistic regression model with generalized estimating equations based on 313 intervals from 121 women who reported having sex with “helping” partner during at least one follow-up interval.
†Adjusted for all variables in column.
The following variables also were analyzed but were not associated with consistent diaphragm use: ethnicity (Kikuyu versus other), education ( ≤9 USD versus >9 USD); important to prevent pregnancy (not at all versus moderately or a lot); worry about pregnancy (not at all versus moderately or a lot); worry about HIV (not at all versus moderately or a lot); new main sex partners in past 2 weeks (0 versus 1-2 versus ≥3); under the influence of alcohol during sex with “helping” partner in past 2 weeks (never versus ≥1 time); under the influence of “bhang” or other drugs during sex with “helping” partner in past 2 weeks (never versus ≥1 time); and the remaining factors that did not load in principal component analysis (listed in the footnote for Table 2).