The accuracy of women’s reports of their partner’s male circumcision status in Rakai, Uganda

Abstract

We assessed whether women had accurate knowledge of their partners’ male circumcision (MC) status using survey data (2010–2011) from Rakai, Uganda, and examined characteristics of women who misreported MC status. Among couples in which men were uncircumcised (N=1744), 8.2% women misreported; and among couples where men were confirmed circumcised (N=759), 1.2% women misreported. Younger women were 2.2 times more likely to misreport compared to older women. Misreporting was not associated with other sociodemographics or behavioral characteristics. If women are to act as advocates for MC acceptance, there is a need to educate women, particularly younger women about the nature and recognition of MC.

Introduction

Male circumcision (MC) is an important strategy for the prevention of heterosexually-acquired HIV infection in men and other sexually transmitted infections in both men and women [1–6]. WHO/UNAIDS have identified thirteen priority sub-Saharan Africa (S-SA) countries for MC scale-up[7]. While the long-term population-level impact of MC on the HIV epidemic depends on the pace of uptake and level of coverage in men, women can play an important role in men’s decision to accept MC. Qualitative studies in S-SA suggest that women’s groups may provide venues to mobilize their husbands or sex partners to accept MC[8–12]. Using women as mobilizers to facilitate MC acceptance by men, however, requires that women have accurate knowledge of their partners’ circumcision status. We examined the accuracy of women’s reports of their partner’s MC status using survey data on linked couples and MC service records from Rakai, Uganda, and identified characteristics of women who misreported their partners’ MC status.

Method

The Rakai Health Sciences Program (RHSP) conducts annual household census in 50 communities of rural Rakai District. Eligible consenting residents aged 15–49 years are then individually interviewed by trained same-sex RHSP staff who collect detailed sociodemographic and behavioral information using structured questionnaires. Men are specifically asked whether, when and where they were circumcised, and women are asked about each of their partner’s MC status (up to four partners). Interview data can be linked for couples who are married or in long-term consensual relationships using household census data on familial relationships. For this analysis, we used couples data from the survey conducted in 2010–2011 which is a period when information about the health benefits of MC for prevention of HIV and sexually transmitted infections had been disseminated, and free MC services were widely available and promoted in Rakai. This analysis is confined to non-Muslim couples, since nearly all Muslim men (who represent ~15% of Rakai men) were circumcised as infants and thus are not eligible for adult MC.

The MC status of the male partner in each couple was determined by the male’s self-report, validated by RHSP MC clinic records. RHSP conducted two MC trials during 2003–07[6, 13], and since 2007, RHSP has been the primary provider of MC services and the sole provider of free MC for the general male population in Rakai. Over 75% of men who self-reporting circumcised could be confirmed by RHSP clinic records (the additional 25% reported circumcision by private practitioners or unidentified sources, and thus could not be verified). All men self-reporting themselves as uncircumcised were verified not to have received MC by RHSP service records. Using data for men whose MC status could be validated by RHSP service records (either as being uncircumcised or circumcised), we assessed whether their female partners correctly reported the men’s MC status by cross tabulating women’s reports versus the validated MC status of men.

We further assessed whether female misreporting was associated with their sociodemiographic characteristics and sexual risk behaviors (condom use, number of sex partners, and sex-related alcohol use during the past year), using Pearson χ²-test or Fisher’s exact test. In addition, the association between availability of electricity in the household and the accuracy of women’s report was examined, as qualitative interviews with Rakai women had identified lack of electric lighting at home as a potential barrier to women’s knowledge about their sex partners’ MC status [JS, pc].

Results

There were 2714 non-Muslim couples identified, and the MC status for 2503 of the couples (92.2%) could be validated using RHSP MC service records. Of the 2503 couples, 759 men (30.3%) were confirmed to be circumcised, and 1.2% of their female partners incorrectly reported them as uncircumcised (n=8) or answered “don’t know” (n=1). Among the 1744 uncircumcised men, 8.2% of their female partners misreported them as circumcised (n=136) or answered “don’t know” (n=7). The overall rate of female misreporting of men’s MC status (including “don’t know” responses) was 6.1%.

Table 1 summarizes the sociodemographic characteristics and sexual risk behavior profiles of women and availability of electric lighting at home by accuracy of women’s report on their partners’ MC status. Women’s age was significantly associated with misreporting: the proportion of misreporting was 9.6% (78/812) in younger women (age<25 years), compared with 4.4% (74/1691) in women aged 25–49. Thus, the missreporting rate in younger women was 2.2 (95% CI 1.6– 3.0) times higher than in older women. The proportions of misreporting did not differ by the education, occupation, and risk behavioral profiles of women, nor by availability of electricity in the household.

Table 1.

Comparisons of sociodemiographic characteristics, sexual risk behaviors and availability of electric lighting at home between women who misreported and those who accurately reported their husbands’ MC status.

		Women who misreport partner’s MC statusa		Women who correctly reported partners MC status
		(n=152)	%	(n=2351)	%	P- value
Age						<.001
15–19		22	10.3	192	89.7
20–24		56	9.4	542	90.6
25–29		30	4.8	595	95.2
30–39		37	4.0	881	96
40+		7	4.7	141	95.3
Education						0.99
No education		9	6.0	141	94
Primary		98	6.0	1528	94
Secondary and plus		45	6.2	680	93.8
Occupation						0.12b
Agriculture related		101	5.6	1694	94.4
Housework or office work		16	5.8	259	94.2
Trading/vending, bar worker		30	9.2	298	90.9
Unemployed or casual laborer		5	4.8	100	95.2
Condum use in the past year						0.11b
Always		5	5.3	89	94.7
Sometimes		33	8.4	358	91.6
Never		114	5.7	1904	94.4
Alcohol use before sex						0.35
Sometimes		135	6.3	2025	93.8
Never		17	5.0	326	95
Number of sex partners in the past year						0.13b
0–1		146	6.0	2306	94
>=2		6	11.8	45	88.2
Having electricity at home						0.82
Yes		11	5.7	182	94.3
No		141	6.1	2169	93.9

^a: Women reporting “don’t know” were included.

^b: P-value was obtained using Fisher’s exact test when any cell size is less than 10.

Discussion

Our data indicate women who are married or in long-term stable consensual relationships may not have accurate knowledge of their sex partner’s MC status. Similar findings were reported for women (including both married and unmarried women in urban and rural Zambia (7% and 5% misreporting, respectively), and in urban Swaziland (13% misreporting)[14]. Our results suggest that MC prevalence estimated from women’s reports would overestimate the true MC prevalence, since female misreporting was more common when men were in fact uncircumcised.

Younger women were found more likely to misreport their partner’s MC status. In RHSP communities, approximately 35% of women aged 15–49 are not in a marital/consensual union, most of whom are younger. If such women were included in population-level surveys, the overall proportion of misreporting is likely to be higher than currently observed among women who were married or in stable consensual relationships. Additionally, compared to women from the 50 RHSP communities, women in the general population have less exposure to MC messages and thus may have an even higher misreporting rate.

Women with uncircumcised sex partners may have misreported them to be circumcised due to a social desirability bias. However, in depth interviews with Rakai women suggested that misreporting was more likely to be due to a lack of a clear understanding of MC [RG pc]. Therefore, using women as potential motivators to increase MC uptake in men will need to first educate women about the nature and recognition of MC, and such education should particularly target younger women. Older women in the communities who traditionally advise younger women on marriage related issues and behaviors [15] may also play the role of educating younger women about MC.