Menu
Menu
Menu
Menu
Introduction: HIV prevalence among pregnant women in Kenya is high. Furthermore, there is a high risk of maternal mortality, as many women do not give birth with a skilled healthcare provider. Previous research suggests that fears of HIV testing and unwanted disclosure of HIV status may be important barriers to utilizing maternity services. We explored relationships between women’s perceptions of HIV-related stigma and their attitudes and intentions regarding facility-based childbirth. Methods: 1,777 pregnant women were interviewed at their first antenatal care visit. We included socio-demographic characteristics, stigma scales, HIV knowledge measures, and an 11-item scale measuring health facility birth attitudes (HFBA). HFBA includes items on cost, transport, comfort, interpersonal relations, and services during delivery at a health facility versus at home. A higher mean HFBA score indicates a more positive attitude towards facility-based childbirth. The mean HFBA score was dichotomized at the median and analyses were conducted with this dichotomized HFBA score using mixed effects logit models. Results: Women who anticipated HIV-related stigma from their male partner had lower adjusted odds of having positive attitudes about giving birth at the health facility (adjusted OR =. 63, 95% CI 0.50-0.78) and less positive attitudes about health facility birth were strongly related to women’s intention to give birth outside a health facility (adjusted OR = 5.56, 95% CI 2.69-11.51). Conclusions: In this sample of pregnant women in rural Kenya, those who anticipated HIV-related stigma were less likely to have positive attitudes towards facility-based childbirth. Furthermore, negative attitudes about facility-based childbirth were associated with the intention to deliver outside a health facility. Thus, HIV-related stigma reduction efforts might result in more positive attitudes towards facility-based childbirth, and thereby lead to an increased level of skilled birth attendance, and reductions in maternal and infant mortality. © 2012 Medema-Wijnveen et al.
This study received ethical approval from the Kenya Medical Research Institute (KEMRI) Ethical Review Committee and the Committee on Human Research of the University of California, San Francisco. The Maternity in Migori and AIDS Stigma (MAMAS) Study was conducted at sites supported by Family AIDS Care and Education Services (FACES), an HIV prevention, care, and treatment program operating in Nyanza Province Kenya. HIV prevalence among women of reproductive age in Nyanza Province is estimated at 16% [3]. Women at least 18 years old in their first seven months of pregnancy, who were visiting the antenatal care clinic (ANC) for the first time in their pregnancy and did not know their current HIV status (never tested or tested negative more than 3 months ago) were recruited at nine governmental health facilities. After obtaining signed informed consent, participants were interviewed in their preferred language (Dholuo, Kiswahili, or English) by a trained interviewer. Subsequently all women were offered voluntary HIV counselling and testing during the ANC visit, followed by post-test counseling and antiretroviral drugs for PMTCT for those who tested HIV-positive, as per Kenyan national guidelines [25]. Information on their acceptance of HIV testing and their test result were subsequently obtained from the women’s medical charts. Recruitment, baseline interviews and HIV testing took place between November 2007 and April 2009. To measure perceptions of HIV-related stigma at baseline, two stigma scales developed in sub-Saharan Africa were included in the survey. Anticipated stigma is the anticipation that one will personally experience specific types of stigma or discrimination if one is found to be HIV-positive and others learn of one’s HIV status. This type of stigma was measured by a nine-item scale originally developed in Botswana that also captures from whom they expected this stigma: male partner (break-up of relationship, physical abuse), family members (neglect, denial of care), or others (treatment as outcast, bad treatment at school or work, bad treatment by health workers, loss of friends, loss of job) [26]. Researchers in Botswana found this scale to have high internal reliability consistency (Cronbach’s alpha = 0.77) [26]. In the current study, a total anticipated stigma score was calculated by taking the mean of all responses for women who provided responses to at least 6 of the 9 scale items, and women who provided 5 or fewer responses were coded as missing. We furthermore created dichotomous measures of anticipated stigma for the different ‘sources’ of stigma. Given our previous analyses showing the importance of male partner stigma, as compared to other dimensions of stigma, as a predictor of HIV test refusal in this population [27], we chose this as our primary anticipated stigma variable. Analyses were conducted using all of the different anticipated stigma measures, but we chose to construct a more parsimonious model including the anticipated male stigma variable, which had the strongest relationship with the outcome. To measure general attitudes and perceptions about persons living with HIV (PLWH) and how they are treated in their community, referred to as perceived community stigma in this paper [27], [28], we used a 22-item scale developed by Genberg et al. [29], [30] with items like ‘People living with HIV/AIDS deserve to be punished’ and ‘People living with HIV/AIDS in this community face rejection from their peers’. This scale was found to have high internal consistency validity and good divergent validity in both Thailand and Zimbabwe [30]. As reported elsewhere [27], internal-consistency reliability of these scales in our sample of pregnant women in Kenya was high (Cronbach’s alpha.86 for the anticipated stigma scale and.85 for the perceived community stigma scale). To measure women’s perceived quality of care and attitudes towards giving birth at the health facility, we developed an 11-item scale derived from our previous qualitative research in Nyanza Province. In that study, we identified common beliefs and attitudes towards giving birth at the health facility assisted by a skilled healthcare provider versus giving birth at home assisted by a TBA, based on in-depth interviews with postpartum women, male partners, TBAs, and health workers [31]. This resulted in 11 statements with which women were asked to agree or disagree (coded 1 or 0). A total HFBA score was calculated by taking the mean score for women who provided responses to 8 or more scale items, and women who provided 7 or fewer responses were coded as missing, with a higher mean score indicating a more positive attitude towards giving birth at a health facility. The internal-consistency reliability of this score in this sample was found to be moderate (Cronbach’s α = .60), but acceptable for research in the early stages [32]. The survey also included questions on socio-demographic characteristics. Questions that assessed knowledge regarding HIV transmission and prevention (for example, can a person get HIV by sharing a meal with someone who is infected?) were included and were used to construct a HIV knowledge index based on the number of items that were answered correctly. The primary outcome variables for our analyses of women’s intentions regarding childbirth were based on questions in which participants were asked where they intended to give birth and with what type of assistance. Participants were also asked about expected costs related to their delivery (transportation, supplies, medicines, fees). We considered giving birth in a health facility to be equivalent to giving birth with a skilled health care provider (although recognizing that the quality of birth attendance provided in health facilities may not be optimal) [33]. Initial analyses were conducted using SPSS 16.0. [34]. With HFBA as the dependent variable, initial tests of bivariable associations were conducted using the Pearson’s Chi-square for nominal or ordinal variables, student T-test and ANOVA for continuous variables, and Kruskal-Wallis-test or the Mann-Whitney U-test for continuous variables with non-normal distributions. Following these initial analyses, HFBA was dichotomized into high and low, with the median (.73) chosen as a cut-off point. After conducting unadjusted logistic regression analyses to identify significant associations, we ran a mixed-effects logit model [35] using Stata 11 [36]. This multivariable model accounted for clustering by site and included variables that were significantly associated with the HFBA score in bivariable analyses (p<.05), as well as other variables that have been shown to be important for childbirth decision-making in the literature. Specifically, we included socio-economic factors including age, educational level, and occupation, as these factors related to women’s empowerment have been shown to be important factors in decision-making regarding childbirth in Kenya and similar countries [3], [37]. These analyses were repeated with a continuous version of the HFBA variable (normally transformed) to examine the potential impact of dichotomization on our results. Similar analytic methods were used to examine the predictors of intended type of assistance during delivery, in that variables found to have statistically significant relationships with the outcome in bivariable analyses (p<.05) were included in the multivariable model (education, occupation, being in a polygamous relationship, HFBA, expectations to pay for transport or supplies, and anticipated stigma), as well as household television ownership as a measure of wealth and access to mass media [38]. Binomial regression was used to obtain estimates of risk ratios for the multivariable analyses, with adjustment for clustering by site. Finally, we conducted exploratory mediation analyses to examine if the effects of anticipated male partner HIV-related stigma on intended type of delivery might potentially operate through effects on HFBA. The conditions for mediation were met [39] as: 1) anticipated male partner stigma is significantly associated with HFBA, 2) HFBA independently predicts intended type of delivery assistance, and 3) the association of anticipated male partner stigma with intended type of birth assistance is reduced after adjustment for HFBA.
N/A
