Matern Child Health J DOI 10.1007/s10995-014-1507-y
HIV Counseling and Testing for the Prevention of Mother-toChild Transmission of HIV in Swaziland: A Multilevel Analysis Marguerite L. Sagna • Donald Schopflocher
Ó Springer Science+Business Media New York 2014
Abstract HIV counseling and voluntary testing during antenatal care have been proven to reduce the risk of HIV transmission from mother to child, through increasing knowledge about safe behaviors, ascertaining HIV status and increasing coverage of effective antiretroviral regimens. However, it remains that, in developing countries where 95 % of mother-to-child HIV transmissions (MTCT) take place, such interventions are not widely accessible or available. Using a nationally representative cross-sectional household survey, the present study aimed to examine individual- and contextual-level influences on the receipt of HIV pre-test counseling and uptake of HIV testing during the antenatal care period in Swaziland, a country highly burdened by HIV/ AIDS. The study sample was restricted to women aged 15–49 years with a live birth in the past five years preceding the survey and who received antenatal care for the most recent birth. The findings of this study indicated that only 62 % of women received pre-test counseling for the prevention of MTCT and no more than 56 % of women consented to be tested for HIV during antenatal care. The multilevel regression analysis revealed that the likelihood of receiving HIV pre-test counseling increases significantly with higher parity, education level, household wealth and antenatal visits while it is lower in areas where poverty is pervasive (OR = 0.474) and in rural regions (OR = 0.598) as well. Beyond all the significant predictors, undergoing pre-test counseling has M. L. Sagna (&) D. Schopflocher Centre for Health Promotion Studies, School of Public Health, University of Alberta, 3-270 Edmonton Clinic Health Academy, 11405-87 Avenue, Edmonton, AB T6G 1C9, Canada e-mail: [email protected] D. Schopflocher Faculty of Nursing, University of Alberta, Edmonton, AB, Canada
emerged as an important determinant of HIV testing. Receiving pre-test counseling increases the odds of accepting an HIV test by 77 %. Evidence from this analysis underscores bottlenecks and challenges that persist in increasing the need for and uptake of HIV preventive and treatment services to stop new HIV infections among children. Keywords HIV pre-test counseling HIV testing Mother-to-child transmission Multilevel modeling Swaziland
Introduction Substantial gains in child survival rates are being increasingly erased in many countries in sub-Saharan Africa because of severe pediatric HIV/AIDS. The magnitude of childhood HIV infection is indisputably alarming in this part of the world, which accounts for 91 % of the global 3.4 million HIV positive children under age 15 [1]. Most of these children became HIV positive by vertical transmission or mother to-child-transmission (MTCT), that is, they acquired the virus from their HIV infected mothers during pregnancy, birth or breastfeeding. In recent years in efforts to drive the rates of new pediatric HIV infections down, many sub-Saharan African countries have adopted a model of service delivery, which integrates HIV counseling and testing as part of the antenatal care routine. However, coverage of such interventions is still inadequate across the region, where available estimates show that about 42 % women were tested for HIV in 2010 (ranging from 25 % in western and central Africa to 61 % in eastern and southern Africa [2]). Swaziland is no different. Despite a strong prevention of mother-to-child-transmission (PMTCT) program launched in 2003, the vertical transmission rate is still disconcerting with
123
Matern Child Health J
11 % of infants and children being infected by their HIVpositive mothers in 2011 [3]. In addition, the country is in the midst of the most severe generalized HIV epidemic in the world, with 25.9 % of its adult population aged 15–49 estimated to be HIV positive [4]. Current estimates also indicate that the rate of infection among women of reproductive age continues to be as high as 3.13 % nationally [2], making the proportion of pregnant women living with HIV/AIDS—about 41 %—the largest worldwide. The national strategy to optimize the impact of PMTCT services in Swaziland consisted of a shift from client-initiated to provider-initiated HIV testing and counseling in 2006 [5]. Under the provider-initiated approach, each pregnant woman should be provided with pretest HIV counseling and subsequently offered an HIV test that would be executed unless the woman declines. Yet, all too often national efforts to expand HIV counseling and testing to a greater number of pregnant women during antenatal care are impeded by women’s reluctance to be tested [6–8], thus missing the opportunity to know their sero-status, and to make informed decision about their health and the well-being of their newborns. Disparities in receiving pre-test counseling also occur, even in settings where this intervention has been structurally integrated in the antenatal care package [9, 10]. While factors associated with sporadic pre-test HIV counseling have received little attention [11], acceptability barriers for HIV testing among pregnant women have been well studied [9, 12–24]. Most published research has drawn on data from healthcare facilities (i.e. largely urban antenatal clinics), sentinel surveillance, and clinical trials. Very few attempts have been made to investigate community-level influences. This limitation is important since health care seeking behaviors are not purely dependent on individuals’ choice or circumstances but also determined by the structure of the social environment in which individuals live [25, 26]. In this paper, we use a population-based survey to examine differentials in the receipt of HIV pre-test counseling and uptake of HIV testing among pregnant women in Swaziland, accounting for the influence of both individual-level and contextual-level factors. With the number of HIV infected women of childbearing age continuously increasing, identifying factors associated with missed opportunities for PMTCT is of paramount importance to strengthening HIV prevention interventions and reducing the risk of MTCT and HIV infection in the population at large.
Methods Data The present study is based on a secondary analysis of cross sectional data derived from the Demographic and
123
Health Survey (DHS) conducted in Swaziland between 2006 and 2007. The DHS is a nationally representative household survey designed to provide information on a battery of issues including fertility, reproductive health, sexual behaviors, and HIV/AIDS related knowledge, attitudes and behaviors in resource-poor settings. The 2006/2007 Swaziland Demographic and Health Survey (SDHS) is the first of its kind to be conducted in this country. Of 5,301 eligible women age 15–49 identified for the survey in 5,500 selected households randomly sampled within 275 primary sampling units (PSUs), 4,987 were successfully interviewed, reflecting a response rate of 94 %. Detailed information about the design and implementation of the SDHS is described in the survey report [27]. The present study focuses on a restricted sample of 1,788 women age 15–49 with a live birth in the past 5 years preceding the survey who received antenatal care for the most recent birth. The sample is comprised of 274 PSUs, representing an average number of 6.5 women per PSU. Since this study involved analysis of secondary data, publicly available upon request, IRB approval was not required. Variables The outcome variables of interest were receipt of HIV pretest counseling during antenatal care, and uptake of HIV testing during antenatal care. Receipt of HIV pre-test counseling is coded 1 if the respondent has received essential information about vertical transmission of HIV during prenatal care and 0 otherwise. Uptake of HIV testing is coded 1 if the respondent consented for an HIV test and 0 otherwise. Choosing explanatory variables for predicting HIV pretest counseling, and HIV testing was guided by the Donabedian’s Quality of Care Framework [28, 29] and the Health Belief Model [30, 31], respectively. The Donabedian Framework suggests that quality of care is attributable to three dimensions: structure (i.e. the context in which care is delivered), process (i.e. interactions between patients and providers throughout the delivery of healthcare) and outcome (i.e. the impact of care on patients’ health status). The context in which care is delivered affects processes and outcomes while the outcomes indicate the combined effects of structure and process. The Health Belief Model argues that the likelihood for an individual to adopt a preventive behavior is affected by personal perceptions about a disease and the subjective weighing of the benefits of adopting the behavior. Based on these conceptual frameworks and informed by evidence from published literature, we included these individual-level covariates: age, marital status, parity, education level, household wealth, timing of first antenatal
Matern Child Health J
visit, number of antenatal visits, type of antenatal care medical provider, diagnosis of a sexually transmissible disease (STD) in the last 12 months, knowledge of MTCT of HIV, and exposure to media. Age is grouped into the following categories: 15–19, 20–29, 30–39, and 40–49 years of age. Marital status is classified according to whether the respondent was single, married or separated/ widowed. Parity is grouped into three categories including 1 child, 2 children, and 3 children or more. Education level refers to the highest level of education completed by the respondent and is grouped into four categories including no education, primary level, secondary level and tertiary level. Household wealth is assessed by five categories corresponding to the index wealth quintiles provided in the survey, including poorest, poor, middle, rich, and richest. Timing of first antenatal care is coded as a dichotomous variable indicating whether or not a woman had received her first antenatal visit within the first trimester. The number of antenatal care visits is recorded as three-category variable indicating whether the respondent has made 3 visits of less, 4 visits, and 5 visits or more. The type of medical personnel who provided antenatal care is assessed by three separate binary variables indicating whether the respondent received care from a doctor, midwife or auxiliary midwife. Diagnosis of STD was measured by binary variable expressing whether or not the respondent has been diagnosed with STD in the previous year. Knowledge of MTCT of HIV is analyzed by three separate dichotomous binary variables indicating whether or not the respondent knows that AIDS can be transmitted during pregnancy, childbirth or through breastfeeding. Exposure to media is assessed across three separate dichotomous variables on whether the respondent read newspapers, listened to radio or watched television at least once a week. The community-level covariates included integral variables and aggregate variables. The integral variables were namely community type (i.e. urban/rural) and region as collected by the survey. The primary sampling unit (PSU) was considered the community-level of analysis in the present study. Aggregate variables were derived from individual level responses by aggregating the individual characteristics of respondents within the to the PSU and then assigning the aggregate value back to each individual in the PSU as a separate variable. Given the scarcity of studies addressing community-level determinants of missed opportunities for PMTCT, a number of communitylevel factors were selected based on Stephenson et al. [32] research exploring community influences on HIV testing uptake among married men in Africa. For this analysis, we included variables that capture community economics, community behavior, and knowledge and community features. The community economics variables were percent-
age of women in community with secondary or higher education, and community poverty conceptualized as the percentage of women in community living in households that fall in the two lowest wealth index quintiles. The community behavior and knowledge variables were the percentage of women in community with knowledge of MTCT, the percentage of women in community against domestic violence, and the percentage of women in community who would want to keep secret the HIV/AIDS status of a family member. The community features variables were community type and province. Statistical analysis Our data analysis was conducted in two phases. First, we performed some descriptive analysis to examine the characteristics of the sample. Second, a multilevel regression modeling strategy was used to examine the influence of individual and community characteristics on interventions related to PMTC. This strategy accounts for the hierarchical nature of the data and corrects the estimated standards errors to allow for the clustering of observations within units [33]. We specified two-level multilevel logistic regression models separately to test predictors of HIV pretest counseling and uptake of HIV testing. The logit of the probability of each outcome is modeled as follows: Logit ðpij Þ ¼ logðpij =ð1 pij ÞÞ ¼ b0 þ b1 Iij þ b2 Cj þ lj where pij is the probability that the outcome is 1 for a respondent i in community j; b0 is the intercept; the b’s are the fixed coefficients; I and C represent individual and community variables respectively; lj * N(0, r2j ) represents the random effects for the jth community (274 community-specific random effects).The analytical strategy involves applying two models for each outcome. Model 1 includes only the individual-level variables as predictors while model 2 adds the community-level variables, to examine the influences of each set of variables and, to test whether specific components of the community independently influence HIV counseling and testing after adjusting for individual characteristics. Due to the complex design of the survey, all analyses are adjusted for the multistage sampling design and the sample weights are used to allow for adjustment of how the sample was collected. All statistical analyses are estimated in Stata version 12 [34].
Results Table 1 presents the characteristics of women of reproductive age (15–49) in the study sample. HIV counseling is
123
Matern Child Health J Table 1 Sample characteristics of women aged 15–49 with adjustment for survey design, 2006/2007 Swaziland Demographic and Health Survey (n = 1,788)
Table 1 continued
Variables
SE
Listens to radio at least once a week Watches television at least once a week Yes
Mean
Variables
Yes
Individual-level covariates
Mean
SE
78.6
0.011
38.2
0.018
HIV pre-test counseling
62.4
0.014
HIV testing Current age (in years)
56.1
0.014
\20
19.5
0.010
Percentage of women in community with secondary or higher education
57.1
0.012
20–29
52.8
0.013
Community poverty
36.1
0.018
30–39
24.0
0.011
63.7
0.010
40–49
3.7
0.005
Percentage of women in community with knowledge of mother-to-child transmission of HIV
0.009
39.7
0.014
Percentage of women in community against domestic violence
66.1
Single Married
55.0
0.015
26.6
0.015
5.3
0.005
Percentage of women in community who would want to keep secret the HIV/AIDS status of a family member
1 child
32.1
0.011
Urban
23.3
0.023
2 children
24.3
0.011
76.7
0.023
3 children or more
43.6
0.012
Rural Province Hhohho
26.6
0.024
No education Primary
8.0 33.0
0.008 0.013
Manzini
32.4
0.028
Shiselweni
21.6
0.012
Secondary
52.7
0.015
Lubombo
19.4
0.014
6.3
0.008
Marital status
Separated/widow Parity
Community-level covariates (n = 274 PSUs)
Community type
Education level
Tertiary Household wealth index Poorest
18.9
0.016
Poor
19.4
0.013
Middle
20.5
0.015
Rich
20.3
0.014
Richest
20.9
0.017
25.9
0.011
Timing of first antenatal visit First trimester Number of antenatal visits 3 visits or less
15.9
0.009
4 visits
18.9
0.009
5 visits or more
65.2
0.012
3.1
0.005
Midwife
80.9
0.015
Auxiliary midwife
16.0
0.014
6.4
0.006
77.4
0.011
91.6
0.007
Type of antenatal care provider Doctor
Diagnosis of STD in the past year Yes Knows that AIDS can be transmitted during pregnancy Yes Knows that AIDS can be transmitted during delivery Yes
Knows that AIDS can be transmitted through breastfeeding Yes
87.6
0.009
58.3
0.016
Reads newspaper at least once a week Yes
123
not a universal practice, as only 62 % of women have received information about MTCT of HIV. The uptake of HIV testing is similarly low, with about 56 % of women who accepted to be screened for HIV during antennal care. Table 2 presents results of the multilevel logistic regressions for receiving pre-test HIV counseling during prenatal care. At the individual-level (Model 1), the odds of receiving HIV pre-test counseling are significantly associated with parity, education level, household wealth index, and the number of antenatal visits. Women of parity 3 or higher (OR 1.562) are significantly more likely to receive HIV pre-test counseling than women of parity 1. The odds of receiving a pre-test counseling increase significantly with the level of education; women with a primary (OR 1.542), secondary (OR 2.317) or higher education (OR 2.538) have significantly greater odds of are receiving HIV pre-test counseling than women with no education. Living in households that fall in the poor (OR 1.483), rich (OR 1.552) and richest (OR 1.500) wealth quintiles is associated with higher odds of getting HIV pre-test counseling compared to living in households that fall in the poorest wealth quintile. The odds of receiving HIV pre-test counseling increase by about 61 % for women who attended four antenatal visits and 63 % for women attended five or more prenatal visits.
Matern Child Health J Table 2 Results of the logistic analysis of the predictors of the receipt of HIV counseling during prenatal care, 2006/07 SDHS (n = 1,788) Model 1
Model 2
Odds ratio
95 % CI
Odds ratio
95 % CI
20–29
1.117
0.784, 1.591
1.027
0.721, 1.464
30–39
1.108
0.701, 1.752
1.003
0.634, 1.587
40–49
1.041
0.568, 1.905
0.970
0.529, 1.779
0.859
0.671, 1.100
0.921
0.718, 1.181
0.811
0.486, 1.351
0.782
0.467, 1.307
2 children
1.203
0.891, 1.625
1.230
0.912, 1.661
3 children or more
1.562**
1.110, 2.198
1.606**
1.140, 2.262
Primary
1.542*
1.037, 2.293
1.453
0.978, 2.159
Secondary
2.317***
1.551, 3.462
2.183***
1.453, 3.280
Tertiary
2.538**
1.368, 4.709
2.386**
1.274, 4.468
Poor
1.483*
1.061, 2.072
1.316
0.942, 1.838
Middle
1.287
0.914, 1.812
0.974
0.681, 1.393
Rich
1.552*
1.085, 2.218
0.968
0.666, 1.461
Richest
1.500*
1.029, 2.187
0.778
0.491, 1.232
0.917
0.706, 1.191
0.896
0.690, 1.164
1.610**
1.139, 2.276
1.642**
1.163, 2.318
1.631***
1.207, 2.204
1.669***
1.236, 2.253
Midwife
0.942
0.594, 1.494
0.990
0.625, 1.567
Auxiliary midwife
0.733
0.360, 1.490
1.344
0.842, 2.144
Percentage of women in community with secondary or higher education
0.885
0.391, 2.002
Community poverty
0.473*
0.233, 0.961
0.598**
0.415, 0.863
Manzini
1.619**
1.176, 2.228
Shiselweni
1.243
0.889, 1.738
Lubombo
0.761
0.540, 1.071
Current age (ref = less than 20)
Marital status (ref = single) Married Separated/widow Parity (ref = 1 child)
Education level (ref = no education)
Household wealth index (ref = poorest)
Timing of first antenatal visit (ref = subsequent trimesters) First trimester Number of antenatal visits (ref = 3 visits or less) 4 visits 5 visits or more Type of antenatal care provider (ref = doctor)
Community-level covariates a
Community type (ref = urban) Rural Province (ref = Hhohho)
*** p \ 0.001; ** p \ 0.01; * p \ 0.05;
p \ 0.10
After adjusting for individual-level factors (model 2), we found that community-level covariates such as community poverty, the percentage of women in community with knowledge of MTCT, community type and, province of residence have an influence on the receipt of HIV pretest counseling. The likelihood of undergoing pre-test counseling is significantly lower in areas where a large proportion of women live in poor households. The odds of receiving HIV pre-test counseling decrease by about 40 %
for women living in rural areas compared to women residing in urban areas. Further, living in Manzini (OR 1.619) increases the odds of receiving HIV pre-test counseling than living in Hhohho. Table 3 presents results of multilevel logistic regressions of factors associated with the uptake of HIV testing during antenatal care. Results of Model 1 show that the odds of consenting for an HIV test are significantly lower for women of parity 2 (OR 0.625) and women of parity 3 or
123
Matern Child Health J Table 3 Results of the logistic analysis of the predictors of consenting for HIV testing during prenatal care, 2006/07 SDHS (n = 1,640) Model 1
Model 2
Odds ratio
95 % CI
Odds ratio
95 % CI
20–29
1.178
0.815, 1.702
1.145
0.793, 1.652
30–39
1.137
0.703, 1.837
1.113
0.689, 1.794
40–49
1.567
0.728, 3.370
1.512
0.704, 3.244
0.812
0.614, 1.071
0.832
0.629, 1.099
0.795
0.449, 1.405
0.789
0.445, 1.395
2 children
0.625**
0.442, 0.883
0.623**
0.441, 0.878
3 children or more
0.648*
0.441, 0.953
0.641*
0.436, 0.941
Primary
1.475
0.926, 2.346
1.454
0.915, 2.307
Secondary
1.649*
1.016, 2.674
1.634*
1.002, 2.663
1.915
0.939, 3.902
2.033
0.990, 4.171
Poor
1.104
0.754, 1.613
1.083
0.741, 1.581
Middle
1.155
0.776, 1.717
1.135
0.748, 1.722
Rich
1.516
0.991, 2.317
1.369
0.858, 2.182
Richest
1.024
0.641, 1.634
0.855
0.488, 1.494
1.046
0.781, 1.399
1.047
0.782, 1.401
0.597*
0.401, 0.886
0.615*
0.414, 0.911
0.828
0.579, 1.182
0.839
0.588, 1.195
Midwife
0.488
0.234, 1.016
0.490
0.235, 1.021
Auxiliary midwife
0.382*
0.175, 0.828
0.403*
0.186, 0.874
1.772***
1.374, 2.286
1.667***
1.289, 2.153
0.623*
0.389, 0.997
0.605*
0.378, 0.9679
0.933
0.700, 1.241
0.886
0.663, 1.183
1.285
0.818, 2.016
1.285
0.820, 2.013
1.114
0.773, 1.604
1.105
0.768, 1.588
0.786
0.594, 1.040
0.789
0.596, 1.042
1.067
0.795, 1.431
1.108
0.826, 1.483
0.911
0.670, 1.236
0.948
0.698, 1.286
Percentage of women in community with secondary or higher education
0.793
0.322, 1.949
Community poverty
1.003
0.457, 2.195
Current age (ref = less than 20)
Marital status (ref = single) Married Separated/widow Parity (ref = 1 child)
Education level (ref = no education)
Tertiary
Household wealth index (ref = poorest)
Timing of first antenatal visit (ref = subsequent trimesters) First trimester Number of antenatal visits (ref = 3 visits or less) 4 visits 5 visits or more Type of antenatal care provider (ref = doctor)
HIV pre-test counseling (ref = no) Yes Diagnosis of STD in the past year (ref = no) Yes Knows that AIDS can be transmitted during pregnancy (ref = no) Yes Knows that AIDS can be transmitted during delivery (ref = no) Yes Knows that AIDS can be transmitted through breastfeeding (ref = no) Yes Reads newspaper at least once a week (ref = no) Yes Listens to radio at least once a week (ref = no) Yes Watches television at least once a week (ref = no) Yes Community-level covariates
123
Matern Child Health J Table 3 continued Model 1 Odds ratio
Model 2 95 % CI
Odds ratio
95 % CI
Percentage of women in community with knowledge of mother-to-child transmission of HIV
1.395
0.552, 3.521
Percentage of women in community against domestic violence
0.688
0.253, 1.870
Percentage of women in community who would want to keep secret the HIV/AIDS status of a family member
0.457
0.177, 1.179
0.777
0.515, 1.171
2.004***
1.426, 2.814
Shiselweni
1.334
0.933, 1.907
Lubombo
1.550*
1.047, 2.295
Community type (ref = urban) Rural Province (ref = Hhohho) Manzini
*** p \ 0.001; ** p \ 0.01; * p \ 0.05;
p \ 0.10
more (OR 0.648) than women of parity 1. Women with a secondary level of education have 65 % higher odds of being screened for HIV than women with no education, everything else being equal. Women with four antenatal visits have 40 % lower odds of accepting HIV testing than women with three or less prenatal visits. Women who received prenatal care from an auxiliary midwife (OR 0.382) were less likely to consent for an HIV test than women who were seen by a doctor. The odds of accepting HIV testing are 77 % greater for women who received a pre-test counseling than women who did not get any counseling. A diagnosis of STD in the previous year (OR 0.623) is negatively associated with the likelihood of consenting for an HIV test during prenatal care. Community level influences are presented in Model 2. Net of individual factors, province is significantly associated with the likelihood of consenting for an HIV test. Women in the Manzini province (OR 2.004) and the Lubombo province (OR 1.550) are more likely to accept HIV testing than women living in the Hhohhio province.
Discussion The high levels of pediatric HIV infection in many subSaharan African countries has led to both the implementation of, and a surge of research that examines PMTCT among pregnant women and its uptake. We believe that the present analysis constitutes a significant improvement on previous studies for several reasons. First, individual- and community-level factors associated with receipt of HIV pre-test counseling and HIV testing during antenatal care are explored using multilevel modeling to provide a contextual analysis of issues surrounding PMTCT. In contrast with the majority of previous studies, the data used here
were drawn from a nationally representative population survey designed to collect health information including reproductive health and HIV/AIDS. This research also contrasts with prior studies, because it examines pre-test counseling and HIV testing individually and emphasizes their shared importance in PMTCT. Finally, results presented here have clinical and public health implications and are relevant for informing HIV prevention planning efforts. In Swaziland, although virtually all women (97 %) attended antenatal care from skilled providers in the 5 years preceding the survey [27], no more than 62 % of women received pre-test counseling for the prevention of MTCT and only 56 % of women consented to be tested for HIV during antenatal care. These low rates have implications for the success of PMTCT programs given that the brunt of pre-test counseling trickle down to HIV testing, a gateway for maternal HIV infection screening, modification of infant feeding, and use of antiretroviral medications. Disparities in receiving pre-test counseling are influenced by a range of individual and community factors. At the individual-level, one of the most significant findings is the positive association between education and HIV pre-test counseling. Consistent with previous research [11], we found that women with higher level of education are more likely to undergo pre-test counseling. Since educated women are more likely to assimilate health information, health providers may be more motivated to discuss PMTCT with them. Odds of receiving counseling about HIV during prenatal care are also greater among women of parity 3 or more compared to women of parity 1. This is likely linked to differentials in HIV prevalence between pregnant women in younger age groups and their peers in older age groups. Women of higher parity tend to be older (aged 30–39 and 40–49), and available estimates have shown that as the HIV/AIDS epidemic has progressed in
123
Matern Child Health J
Swaziland, the prevalence of the disease among pregnant women has increased among older pregnant women while it has decreased among the younger age group (15–19 years) [5]. The pervasiveness of the HIV infection among older women may explain the determination of health care providers to offer counseling services to higher parity women. Household wealth is also predictive of HIV pre-test counseling, as high household income level is associated with a greater likelihood of receiving of HIV pre-test counseling. Women living in household with higher economic standing may experience greater financial advantages that allow them to easily navigate the health care system and therefore to seek care in reputable medical facilities or private clinics usually reputed to provide better care than government-operated facilities. The finding that the greater odds of receiving counseling are associated with higher frequency of visits suggests that compliance to the World Health Organization’s recommended number of prenatal checkups (i.e. at least four visits) is protective against missed opportunities for HIV pre-test counseling, as consistent use of antenatal care increasingly exposes a woman to information about PMTCT. The results of this study also demonstrate that pre-test counseling is influenced by contextual factors. The higher the concentration of poverty in a community, the lower the odds of receiving HIV counseling during prenatal care, likely because impoverished neighborhoods tend to harbor medical facilities that are poorly staffed. Since health providers are burdened with greater workload there will be less time to provide one-to-one counseling. Moreover, women in rural areas are less likely to be counseled about HIV transmission and prevention during prenatal care. A gap in the quality of care may also account for this difference. Counseling requires time and interaction with patients, and for many health providers in rural areas addressing the need for individual counseling is a difficult challenge. The reality is that majority of medical facilities in rural areas suffer from lack of resources, lack of trained staff to shortage of personnel, putting medical providers to a certain level of stress that impede them from providing adequate or comprehensive care to their patients. Pre-test counseling is also influenced by the province of residence; likely attributable to differences in content and quality of care dispensed at health facilities across the country. The present study also identified barriers and facilitators to HIV testing among pregnant women in Swaziland at multiple levels. At the individual-level, a number of factors are found to be associated with the uptake of HIV testing. Women with high parity are more likely to refuse HIV testing, likely because they do not perceive themselves at a high risk, especially if they have been tested negative during their previous pregnancies. Consistent with prior work [12, 21], uptake of HIV testing is greater among
123
educated women, likely due to the fact they are better informed and understand the benefits of HIV prevention interventions and are thus more amenable to informed choices about their health and health care. Though increasing numbers of antenatal checkups up to four visits was positively associated with pre-test counseling in the previous analysis, it was negatively associated with the uptake of HIV testing. This suggests that considerable efforts to delineate the importance of HIV testing need to be deployed during the very first antenatal checkup to minimize missed opportunities for testing. Women who received antenatal care from an auxiliary midwife/nurse are also more likely to decline HIV testing. This may reflect a lack of trust patients may have in this category of health providers who may be considered less qualified. It may also reflect the quality of information on PMTC provided by an auxiliary midwife/nurse. The association between pre-test counseling and HIV testing is noteworthy. Women who received pre-test counseling are significantly more likely to consent to an HIV test. This underscores the importance of pre-test counseling as a springboard for HIV testing, and is consistent with work from several countries including Zambia, Burundi, Kenya, Rwanda, and Uganda [18]. Pre-test counseling goes beyond providing information about the importance of HIV testing, to raising awareness about risk reduction strategies, family planning options, and connecting patients to appropriate treatments as needed. Further, in agreement with a previous study [14], a history of a prior STD is significantly associated with refusal of HIV testing. Reasons for this may have to do with patients’ perception of already being HIV infected, the fact that they have been previously tested as part of the medical care following STD diagnosis, or simply because they may be in denial about their risk for HIV. At the community-level after adjusting for individual characteristics, province of residence was associated with the uptake of testing. Women living in Manzini and Lubombo are more likely to consent for HIV testing than their Hhohho counterparts. This is noteworthy considering that Hhohho is the most urbanized region and has the highest nationwide HIV prevalence with 21.0 % of the population 2 years or older being HIV positive, while Manzini had 19.1 % and Lubombo had 18.1 % [27]. No other community-level variables were found to influence the uptake of HIV testing during antenatal care. This contrasts with a previous study examining contextual influences of HIV testing among men in several sub-Saharan African countries [32] where community economics, community knowledge of HIV, and community tolerance of violence towards women were associated with HIV testing. There were number of limitations in this study. First, this analysis uses PSUs as the higher-level unit of analysis.
Matern Child Health J
This may be problematic since the PSUs are created for enumeration purpose and do not necessarily correspond to actual communities. Second, the present research is based on data collected in 2006/07, and it is likely much has changed in Swaziland since then. In fact, a recent report has shown that the percentage of child HIV infections from HIV positive women has decreased from 18 % in 2009 to 11 % in 2011 [35], mirroring changes in preventive careseeking behaviors as well as improvement of quality of PMTCT interventions and access to services. As well, our study was unable to examine the effects of factors such as fear of repercussions (e.g. discrimination, divorce or domestic violence) [13, 24], male support/consent [36], and issues related to privacy and confidentiality in health care [20]. Finally, the survey was cross-sectional and therefore causal inferences are difficult to draw and impossible to confirm. Despite these drawbacks, this study does critically examine factors influencing disparities in receiving pre-test counseling and uptake of HIV testing among pregnant women at multiple levels. This is relevant to HIV prevention programs in a developing country context, since pre-test counseling is conducive to HIV testing, and getting tested for HIV has ripple effects. Such testing allows a woman to become aware of her HIV status, to start using antiretroviral medications, to make decisions about breastfeeding options and to obtain early HIV testing for her baby. Disparities in pre-test counseling suggest that better compliance of health professionals in providing adequate prenatal care is worthy striving for. Of course, this will remain a challenge if the shortage of health care workers endemic in most sub-Saharan African countries cannot be addressed. In conclusion, this investigation underscores the need for multilevel public health interventions to further promote HIV counseling and testing in order to prevent HIV transmission in all segments of the population.
References 1. UNAIDS. (2001b). Counselling and voluntary HIV testing for pregnant women in high HIV prevalence countries: Elements and issues. Retrieved October 2, 2012, from data.unaids.org/Publica tions/IRC-pub01/jc245-couns_test_en.pdf. 2. WHO, UNAIDS, UNICEF. (2011). Global HIV/AIDS responseEpidemic update and health sector progress towards universal access. Retrieved October 2, 2012, from http://www.unaids.org/ en/media/unaids/contentassets/documents/unaidspublication/2011/ 20111130_UA_Report_en.pdf. 3. UNICEF. (2012). Countdown to zero: Elimination of new HIV infections among children by 2015 and keeping their mothers alive. Retrieved October 2, 2012, from http://www.unicef.org/ aids/files/hiv_pmtctfactsheetSwaziland.pdf. 4. UNAIDS. (2010). Globla report: UNAIDS report on the global AIDS epidemic 2010. Retrieved October 2, 2012, from http:// www.unaids.org/documents/20101123_globalreport_em.pdf.
5. UNAIDS. (2012a). Swaziland country report on monitoring the political declaration on HIV and AIDS. Retrieved October 2, 2012, from http://www.unaids.org/en/dataanalysis/knowyourre sponse/countryprogressreports/2012countries/ce_SZ_Narrative_ Report%5B1%5D.pdf. 6. Dabis, F., & Ekpini, E. R. (2002). HIV-1/AIDS and maternal and child health in Africa. The Lancet, 359, 2097–2104. 7. Bassett, M. T. (2001). Keeping the M in MTCT: Women, mothers, and HIV prevention. American Journal of Public Health, 91, 701–703. 8. UNAIDS. (2001a). The impact of voluntary counseling and testing: The global review of the benefits and challenges. Retrieved October 2, 2012, from http://data.unaids.org/Publica tions/IRC-pub02/jc580-vct_en.pdf. 9. Karamagi, C. A. S., Tumwine, J. K., Thorkild, T., & Heggenhougen, K. (2006). Antenatal HIV testing in rural eastern Uganda in 2003: Incomplete rollout of the prevention of motherto-child transmission of HIV programme? Retrieved October 2, 2012 from http://www.biomedcentral.com/1472-698X/6/6. 10. Temmerman, M., Quaghebeur, A., Mwanyumba, F., & Mandaliya, K. (2003). Mother-to-child HIV transmission in resource poor settings: How to improve coverage? Aids, 17, 1239–1242. 11. Hardon, A., Vernooij, E., Bongololo-Mbera, G., Cherutich, P., Desclaux, A., Kyaddondo, D., et al. (2012). Women’s views on consent, counseling and confidentiality in PMTCT: a mixed-methods study in four African countries. BMC Public Health, 2012, 26. 12. Fanta, W., & Worku, A. (2012). Determinants for refusal of HIV testing among women attending for antenatal care in Gambella Region, Ethiopia. Retrieved October 2, 2012 from http://www. reproductive-health-journal.com/content/9/1/8. 13. Pool, R., Nyanzi, S., & Whitworth, J. G. (2001). Attitudes to voluntary counselling and testing for HIV among pregnant women in rural south-west Uganda. AIDS Care, 13, 605–615. 14. Holmes, C., Preko, P., Bolds, R., Baidoo, J., & Jolly, P. (2008). Acceptance of voluntary counselling, testing and treatment for HIV among Pregnant Women in Kumasi Ghana. Ghana Medical Journal, 42, 8–15. 15. Bassett, M. T. (2002). Ensuring a public health impact of programs to reduce HIV transmission from mothers to infants: The place of voluntary counseling and testing. American Journal of Public Health, 92, 347–351. 16. Eide, M., Myhre, M., Lindbaek, M., Sundby, J., Arimi, P., & Thior, I. (2006). Social consequences of HIV-positive women’s participation in prevention of mother-to-child transmission programmes. Patient Education and Counseling, 60, 146–151. 17. Kilewo, C., Massawe, A., Lyamuya, E., Semali, I., Kalokola, F., Urassa, E., et al. (2001). HIV counseling and testing of pregnant women in sub-Saharan Africa—Experiences from a study on prevention of mother-to-child HIV-1 transmission in Dar es Salaam, Tanzania. Journal of Acquired Immune Deficiency Syndromes, 28, 458–462. 18. Rutenberg, N., Baek, C., Kalibala, S., & Rosen, J. (2003). Evaluation of United Nations-supported pilot projects for the prevention of mother-to-child transmission of HIV: Overview of findings. Retrieved October 2, 2012, from http://www.unicef.org/ evaldatabase/files/Global_2003_UN_Supported_PMTCT_Projects. pdf. 19. Baiden, F., Remes, P., Baiden, R., Williams, J., Hodgson, A., Boelaert, M., et al. (2005). Voluntary counseling and HIV testing for pregnant women in the Kassena–Nankana district of northern Ghana: Is couple counseling the way forward? AIDS Care, 17, 648–657. 20. Fylkesnes, K., Haworth, A., Rosensvard, C., & Kwapa, P. M. (1999). HIV counseling and testing: Over- emphasizing high acceptance rates as a threat to confidentiality and the right not to know. Aids, 13, 2469–2474.
123
Matern Child Health J 21. Fabiani, M., Cawthorne, A., Nattabi, B., Ayella, E. O., Ogwang, M., & Declich, S. (2007). Investigating factors associated with uptake of HIV voluntary counselling and testing among pregnant women living in North Uganda. Aids Care-Psychological and Socio-Medical Aspects of Aids/Hiv, 19, 733–739. 22. de Paoli, M. M., Manongi, R., & Klepp, K. I. (2004). Factors influencing acceptability of voluntary counselling and HIV-testing among pregnant women in Northern Tanzania. Aids CarePsychological and Socio-Medical Aspects of Aids/Hiv, 16, 411–425. 23. Painter, T. M. (2001). Voluntary counseling and testing for couples: A high-leverage intervention for HIV/AIDS prevention in sub-Saharan Africa. Social Science and Medicine, 53, 1397–1411. 24. Turan, J. M., Bukusi, E. A., Onono, M., Holzemer, W. L., Miller, S., & Cohen, C. R. (2011). HIV/AIDS stigma and refusal of HIV testing among pregnant women in rural Kenya: Results from the MAMAS Study. AIDS Behavior, 15, 1111–1120. 25. Kroeger, A. (1983). Anthropological and socio-medical health care research in developing countries. Social Science and Medicine, 17, 147–161. 26. Andersen, R. M. (1995). Revisiting the behavioral model and access to medical care: Does it matter? Journal of Health and Social Behavior, 36, 1–10. 27. Central Statistical Office. (CSO) [Swaziland] (2008). Macro International Inc. Swaziland Demographic and Health Survey 2006–2007. In: Mbabane, Swaziland. Calverton, MD: Central Statistical Office and Macro International Inc. 28. Donabedian, A. (1980). Exploration in quality assessment and monitoring: The definition of quality and approaches to its
123
29. 30.
31.
32.
33. 34. 35.
36.
management (vol. 1). Ann Arbor, MA: Health Administration Press. Donabedian, A. (1986). Criteria and standards for quality assessment and monitoring. Quality Review Bulletin, 3, 99–108. Rosenstock, I. M., Strecher, V. J., & Becker, M. (1994). The health belief model and HIV risk behavior change. In R. J. DiClemente, & J. L. Peterson E. (Eds.), Preventing AIDS: Theories and methods of behavioral interventions (pp. 5–24). New York: Plenum Press. Strecher, V. J., & Rosenstock, I. M. (1997). The health belief model. In K. L. F. Glanz, & B. K. Rimer (Eds.), Health behavior and health education: Theory, research and practice (pp. 41–59). San Francisco: Jossey-Bass. Stephenson, R., Elfstrom, K. M., & Winter, A. (2013). Community influences on married men’s uptake of HIV testing in eight African Countries. AIDS Behavior, 17, 2352–2366. Goldstein, H. (2003). Multilevel Statistical Models. London: Edward Arnold. StataCorp. (2011). Stata statistical software: Release 12. College Station, TX: StataCorp LP. UNAIDS. (2012b). Together we will end AIDS. Retrieved December 21, 2012, from http://www.unaids.org/en/media/ unaids/contentassets/documents/epidemiology/2012/jc2296_unaids_ togetherreport_2012_en.pdf. Homsy, J., King, R., Malamba, S. S., Opio, C., Okallany, A., Kalamya, J. N., et al. (2007). The need for partner consent is a main reason for opting out of routine HIV testing for prevention of mother-to-child transmission in a rural Ugandan Hospital. Journal of Acquired Immune Deficiency Syndromes, 44, 366–369.
HIV Counseling and Testing for the Prevention of Mother-toChild Transmission of HIV in Swaziland: A Multilevel Analysis Marguerite L. Sagna • Donald Schopflocher
Ó Springer Science+Business Media New York 2014
Abstract HIV counseling and voluntary testing during antenatal care have been proven to reduce the risk of HIV transmission from mother to child, through increasing knowledge about safe behaviors, ascertaining HIV status and increasing coverage of effective antiretroviral regimens. However, it remains that, in developing countries where 95 % of mother-to-child HIV transmissions (MTCT) take place, such interventions are not widely accessible or available. Using a nationally representative cross-sectional household survey, the present study aimed to examine individual- and contextual-level influences on the receipt of HIV pre-test counseling and uptake of HIV testing during the antenatal care period in Swaziland, a country highly burdened by HIV/ AIDS. The study sample was restricted to women aged 15–49 years with a live birth in the past five years preceding the survey and who received antenatal care for the most recent birth. The findings of this study indicated that only 62 % of women received pre-test counseling for the prevention of MTCT and no more than 56 % of women consented to be tested for HIV during antenatal care. The multilevel regression analysis revealed that the likelihood of receiving HIV pre-test counseling increases significantly with higher parity, education level, household wealth and antenatal visits while it is lower in areas where poverty is pervasive (OR = 0.474) and in rural regions (OR = 0.598) as well. Beyond all the significant predictors, undergoing pre-test counseling has M. L. Sagna (&) D. Schopflocher Centre for Health Promotion Studies, School of Public Health, University of Alberta, 3-270 Edmonton Clinic Health Academy, 11405-87 Avenue, Edmonton, AB T6G 1C9, Canada e-mail: [email protected] D. Schopflocher Faculty of Nursing, University of Alberta, Edmonton, AB, Canada
emerged as an important determinant of HIV testing. Receiving pre-test counseling increases the odds of accepting an HIV test by 77 %. Evidence from this analysis underscores bottlenecks and challenges that persist in increasing the need for and uptake of HIV preventive and treatment services to stop new HIV infections among children. Keywords HIV pre-test counseling HIV testing Mother-to-child transmission Multilevel modeling Swaziland
Introduction Substantial gains in child survival rates are being increasingly erased in many countries in sub-Saharan Africa because of severe pediatric HIV/AIDS. The magnitude of childhood HIV infection is indisputably alarming in this part of the world, which accounts for 91 % of the global 3.4 million HIV positive children under age 15 [1]. Most of these children became HIV positive by vertical transmission or mother to-child-transmission (MTCT), that is, they acquired the virus from their HIV infected mothers during pregnancy, birth or breastfeeding. In recent years in efforts to drive the rates of new pediatric HIV infections down, many sub-Saharan African countries have adopted a model of service delivery, which integrates HIV counseling and testing as part of the antenatal care routine. However, coverage of such interventions is still inadequate across the region, where available estimates show that about 42 % women were tested for HIV in 2010 (ranging from 25 % in western and central Africa to 61 % in eastern and southern Africa [2]). Swaziland is no different. Despite a strong prevention of mother-to-child-transmission (PMTCT) program launched in 2003, the vertical transmission rate is still disconcerting with
123
Matern Child Health J
11 % of infants and children being infected by their HIVpositive mothers in 2011 [3]. In addition, the country is in the midst of the most severe generalized HIV epidemic in the world, with 25.9 % of its adult population aged 15–49 estimated to be HIV positive [4]. Current estimates also indicate that the rate of infection among women of reproductive age continues to be as high as 3.13 % nationally [2], making the proportion of pregnant women living with HIV/AIDS—about 41 %—the largest worldwide. The national strategy to optimize the impact of PMTCT services in Swaziland consisted of a shift from client-initiated to provider-initiated HIV testing and counseling in 2006 [5]. Under the provider-initiated approach, each pregnant woman should be provided with pretest HIV counseling and subsequently offered an HIV test that would be executed unless the woman declines. Yet, all too often national efforts to expand HIV counseling and testing to a greater number of pregnant women during antenatal care are impeded by women’s reluctance to be tested [6–8], thus missing the opportunity to know their sero-status, and to make informed decision about their health and the well-being of their newborns. Disparities in receiving pre-test counseling also occur, even in settings where this intervention has been structurally integrated in the antenatal care package [9, 10]. While factors associated with sporadic pre-test HIV counseling have received little attention [11], acceptability barriers for HIV testing among pregnant women have been well studied [9, 12–24]. Most published research has drawn on data from healthcare facilities (i.e. largely urban antenatal clinics), sentinel surveillance, and clinical trials. Very few attempts have been made to investigate community-level influences. This limitation is important since health care seeking behaviors are not purely dependent on individuals’ choice or circumstances but also determined by the structure of the social environment in which individuals live [25, 26]. In this paper, we use a population-based survey to examine differentials in the receipt of HIV pre-test counseling and uptake of HIV testing among pregnant women in Swaziland, accounting for the influence of both individual-level and contextual-level factors. With the number of HIV infected women of childbearing age continuously increasing, identifying factors associated with missed opportunities for PMTCT is of paramount importance to strengthening HIV prevention interventions and reducing the risk of MTCT and HIV infection in the population at large.
Methods Data The present study is based on a secondary analysis of cross sectional data derived from the Demographic and
123
Health Survey (DHS) conducted in Swaziland between 2006 and 2007. The DHS is a nationally representative household survey designed to provide information on a battery of issues including fertility, reproductive health, sexual behaviors, and HIV/AIDS related knowledge, attitudes and behaviors in resource-poor settings. The 2006/2007 Swaziland Demographic and Health Survey (SDHS) is the first of its kind to be conducted in this country. Of 5,301 eligible women age 15–49 identified for the survey in 5,500 selected households randomly sampled within 275 primary sampling units (PSUs), 4,987 were successfully interviewed, reflecting a response rate of 94 %. Detailed information about the design and implementation of the SDHS is described in the survey report [27]. The present study focuses on a restricted sample of 1,788 women age 15–49 with a live birth in the past 5 years preceding the survey who received antenatal care for the most recent birth. The sample is comprised of 274 PSUs, representing an average number of 6.5 women per PSU. Since this study involved analysis of secondary data, publicly available upon request, IRB approval was not required. Variables The outcome variables of interest were receipt of HIV pretest counseling during antenatal care, and uptake of HIV testing during antenatal care. Receipt of HIV pre-test counseling is coded 1 if the respondent has received essential information about vertical transmission of HIV during prenatal care and 0 otherwise. Uptake of HIV testing is coded 1 if the respondent consented for an HIV test and 0 otherwise. Choosing explanatory variables for predicting HIV pretest counseling, and HIV testing was guided by the Donabedian’s Quality of Care Framework [28, 29] and the Health Belief Model [30, 31], respectively. The Donabedian Framework suggests that quality of care is attributable to three dimensions: structure (i.e. the context in which care is delivered), process (i.e. interactions between patients and providers throughout the delivery of healthcare) and outcome (i.e. the impact of care on patients’ health status). The context in which care is delivered affects processes and outcomes while the outcomes indicate the combined effects of structure and process. The Health Belief Model argues that the likelihood for an individual to adopt a preventive behavior is affected by personal perceptions about a disease and the subjective weighing of the benefits of adopting the behavior. Based on these conceptual frameworks and informed by evidence from published literature, we included these individual-level covariates: age, marital status, parity, education level, household wealth, timing of first antenatal
Matern Child Health J
visit, number of antenatal visits, type of antenatal care medical provider, diagnosis of a sexually transmissible disease (STD) in the last 12 months, knowledge of MTCT of HIV, and exposure to media. Age is grouped into the following categories: 15–19, 20–29, 30–39, and 40–49 years of age. Marital status is classified according to whether the respondent was single, married or separated/ widowed. Parity is grouped into three categories including 1 child, 2 children, and 3 children or more. Education level refers to the highest level of education completed by the respondent and is grouped into four categories including no education, primary level, secondary level and tertiary level. Household wealth is assessed by five categories corresponding to the index wealth quintiles provided in the survey, including poorest, poor, middle, rich, and richest. Timing of first antenatal care is coded as a dichotomous variable indicating whether or not a woman had received her first antenatal visit within the first trimester. The number of antenatal care visits is recorded as three-category variable indicating whether the respondent has made 3 visits of less, 4 visits, and 5 visits or more. The type of medical personnel who provided antenatal care is assessed by three separate binary variables indicating whether the respondent received care from a doctor, midwife or auxiliary midwife. Diagnosis of STD was measured by binary variable expressing whether or not the respondent has been diagnosed with STD in the previous year. Knowledge of MTCT of HIV is analyzed by three separate dichotomous binary variables indicating whether or not the respondent knows that AIDS can be transmitted during pregnancy, childbirth or through breastfeeding. Exposure to media is assessed across three separate dichotomous variables on whether the respondent read newspapers, listened to radio or watched television at least once a week. The community-level covariates included integral variables and aggregate variables. The integral variables were namely community type (i.e. urban/rural) and region as collected by the survey. The primary sampling unit (PSU) was considered the community-level of analysis in the present study. Aggregate variables were derived from individual level responses by aggregating the individual characteristics of respondents within the to the PSU and then assigning the aggregate value back to each individual in the PSU as a separate variable. Given the scarcity of studies addressing community-level determinants of missed opportunities for PMTCT, a number of communitylevel factors were selected based on Stephenson et al. [32] research exploring community influences on HIV testing uptake among married men in Africa. For this analysis, we included variables that capture community economics, community behavior, and knowledge and community features. The community economics variables were percent-
age of women in community with secondary or higher education, and community poverty conceptualized as the percentage of women in community living in households that fall in the two lowest wealth index quintiles. The community behavior and knowledge variables were the percentage of women in community with knowledge of MTCT, the percentage of women in community against domestic violence, and the percentage of women in community who would want to keep secret the HIV/AIDS status of a family member. The community features variables were community type and province. Statistical analysis Our data analysis was conducted in two phases. First, we performed some descriptive analysis to examine the characteristics of the sample. Second, a multilevel regression modeling strategy was used to examine the influence of individual and community characteristics on interventions related to PMTC. This strategy accounts for the hierarchical nature of the data and corrects the estimated standards errors to allow for the clustering of observations within units [33]. We specified two-level multilevel logistic regression models separately to test predictors of HIV pretest counseling and uptake of HIV testing. The logit of the probability of each outcome is modeled as follows: Logit ðpij Þ ¼ logðpij =ð1 pij ÞÞ ¼ b0 þ b1 Iij þ b2 Cj þ lj where pij is the probability that the outcome is 1 for a respondent i in community j; b0 is the intercept; the b’s are the fixed coefficients; I and C represent individual and community variables respectively; lj * N(0, r2j ) represents the random effects for the jth community (274 community-specific random effects).The analytical strategy involves applying two models for each outcome. Model 1 includes only the individual-level variables as predictors while model 2 adds the community-level variables, to examine the influences of each set of variables and, to test whether specific components of the community independently influence HIV counseling and testing after adjusting for individual characteristics. Due to the complex design of the survey, all analyses are adjusted for the multistage sampling design and the sample weights are used to allow for adjustment of how the sample was collected. All statistical analyses are estimated in Stata version 12 [34].
Results Table 1 presents the characteristics of women of reproductive age (15–49) in the study sample. HIV counseling is
123
Matern Child Health J Table 1 Sample characteristics of women aged 15–49 with adjustment for survey design, 2006/2007 Swaziland Demographic and Health Survey (n = 1,788)
Table 1 continued
Variables
SE
Listens to radio at least once a week Watches television at least once a week Yes
Mean
Variables
Yes
Individual-level covariates
Mean
SE
78.6
0.011
38.2
0.018
HIV pre-test counseling
62.4
0.014
HIV testing Current age (in years)
56.1
0.014
\20
19.5
0.010
Percentage of women in community with secondary or higher education
57.1
0.012
20–29
52.8
0.013
Community poverty
36.1
0.018
30–39
24.0
0.011
63.7
0.010
40–49
3.7
0.005
Percentage of women in community with knowledge of mother-to-child transmission of HIV
0.009
39.7
0.014
Percentage of women in community against domestic violence
66.1
Single Married
55.0
0.015
26.6
0.015
5.3
0.005
Percentage of women in community who would want to keep secret the HIV/AIDS status of a family member
1 child
32.1
0.011
Urban
23.3
0.023
2 children
24.3
0.011
76.7
0.023
3 children or more
43.6
0.012
Rural Province Hhohho
26.6
0.024
No education Primary
8.0 33.0
0.008 0.013
Manzini
32.4
0.028
Shiselweni
21.6
0.012
Secondary
52.7
0.015
Lubombo
19.4
0.014
6.3
0.008
Marital status
Separated/widow Parity
Community-level covariates (n = 274 PSUs)
Community type
Education level
Tertiary Household wealth index Poorest
18.9
0.016
Poor
19.4
0.013
Middle
20.5
0.015
Rich
20.3
0.014
Richest
20.9
0.017
25.9
0.011
Timing of first antenatal visit First trimester Number of antenatal visits 3 visits or less
15.9
0.009
4 visits
18.9
0.009
5 visits or more
65.2
0.012
3.1
0.005
Midwife
80.9
0.015
Auxiliary midwife
16.0
0.014
6.4
0.006
77.4
0.011
91.6
0.007
Type of antenatal care provider Doctor
Diagnosis of STD in the past year Yes Knows that AIDS can be transmitted during pregnancy Yes Knows that AIDS can be transmitted during delivery Yes
Knows that AIDS can be transmitted through breastfeeding Yes
87.6
0.009
58.3
0.016
Reads newspaper at least once a week Yes
123
not a universal practice, as only 62 % of women have received information about MTCT of HIV. The uptake of HIV testing is similarly low, with about 56 % of women who accepted to be screened for HIV during antennal care. Table 2 presents results of the multilevel logistic regressions for receiving pre-test HIV counseling during prenatal care. At the individual-level (Model 1), the odds of receiving HIV pre-test counseling are significantly associated with parity, education level, household wealth index, and the number of antenatal visits. Women of parity 3 or higher (OR 1.562) are significantly more likely to receive HIV pre-test counseling than women of parity 1. The odds of receiving a pre-test counseling increase significantly with the level of education; women with a primary (OR 1.542), secondary (OR 2.317) or higher education (OR 2.538) have significantly greater odds of are receiving HIV pre-test counseling than women with no education. Living in households that fall in the poor (OR 1.483), rich (OR 1.552) and richest (OR 1.500) wealth quintiles is associated with higher odds of getting HIV pre-test counseling compared to living in households that fall in the poorest wealth quintile. The odds of receiving HIV pre-test counseling increase by about 61 % for women who attended four antenatal visits and 63 % for women attended five or more prenatal visits.
Matern Child Health J Table 2 Results of the logistic analysis of the predictors of the receipt of HIV counseling during prenatal care, 2006/07 SDHS (n = 1,788) Model 1
Model 2
Odds ratio
95 % CI
Odds ratio
95 % CI
20–29
1.117
0.784, 1.591
1.027
0.721, 1.464
30–39
1.108
0.701, 1.752
1.003
0.634, 1.587
40–49
1.041
0.568, 1.905
0.970
0.529, 1.779
0.859
0.671, 1.100
0.921
0.718, 1.181
0.811
0.486, 1.351
0.782
0.467, 1.307
2 children
1.203
0.891, 1.625
1.230
0.912, 1.661
3 children or more
1.562**
1.110, 2.198
1.606**
1.140, 2.262
Primary
1.542*
1.037, 2.293
1.453
0.978, 2.159
Secondary
2.317***
1.551, 3.462
2.183***
1.453, 3.280
Tertiary
2.538**
1.368, 4.709
2.386**
1.274, 4.468
Poor
1.483*
1.061, 2.072
1.316
0.942, 1.838
Middle
1.287
0.914, 1.812
0.974
0.681, 1.393
Rich
1.552*
1.085, 2.218
0.968
0.666, 1.461
Richest
1.500*
1.029, 2.187
0.778
0.491, 1.232
0.917
0.706, 1.191
0.896
0.690, 1.164
1.610**
1.139, 2.276
1.642**
1.163, 2.318
1.631***
1.207, 2.204
1.669***
1.236, 2.253
Midwife
0.942
0.594, 1.494
0.990
0.625, 1.567
Auxiliary midwife
0.733
0.360, 1.490
1.344
0.842, 2.144
Percentage of women in community with secondary or higher education
0.885
0.391, 2.002
Community poverty
0.473*
0.233, 0.961
0.598**
0.415, 0.863
Manzini
1.619**
1.176, 2.228
Shiselweni
1.243
0.889, 1.738
Lubombo
0.761
0.540, 1.071
Current age (ref = less than 20)
Marital status (ref = single) Married Separated/widow Parity (ref = 1 child)
Education level (ref = no education)
Household wealth index (ref = poorest)
Timing of first antenatal visit (ref = subsequent trimesters) First trimester Number of antenatal visits (ref = 3 visits or less) 4 visits 5 visits or more Type of antenatal care provider (ref = doctor)
Community-level covariates a
Community type (ref = urban) Rural Province (ref = Hhohho)
*** p \ 0.001; ** p \ 0.01; * p \ 0.05;
p \ 0.10
After adjusting for individual-level factors (model 2), we found that community-level covariates such as community poverty, the percentage of women in community with knowledge of MTCT, community type and, province of residence have an influence on the receipt of HIV pretest counseling. The likelihood of undergoing pre-test counseling is significantly lower in areas where a large proportion of women live in poor households. The odds of receiving HIV pre-test counseling decrease by about 40 %
for women living in rural areas compared to women residing in urban areas. Further, living in Manzini (OR 1.619) increases the odds of receiving HIV pre-test counseling than living in Hhohho. Table 3 presents results of multilevel logistic regressions of factors associated with the uptake of HIV testing during antenatal care. Results of Model 1 show that the odds of consenting for an HIV test are significantly lower for women of parity 2 (OR 0.625) and women of parity 3 or
123
Matern Child Health J Table 3 Results of the logistic analysis of the predictors of consenting for HIV testing during prenatal care, 2006/07 SDHS (n = 1,640) Model 1
Model 2
Odds ratio
95 % CI
Odds ratio
95 % CI
20–29
1.178
0.815, 1.702
1.145
0.793, 1.652
30–39
1.137
0.703, 1.837
1.113
0.689, 1.794
40–49
1.567
0.728, 3.370
1.512
0.704, 3.244
0.812
0.614, 1.071
0.832
0.629, 1.099
0.795
0.449, 1.405
0.789
0.445, 1.395
2 children
0.625**
0.442, 0.883
0.623**
0.441, 0.878
3 children or more
0.648*
0.441, 0.953
0.641*
0.436, 0.941
Primary
1.475
0.926, 2.346
1.454
0.915, 2.307
Secondary
1.649*
1.016, 2.674
1.634*
1.002, 2.663
1.915
0.939, 3.902
2.033
0.990, 4.171
Poor
1.104
0.754, 1.613
1.083
0.741, 1.581
Middle
1.155
0.776, 1.717
1.135
0.748, 1.722
Rich
1.516
0.991, 2.317
1.369
0.858, 2.182
Richest
1.024
0.641, 1.634
0.855
0.488, 1.494
1.046
0.781, 1.399
1.047
0.782, 1.401
0.597*
0.401, 0.886
0.615*
0.414, 0.911
0.828
0.579, 1.182
0.839
0.588, 1.195
Midwife
0.488
0.234, 1.016
0.490
0.235, 1.021
Auxiliary midwife
0.382*
0.175, 0.828
0.403*
0.186, 0.874
1.772***
1.374, 2.286
1.667***
1.289, 2.153
0.623*
0.389, 0.997
0.605*
0.378, 0.9679
0.933
0.700, 1.241
0.886
0.663, 1.183
1.285
0.818, 2.016
1.285
0.820, 2.013
1.114
0.773, 1.604
1.105
0.768, 1.588
0.786
0.594, 1.040
0.789
0.596, 1.042
1.067
0.795, 1.431
1.108
0.826, 1.483
0.911
0.670, 1.236
0.948
0.698, 1.286
Percentage of women in community with secondary or higher education
0.793
0.322, 1.949
Community poverty
1.003
0.457, 2.195
Current age (ref = less than 20)
Marital status (ref = single) Married Separated/widow Parity (ref = 1 child)
Education level (ref = no education)
Tertiary
Household wealth index (ref = poorest)
Timing of first antenatal visit (ref = subsequent trimesters) First trimester Number of antenatal visits (ref = 3 visits or less) 4 visits 5 visits or more Type of antenatal care provider (ref = doctor)
HIV pre-test counseling (ref = no) Yes Diagnosis of STD in the past year (ref = no) Yes Knows that AIDS can be transmitted during pregnancy (ref = no) Yes Knows that AIDS can be transmitted during delivery (ref = no) Yes Knows that AIDS can be transmitted through breastfeeding (ref = no) Yes Reads newspaper at least once a week (ref = no) Yes Listens to radio at least once a week (ref = no) Yes Watches television at least once a week (ref = no) Yes Community-level covariates
123
Matern Child Health J Table 3 continued Model 1 Odds ratio
Model 2 95 % CI
Odds ratio
95 % CI
Percentage of women in community with knowledge of mother-to-child transmission of HIV
1.395
0.552, 3.521
Percentage of women in community against domestic violence
0.688
0.253, 1.870
Percentage of women in community who would want to keep secret the HIV/AIDS status of a family member
0.457
0.177, 1.179
0.777
0.515, 1.171
2.004***
1.426, 2.814
Shiselweni
1.334
0.933, 1.907
Lubombo
1.550*
1.047, 2.295
Community type (ref = urban) Rural Province (ref = Hhohho) Manzini
*** p \ 0.001; ** p \ 0.01; * p \ 0.05;
p \ 0.10
more (OR 0.648) than women of parity 1. Women with a secondary level of education have 65 % higher odds of being screened for HIV than women with no education, everything else being equal. Women with four antenatal visits have 40 % lower odds of accepting HIV testing than women with three or less prenatal visits. Women who received prenatal care from an auxiliary midwife (OR 0.382) were less likely to consent for an HIV test than women who were seen by a doctor. The odds of accepting HIV testing are 77 % greater for women who received a pre-test counseling than women who did not get any counseling. A diagnosis of STD in the previous year (OR 0.623) is negatively associated with the likelihood of consenting for an HIV test during prenatal care. Community level influences are presented in Model 2. Net of individual factors, province is significantly associated with the likelihood of consenting for an HIV test. Women in the Manzini province (OR 2.004) and the Lubombo province (OR 1.550) are more likely to accept HIV testing than women living in the Hhohhio province.
Discussion The high levels of pediatric HIV infection in many subSaharan African countries has led to both the implementation of, and a surge of research that examines PMTCT among pregnant women and its uptake. We believe that the present analysis constitutes a significant improvement on previous studies for several reasons. First, individual- and community-level factors associated with receipt of HIV pre-test counseling and HIV testing during antenatal care are explored using multilevel modeling to provide a contextual analysis of issues surrounding PMTCT. In contrast with the majority of previous studies, the data used here
were drawn from a nationally representative population survey designed to collect health information including reproductive health and HIV/AIDS. This research also contrasts with prior studies, because it examines pre-test counseling and HIV testing individually and emphasizes their shared importance in PMTCT. Finally, results presented here have clinical and public health implications and are relevant for informing HIV prevention planning efforts. In Swaziland, although virtually all women (97 %) attended antenatal care from skilled providers in the 5 years preceding the survey [27], no more than 62 % of women received pre-test counseling for the prevention of MTCT and only 56 % of women consented to be tested for HIV during antenatal care. These low rates have implications for the success of PMTCT programs given that the brunt of pre-test counseling trickle down to HIV testing, a gateway for maternal HIV infection screening, modification of infant feeding, and use of antiretroviral medications. Disparities in receiving pre-test counseling are influenced by a range of individual and community factors. At the individual-level, one of the most significant findings is the positive association between education and HIV pre-test counseling. Consistent with previous research [11], we found that women with higher level of education are more likely to undergo pre-test counseling. Since educated women are more likely to assimilate health information, health providers may be more motivated to discuss PMTCT with them. Odds of receiving counseling about HIV during prenatal care are also greater among women of parity 3 or more compared to women of parity 1. This is likely linked to differentials in HIV prevalence between pregnant women in younger age groups and their peers in older age groups. Women of higher parity tend to be older (aged 30–39 and 40–49), and available estimates have shown that as the HIV/AIDS epidemic has progressed in
123
Matern Child Health J
Swaziland, the prevalence of the disease among pregnant women has increased among older pregnant women while it has decreased among the younger age group (15–19 years) [5]. The pervasiveness of the HIV infection among older women may explain the determination of health care providers to offer counseling services to higher parity women. Household wealth is also predictive of HIV pre-test counseling, as high household income level is associated with a greater likelihood of receiving of HIV pre-test counseling. Women living in household with higher economic standing may experience greater financial advantages that allow them to easily navigate the health care system and therefore to seek care in reputable medical facilities or private clinics usually reputed to provide better care than government-operated facilities. The finding that the greater odds of receiving counseling are associated with higher frequency of visits suggests that compliance to the World Health Organization’s recommended number of prenatal checkups (i.e. at least four visits) is protective against missed opportunities for HIV pre-test counseling, as consistent use of antenatal care increasingly exposes a woman to information about PMTCT. The results of this study also demonstrate that pre-test counseling is influenced by contextual factors. The higher the concentration of poverty in a community, the lower the odds of receiving HIV counseling during prenatal care, likely because impoverished neighborhoods tend to harbor medical facilities that are poorly staffed. Since health providers are burdened with greater workload there will be less time to provide one-to-one counseling. Moreover, women in rural areas are less likely to be counseled about HIV transmission and prevention during prenatal care. A gap in the quality of care may also account for this difference. Counseling requires time and interaction with patients, and for many health providers in rural areas addressing the need for individual counseling is a difficult challenge. The reality is that majority of medical facilities in rural areas suffer from lack of resources, lack of trained staff to shortage of personnel, putting medical providers to a certain level of stress that impede them from providing adequate or comprehensive care to their patients. Pre-test counseling is also influenced by the province of residence; likely attributable to differences in content and quality of care dispensed at health facilities across the country. The present study also identified barriers and facilitators to HIV testing among pregnant women in Swaziland at multiple levels. At the individual-level, a number of factors are found to be associated with the uptake of HIV testing. Women with high parity are more likely to refuse HIV testing, likely because they do not perceive themselves at a high risk, especially if they have been tested negative during their previous pregnancies. Consistent with prior work [12, 21], uptake of HIV testing is greater among
123
educated women, likely due to the fact they are better informed and understand the benefits of HIV prevention interventions and are thus more amenable to informed choices about their health and health care. Though increasing numbers of antenatal checkups up to four visits was positively associated with pre-test counseling in the previous analysis, it was negatively associated with the uptake of HIV testing. This suggests that considerable efforts to delineate the importance of HIV testing need to be deployed during the very first antenatal checkup to minimize missed opportunities for testing. Women who received antenatal care from an auxiliary midwife/nurse are also more likely to decline HIV testing. This may reflect a lack of trust patients may have in this category of health providers who may be considered less qualified. It may also reflect the quality of information on PMTC provided by an auxiliary midwife/nurse. The association between pre-test counseling and HIV testing is noteworthy. Women who received pre-test counseling are significantly more likely to consent to an HIV test. This underscores the importance of pre-test counseling as a springboard for HIV testing, and is consistent with work from several countries including Zambia, Burundi, Kenya, Rwanda, and Uganda [18]. Pre-test counseling goes beyond providing information about the importance of HIV testing, to raising awareness about risk reduction strategies, family planning options, and connecting patients to appropriate treatments as needed. Further, in agreement with a previous study [14], a history of a prior STD is significantly associated with refusal of HIV testing. Reasons for this may have to do with patients’ perception of already being HIV infected, the fact that they have been previously tested as part of the medical care following STD diagnosis, or simply because they may be in denial about their risk for HIV. At the community-level after adjusting for individual characteristics, province of residence was associated with the uptake of testing. Women living in Manzini and Lubombo are more likely to consent for HIV testing than their Hhohho counterparts. This is noteworthy considering that Hhohho is the most urbanized region and has the highest nationwide HIV prevalence with 21.0 % of the population 2 years or older being HIV positive, while Manzini had 19.1 % and Lubombo had 18.1 % [27]. No other community-level variables were found to influence the uptake of HIV testing during antenatal care. This contrasts with a previous study examining contextual influences of HIV testing among men in several sub-Saharan African countries [32] where community economics, community knowledge of HIV, and community tolerance of violence towards women were associated with HIV testing. There were number of limitations in this study. First, this analysis uses PSUs as the higher-level unit of analysis.
Matern Child Health J
This may be problematic since the PSUs are created for enumeration purpose and do not necessarily correspond to actual communities. Second, the present research is based on data collected in 2006/07, and it is likely much has changed in Swaziland since then. In fact, a recent report has shown that the percentage of child HIV infections from HIV positive women has decreased from 18 % in 2009 to 11 % in 2011 [35], mirroring changes in preventive careseeking behaviors as well as improvement of quality of PMTCT interventions and access to services. As well, our study was unable to examine the effects of factors such as fear of repercussions (e.g. discrimination, divorce or domestic violence) [13, 24], male support/consent [36], and issues related to privacy and confidentiality in health care [20]. Finally, the survey was cross-sectional and therefore causal inferences are difficult to draw and impossible to confirm. Despite these drawbacks, this study does critically examine factors influencing disparities in receiving pre-test counseling and uptake of HIV testing among pregnant women at multiple levels. This is relevant to HIV prevention programs in a developing country context, since pre-test counseling is conducive to HIV testing, and getting tested for HIV has ripple effects. Such testing allows a woman to become aware of her HIV status, to start using antiretroviral medications, to make decisions about breastfeeding options and to obtain early HIV testing for her baby. Disparities in pre-test counseling suggest that better compliance of health professionals in providing adequate prenatal care is worthy striving for. Of course, this will remain a challenge if the shortage of health care workers endemic in most sub-Saharan African countries cannot be addressed. In conclusion, this investigation underscores the need for multilevel public health interventions to further promote HIV counseling and testing in order to prevent HIV transmission in all segments of the population.
References 1. UNAIDS. (2001b). Counselling and voluntary HIV testing for pregnant women in high HIV prevalence countries: Elements and issues. Retrieved October 2, 2012, from data.unaids.org/Publica tions/IRC-pub01/jc245-couns_test_en.pdf. 2. WHO, UNAIDS, UNICEF. (2011). Global HIV/AIDS responseEpidemic update and health sector progress towards universal access. Retrieved October 2, 2012, from http://www.unaids.org/ en/media/unaids/contentassets/documents/unaidspublication/2011/ 20111130_UA_Report_en.pdf. 3. UNICEF. (2012). Countdown to zero: Elimination of new HIV infections among children by 2015 and keeping their mothers alive. Retrieved October 2, 2012, from http://www.unicef.org/ aids/files/hiv_pmtctfactsheetSwaziland.pdf. 4. UNAIDS. (2010). Globla report: UNAIDS report on the global AIDS epidemic 2010. Retrieved October 2, 2012, from http:// www.unaids.org/documents/20101123_globalreport_em.pdf.
5. UNAIDS. (2012a). Swaziland country report on monitoring the political declaration on HIV and AIDS. Retrieved October 2, 2012, from http://www.unaids.org/en/dataanalysis/knowyourre sponse/countryprogressreports/2012countries/ce_SZ_Narrative_ Report%5B1%5D.pdf. 6. Dabis, F., & Ekpini, E. R. (2002). HIV-1/AIDS and maternal and child health in Africa. The Lancet, 359, 2097–2104. 7. Bassett, M. T. (2001). Keeping the M in MTCT: Women, mothers, and HIV prevention. American Journal of Public Health, 91, 701–703. 8. UNAIDS. (2001a). The impact of voluntary counseling and testing: The global review of the benefits and challenges. Retrieved October 2, 2012, from http://data.unaids.org/Publica tions/IRC-pub02/jc580-vct_en.pdf. 9. Karamagi, C. A. S., Tumwine, J. K., Thorkild, T., & Heggenhougen, K. (2006). Antenatal HIV testing in rural eastern Uganda in 2003: Incomplete rollout of the prevention of motherto-child transmission of HIV programme? Retrieved October 2, 2012 from http://www.biomedcentral.com/1472-698X/6/6. 10. Temmerman, M., Quaghebeur, A., Mwanyumba, F., & Mandaliya, K. (2003). Mother-to-child HIV transmission in resource poor settings: How to improve coverage? Aids, 17, 1239–1242. 11. Hardon, A., Vernooij, E., Bongololo-Mbera, G., Cherutich, P., Desclaux, A., Kyaddondo, D., et al. (2012). Women’s views on consent, counseling and confidentiality in PMTCT: a mixed-methods study in four African countries. BMC Public Health, 2012, 26. 12. Fanta, W., & Worku, A. (2012). Determinants for refusal of HIV testing among women attending for antenatal care in Gambella Region, Ethiopia. Retrieved October 2, 2012 from http://www. reproductive-health-journal.com/content/9/1/8. 13. Pool, R., Nyanzi, S., & Whitworth, J. G. (2001). Attitudes to voluntary counselling and testing for HIV among pregnant women in rural south-west Uganda. AIDS Care, 13, 605–615. 14. Holmes, C., Preko, P., Bolds, R., Baidoo, J., & Jolly, P. (2008). Acceptance of voluntary counselling, testing and treatment for HIV among Pregnant Women in Kumasi Ghana. Ghana Medical Journal, 42, 8–15. 15. Bassett, M. T. (2002). Ensuring a public health impact of programs to reduce HIV transmission from mothers to infants: The place of voluntary counseling and testing. American Journal of Public Health, 92, 347–351. 16. Eide, M., Myhre, M., Lindbaek, M., Sundby, J., Arimi, P., & Thior, I. (2006). Social consequences of HIV-positive women’s participation in prevention of mother-to-child transmission programmes. Patient Education and Counseling, 60, 146–151. 17. Kilewo, C., Massawe, A., Lyamuya, E., Semali, I., Kalokola, F., Urassa, E., et al. (2001). HIV counseling and testing of pregnant women in sub-Saharan Africa—Experiences from a study on prevention of mother-to-child HIV-1 transmission in Dar es Salaam, Tanzania. Journal of Acquired Immune Deficiency Syndromes, 28, 458–462. 18. Rutenberg, N., Baek, C., Kalibala, S., & Rosen, J. (2003). Evaluation of United Nations-supported pilot projects for the prevention of mother-to-child transmission of HIV: Overview of findings. Retrieved October 2, 2012, from http://www.unicef.org/ evaldatabase/files/Global_2003_UN_Supported_PMTCT_Projects. pdf. 19. Baiden, F., Remes, P., Baiden, R., Williams, J., Hodgson, A., Boelaert, M., et al. (2005). Voluntary counseling and HIV testing for pregnant women in the Kassena–Nankana district of northern Ghana: Is couple counseling the way forward? AIDS Care, 17, 648–657. 20. Fylkesnes, K., Haworth, A., Rosensvard, C., & Kwapa, P. M. (1999). HIV counseling and testing: Over- emphasizing high acceptance rates as a threat to confidentiality and the right not to know. Aids, 13, 2469–2474.
123
Matern Child Health J 21. Fabiani, M., Cawthorne, A., Nattabi, B., Ayella, E. O., Ogwang, M., & Declich, S. (2007). Investigating factors associated with uptake of HIV voluntary counselling and testing among pregnant women living in North Uganda. Aids Care-Psychological and Socio-Medical Aspects of Aids/Hiv, 19, 733–739. 22. de Paoli, M. M., Manongi, R., & Klepp, K. I. (2004). Factors influencing acceptability of voluntary counselling and HIV-testing among pregnant women in Northern Tanzania. Aids CarePsychological and Socio-Medical Aspects of Aids/Hiv, 16, 411–425. 23. Painter, T. M. (2001). Voluntary counseling and testing for couples: A high-leverage intervention for HIV/AIDS prevention in sub-Saharan Africa. Social Science and Medicine, 53, 1397–1411. 24. Turan, J. M., Bukusi, E. A., Onono, M., Holzemer, W. L., Miller, S., & Cohen, C. R. (2011). HIV/AIDS stigma and refusal of HIV testing among pregnant women in rural Kenya: Results from the MAMAS Study. AIDS Behavior, 15, 1111–1120. 25. Kroeger, A. (1983). Anthropological and socio-medical health care research in developing countries. Social Science and Medicine, 17, 147–161. 26. Andersen, R. M. (1995). Revisiting the behavioral model and access to medical care: Does it matter? Journal of Health and Social Behavior, 36, 1–10. 27. Central Statistical Office. (CSO) [Swaziland] (2008). Macro International Inc. Swaziland Demographic and Health Survey 2006–2007. In: Mbabane, Swaziland. Calverton, MD: Central Statistical Office and Macro International Inc. 28. Donabedian, A. (1980). Exploration in quality assessment and monitoring: The definition of quality and approaches to its
123
29. 30.
31.
32.
33. 34. 35.
36.
management (vol. 1). Ann Arbor, MA: Health Administration Press. Donabedian, A. (1986). Criteria and standards for quality assessment and monitoring. Quality Review Bulletin, 3, 99–108. Rosenstock, I. M., Strecher, V. J., & Becker, M. (1994). The health belief model and HIV risk behavior change. In R. J. DiClemente, & J. L. Peterson E. (Eds.), Preventing AIDS: Theories and methods of behavioral interventions (pp. 5–24). New York: Plenum Press. Strecher, V. J., & Rosenstock, I. M. (1997). The health belief model. In K. L. F. Glanz, & B. K. Rimer (Eds.), Health behavior and health education: Theory, research and practice (pp. 41–59). San Francisco: Jossey-Bass. Stephenson, R., Elfstrom, K. M., & Winter, A. (2013). Community influences on married men’s uptake of HIV testing in eight African Countries. AIDS Behavior, 17, 2352–2366. Goldstein, H. (2003). Multilevel Statistical Models. London: Edward Arnold. StataCorp. (2011). Stata statistical software: Release 12. College Station, TX: StataCorp LP. UNAIDS. (2012b). Together we will end AIDS. Retrieved December 21, 2012, from http://www.unaids.org/en/media/ unaids/contentassets/documents/epidemiology/2012/jc2296_unaids_ togetherreport_2012_en.pdf. Homsy, J., King, R., Malamba, S. S., Opio, C., Okallany, A., Kalamya, J. N., et al. (2007). The need for partner consent is a main reason for opting out of routine HIV testing for prevention of mother-to-child transmission in a rural Ugandan Hospital. Journal of Acquired Immune Deficiency Syndromes, 44, 366–369.
