Utilization of delivery services in the context of Prevention of HIV from Mother-To-Child ( PMTCT ) in a rural community , South Africa

Curationis 29(1): 54-61 The aim of this study was to investigate the utilization of delivery services in the context of PMTCT in a rural community in South Africa. Based on a cross-sectional survey, the sample included 870 pregnant women who had delivered before recruited from five PMTCT clinics and surrounding communities. Results indicated that 55.9% had delivered their last child in a health care facility and 44.1% at home (mostly without assistance from a traditional birth attendant). The odds of access to the health facility were (1) women who stayed close to the hospital (OR=2.87), (2) those who had higher formal education (OR=l .55), (3) higher traveling costs (affordability) to get to nearest clinic (OR=1.77), and (4) those who were single (OR=1.58). Childbirth experiences of the mother or mother-in-law greatly influenced the delivery choices in terms of home delivery. The majority of the pregnant women were aware of mother-to-child HIV transmission but only 9% of the pregnant women had ever been tested for HIV. HIV knowledge, HIV testing behaviour and attitudes were found to be not associated with the delivery option. T Mosala Human Sciences Research Council


Introduction
Lack of access to maternal health services resulted in poor utilization of health services by rural communities.The rate of home deliveries ranged from 40% to 92% in rural South Africa (Tsoka, Sueur & Sharp 2003:70).For example, O' Mahony and Steinberg (1995) conducted a survey of two hundred wom en on place of delivery, home delivery practices and an ten atal care fo r the m ost recent delivery (within the previous 5 years) from randomly selected clusters, obtained from a m ultistage random sam pling 54

Curationis March 2006
process, in rural Transkei (Eastern Cape including both form er T ranskei and Ciskei, and South Africa).Two-thirds had delivered at home (home-based) and onethird within the health services (settingbased).Of those who delivered at home, 62 (47%) were alone at the time of delivery while the remainder were assisted by a close relative or neighbour; 38% had one or m ore risk facto rs fo r o b stetric complications.Uyirwoth, Itsweng, Mpai, Nchabeleng and Nkoane (1996:91) found in a rural community in the Limpopo Province among mothers who delivered within 1 2 months before the date of interview that the proportion of health facility deliveries was 74.6% while 26.3% o f all b irth s o ccu rred at hom e. Inaccessibility to maternity services, lack of money, negative staff attitudes and lack of privacy were the common reasons given for preference of home delivery.
A ccessibility factors (cost, distance, transport, availability of health facilities, and nurses' attitudes) were also major barriers, whereas traditional beliefs were reported as less significant, in a study in N orthern Ghana (M ills & Bertrand, 2005:45).Other studies found that the attitudes of mothers or mother-in-laws and social influence from the spouse played a role in the delivery options (Amooti-Kaguna & Nuwaha 2000:203;Duong et al. 2005:172).
Since the im p lem en tatio n o f free maternity services for pregnant women and children under six years of age in 1994 in South Africa, McCoy (1996:5) noted that although more m aternity services were provided in rural areas, they were still inaccessible to many women.
Use of skilled professional childbirth attendants by all women remains a goal of safe motherhood programmes, despite well-known barriers.

Method Design and setting
A cross-sectional sample of pregnant women recruited from primary health care clinics (w ith PMTCT) and from the com m unity.In addition, m others or mothers-in-law and husbands or partners of the pregnant women were recruited into the study.

Data analysis
Descriptive statistics were calculated using SPSS package (version 12.0).Chisquare tests were used to compare the hom e-and settin g -b ase d groups.
M u ltiv a riate lo g istic an aly sis was conducted to examine the relationships betw een deliv ery optio n and independent variables.

Sample characteristics
Almost all of the pregnant women were Xhosa by ethnicity (97%).The mean age of pregnant w om en was 29.3 years (SD=6.9years), with a range of 15 to 53 years and 8 % 19 years and younger.On average the pregnant women had 2 .8natural children and 6.4 dependents.The 56

Curationis March 2006
majority of the pregnant women were married or cohabitating (54.7% ) and 41.5% were never married.Two in five pregnant women were living with their parents, 34% w ith th eir p artn e r or husband and 14% with their in-laws.Eleven percent of the women had no formal education, 45% had Grade 5 to 8 , and 28% had G rade 1 to 4 form al education.Three-quarters lived in a traditional African house while onequarter lived in a brick house.Only a few of the wom en (14% ) were form ally employed (see Table 1).

Delivery options
From the 870 pregnant w om en and mothers, 55.3% had delivered their last child in a hospital, 0 .6 % in a clinic and 44.1 % at home; from those who delivered at hom e 38% w ere a ssiste d by a traditional birth attendant (TBA) and 62% delivered without a TBA (see Table 2).
The analysis of the comparison between pregnant women who had delivered their last child at home (home-based) or in a health facility (setting-based), found in terms of sociodemographic variables that m ore yo u n g er w om en, w ith higher educational levels, higher econom ic levels (regarding the type of house), being single and having the first child had delivered their last child in a health facility, and in terms of access to health facility pregnant women with longer time periods to get to the nearest health facility (clinic and hospital), having less money for transport to get to the health facility and having no access to telephone were more likely to have delivered their last child at home.The maternal age among the home-based group (M =30.6) was higher than among the settings-based group (M=28.3)(t=496, pc.001) (see Table 3).

Delivery intentions
W hen pregnant wom en were asked about where they intend to deliver their baby, most (93.7%)indicated in the health facility and 6.3% preferred to deliver at hom e. T here seem s to be a large difference between 44% home delivery with the last child and the intention to deliver at home of 6 % with the current pregnancy.Although almost two-thirds (63%) of the mothers or mother-in-laws of the pregnant women indicated that they had delivered their last bom child at home, almost all of them (97%) wished that their daughter would deliver in a health facility.Likewise almost all the partners or husbands (97% ) o f the participants also indicated that they wished their partner of wife to deliver the current pregnancy in a health facility.
Major reasons for the intention to deliver in a health facility mentioned by mothers, in-laws and partners were the belief in good health care from a doctor or nurse and to avoid complications.Comparing pregnant women who had delivered their last child at home or in a health facility, past home delivery of one self, home delivery of the mother or in-law, mother or in-law s and partner or husbands intention of home delivery were associated with home delivery intention (see Table 4).

Logistic regression
We explored the relationship between delivery options (in the past) and independent variables including intention to deliver, mother or mother-in-law's delivery practice with last bom, mother or m o th e r-in -la w 's in ten tio n to d eliv er pregnancy and p artn er or h u sb a n d 's in ten tio n to d eliv er pregnancy.Stepwise logistic regression analysis resulted in two significant factors for home delivery, mother or mother-in-law delivered her last bom at home (OR=3.19;95%CI=2.01-5.08)and pregnant woman's intention to deliver at home (OR=22.91;95%CI=2.96-177.47);for the latter the sample size was too small.

HIV/AIDS knowledge and HIV testing
The majority of the pregnant women (81%), mothers or mother-in-laws (74%) and partners or husbands (74%) knew that a pregnant woman infected with HIV or AIDS can transmit the virus to her unbom child.Only 9% of the pregnant women and 13% of their partners or husbands ever had an HIV test.Major barriers of pregnant women for not having had an HIV test were fear of being HIV positive, not aware where to get tested and unsure if the test results will rem ain co n fid e n tia l.H aving been provided with HIV/AIDS information, wanting to know their HIV status and concern for the transmission of HIV from mother to the unbom child were given as major factors that would encourage them to go for an HIV test.Further, almost all mothers or mother-in-laws and partners or husbands of the pregnant women w ould, how ever, encourage th eir daughter or wife to have an HIV test done.HIV know ledge, HIV testing behaviour and attitudes were found to be not associated with the delivery option (home-based or setting-based).Having had contact with an HIV community worker was significantly related with home delivery (see Table 5).

Discussion
The study found among this rural sample of pregnant women who had delivered before that 55.3% had delivered their last child in a hospital, 0 .6 % in a clinic and 57  S teinberg 1995: 1168).Similarly, the m others or m o th er-in -la w s o f the pregnant women studied indicated that two-thirds of their last born child was delivered at home.More re cen t stu d ies seem to show a trend to increased health facility delivery, for example in a recent study am ong w om en in ru ral KwaZulu-Natal 77% had delivered in a health facility (Tsokaetal. 2003:70).This study found that the odds of reaching a health care facility increase by almost 3 fold if women stay nearby than those who stay far from the h o sp ital (OR=2.87),also women who are more educated have 1.5 chance of accessing the 58 S ocial facto rs appeared to contribute to the low utilization of maternity health services.Childbirth experiences of the mother or mother-inlaw greatly in terms of home delivery influenced the delivery choices of the pregnant women in this study.Having support from the mother-in-law was associated with home delivery.However, almost all pregnant women, their mothers or m o th er-in -law s and partners or husbands intended to deliver the current pregnancy in a health facility but only the pregnant women's intentions were significantly related with the past delivery option.Experience elsewhere has shown that most of these women intend to deliver at health facilities but cannot find transport when they go into labour at home (Tsoka et al. 2003:72).It appears that the attitudes of mothers or motherin-laws did not play a significant role in the delivery option in this study, while the influence of in-laws and extended family in rural Vietnam (Duong et al.The majority of the pregnant women, their mothers or m other-in-laws and p artn ers or hu sb ands knew that a pregnant woman infected with HIV or AIDS can transmit the virus to her unborn child, but only 9% of the pregnant women and 13% of their partners or husbands ever had an HIV test.Further, almost all mothers or mother-in-laws and partners or husbands of the pregnant women w ould, how ever, encourage their daughter or wife to have an HIV test done.HIV know ledge, HIV testing behaviour and attitudes were found to be not associated with the delivery option (home-based or setting-based).

Conclusion and recommendations
In order to make delivery safer, there is a need to improve access to maternity services by reducing the distance to maternity facilities, increase the number of skilled personnel, provide emergency transport (those that come from far, arrangem ents for "w aiting m others" accommodation near the point of delivery in the absence of suitable or reliable transport should be made), train TBAs and equip them with delivery kits in case of emergency delivery, and health and social improvements such as increased female health education.Primary care providers have a strong influence over women's perceptions of antenatal care and are in a good position to provide appropriate edu catio n al messages.Health promotion strategies should be based on exploring patients' ex p lan a to ry m odels o f h ea lth in pregnancy and childbirth.

Limitations
The quality of maternity services in a health facility and at hom e (with or without a TBA) and prenatal services was not assessed (Duong et al. 2004:167) in this study but can greatly contribute to delivery choices.Data were collected from the self-report of respondents.Such incur recall bias, especially with regard to travelling time and costs of transportation to the nearest health facility.

Table 1 : Sample characteristics
Peltzer et al. 2005:5-6)including the follow ing sections containing both closed and open-ended questions for: A) the pregnant woman: demographic and socio economic data (9 items), access to health facility (4 items), pregnancy and delivery related 55 Curationis March 2006 Q the husband or partner: demographic information (4 items), wife's/partner's delivery related questions (4 items), HIV knowledge, HIV testing behaviour and attitudes (8 items, Cronbach alpha was .65)

Table 3 : Sociodemographic and access variables by home-based and setting-based delivery groups
***p<.001, **p<.01,*p<.05; # Pregnant women recruited from three clinic areas were on average less than 35 km (measured from the location of the clinic) and from two clinic areas on average more than 35 km away from the nearest hospital.