Health system weaknesses constrain access to PMTCT and maternal HIV services in South Africa: a qualitative enquiry

HIV testing within ANC is the entry point to the care continuum for pregnant women. Across the facilities studied, a significant proportion of the HIV-positive pregnant or postnatal women interviewed failed to receive an HIV test during their first ANC visit, mainly due to shortages in staff and supplies. In both Eastern Cape hospitals, nurses provided all counselling and related HIV services, with a single nurse per facility running the PMTCT programme and offering all HIV counselling. In addition to their other duties, the 'assigned' nurses provided PMTCT services for about five hours a day (8:30 am to 1:30 pm) from Monday to Friday. As a means of coping with this workload, one nurse explained: "I provide five counselling sessions per day, and then I stop [because] I have other work to do" (Eastern Cape hospital, October 2008). If this nurse was ill or undergoing training elsewhere, HIV services were simply not available. Infrequently, an HIV counsellor or doctor would assist in providing some counselling, although many respondents believed that doctors were too busy to provide optimal counselling. The nurses acknowledged that there was generally, then, no HIV testing and counselling provided for patients admitted during the afternoons, weekends, or on public holidays.

Such nursing shortages are evident throughout the country. In 2008, for example, Health Systems Trust data documented a nurse staffing gap of 36% for public sector posts nationwide and 40% for the Eastern Cape, with some provincial deficits registering upwards of 50% [20]. Nurses interviewed spoke of the challenge of attracting and retaining health personnel in the Eastern Cape, especially in certain peri-rural towns. At the Eastern Cape tertiary-facility, only half of the 600 nursing posts were filled. Respondents there stated that it was commonplace for nurses to depart for more promising posts in the private sector or overseas. In the month when interviews took place, three nurses at the facility were leaving at that month-end alone (interviews with key informants, Eastern Cape academic hospital, October 2008).

In the Eastern Cape clinic, shortages of HIV test kits and stock-outs of nevirapine, were reported by staff. The popular press, together with academic sources, found similar problems with drug procurement and supply bottlenecks in other parts of the country [21‐23] (see Table 1, Table 2). By contrast, according to both patients and key informants, the hospitals in the Eastern Cape and Johannesburg had no such supply problems. However, in Johannesburg, systems' failures took the form of frequent delays in payment to lay HIV counsellors who were responsible for testing and counselling. Absenteeism and low staff morale were common. Indeed, over the past few years a leading South African NGO, the AIDS Law Project (now operating under the name SECTION27), had called for the Department of Health to address the poor employment conditions of lay counsellors, pressing for legal action to address this chronic problem [24‐26].

Another consistent delay for HIV-positive women concerned the timely receipt of their CD4 cell count results, a necessary step for discerning ART eligibility. When women attended their second or third ANC visit, they often could not commence ART as their CD4 cell counts were still unavailable. Patient files indicated that many HIV-infected women, though eligible for ART, had already delivered before initiating ART or PMTCT prophylaxis, either due to the above-mentioned systems' failures, or, in some instances, preterm delivery. A further group of women began ART late - just prior to childbirth - making optimal prevention and treatment outcomes less likely [27].

Postnatal care constitutes the next component of the care continuum, where there are a number of opportunities for protecting the health of the woman and her newborn by optimizing HIV prevention and treatment. During breastfeeding, for example, the efficacy of ARV drugs taken in pregnancy and during labour is reduced over time, [28, 29], with postnatal HIV transmission responsible for up to half of HIV infections in South African children. Mixed feeding carries a particularly high risk [30, 31]. Feeding options need to be clearly explained and women counselled on the implications of their feeding choices during the early postnatal period. This study found that one of the weakest aspects of PMTCT interventions is counselling women on infant feeding. Across the facilities, many HIV-positive women struggled with feeding choices, with a number practicing mixed feeding, unaware of the increased risks of transmission. This reflects the poor and ad hoc counselling received by women during ANC and postnatally.

Based on interviews with pregnant or postnatal women, during 'counselling' about infant feeding options, healthcare workers in many instances appeared to 'steer' women towards their own preference, encouraging women to do what the health personnel believed to be 'right' or 'proper'. This often resulted in inappropriate choices given women's available resources - in terms of money, time, and access to safe water. For example, in one of the Eastern Cape hospitals, records showed that 97% of women in August 2008 and even 100% of women in September of that year elected to formula feed. While free formula is available in clinics across the country, only 9% of households in the surrounding district have potable water [32] - meaning that women in this district would struggle to ensure safe formula feeding. One woman observed: "I wasn't given feeding options - I was simply told to formula feed" (Johannesburg, May 2008). Another said: "The nurse told me that formula feed was the only safe option - she did not give me a choice" (Johannesburg, June 2008).

Infant HIV diagnosis is critical, especially early diagnosis (and subsequently ART if required), but has proved challenging in South Africa [33]. Organisation of services in a vertical manner accounts for much of these difficulties, together with the related problem of limited locations for testing infants in the these peri-urban facilities.

In the Johannesburg site, polymerase chain reaction (PCR) testing required mothers to take their infants to the paediatric virology ward, a different location from where they had attended antenatal and postnatal care but within the same facility. In the Eastern Cape, women were required to attend an entirely different hospital, as only the academic hospital in the district offered PCR testing. In both settings, health personnel were meant to direct women accordingly, however, many women appeared unaware of this information.

A former nurse interviewed in the Eastern Cape clinic noted that the 'tins' used for formula feeding were associated with stigma (October 2008). This was confirmed by patients and health personnel interviewed, and has been identified in previous studies [34]. One woman noted: "I hide it [her HIV-positive status]. I say the baby doesn't like breast milk to anyone who asks why I am not breastfeeding" (Johannesburg, August 2008). On this theme, another woman, when asked how she managed formula feeding, said: "I put the formula in a canister without a label (e.g., a can for instant coffee). I worry about what people think, so I cannot tell anyone about my status outside my family. I keep it to myself" (Johannesburg, August 2008).

One health worker also noted: "People are scared of themselves" and "stigma prevents people from testing" (Johannesburg, August 2008). Attributing her experiences to discrimination, one pregnant woman stated: "I was turned away at X and Y clinics. I was already on ARVs. Maybe they turned me away because I was HIV positive?" (Johannesburg, June 2008).

While patients, healthcare workers and researchers agree that stigma is abating somewhat, it remains pervasive. Human Rights Watch noted: "People living with HIV and AIDS in South Africa continue to fear discrimination and victimisation. Few people choose to publicly disclose an HIV-positive status, fearing that this will cause stigmatisation in their community and loss of their jobs" [35].

Women's HIV status has an impact on their mental health, which can then affect their willingness and ability to seek health services and care [36]. Shock, denial or uncertainty can delay women's return to health facilities for the next step in HIV service provision, namely, ART initiation. While such individual barriers may be difficult to obviate completely, comprehensive counselling can mitigate this. One woman stated: "There is stigma attached to HIV. I cope by not telling people because people will criticize. I gain support from one of the counsellors at the hospital when I feel low" (postnatal patient, Johannesburg, June 2008). An antenatal patient said: "Being HIV positive was difficult at first. But since I have had HIV counselling here I feel strong because of the counselling" (Johannesburg, April 2008). Patients described how, when they did see a nurse or counsellor, health personnel would share strategies about adherence, disclosure and how to deal with in-laws who disapprove of women who do not breastfeed. At other times, however, health personnel played a more directive, even invasive, role. One patient reflected: "I was in denial about going onto ARVs and refused at first. It was only when I went back to a second HIV counsellor that she said 'you are killing your child by not taking the medication'". The patient then "became very worried about the baby's health and I was frightened into action". After she told her husband she was taking ART: "He reacted in a violent manner and threw the pills away". The counsellor then helped the woman put "the pills in a different place to take the pills in secret" (Johannesburg, July 2008).

Thus, though there were many examples of health workers seemingly being overwhelmed by their workload and working conditions, this did not always prevent them from offering assistance to women, often drawing on years of experience from previous interactions with HIV-infected women. Health personnel who knew the patient's status - and offered guidance about the complex challenges facing HIV-positive women in South Africa - were thus able to play a critical support role for some women in this study.

The facility in Johannesburg kept routine statistics, which were computerised. However, in the Eastern Cape, there were no computers at two of the three facilities, and information was recorded manually. The type of indicators recorded - and the actual figures tallied - seemed to reflect only a portion of the actual PMTCT and ART activity, and the data were generally of poor quality. Consequently and likely perpetuating these poor monitoring practices, what little information health staff collected was not being used to improve current practices and systems: "There are no feedback loops for quality improvement" as one key informant noted (Eastern Cape hospital, June 2009).

In the Eastern Cape (non-academic) hospital, in terms of HIV testing, the numbers of women who apparently tested for HIV were 24 in August and 25 in September (Figure 1). Yet there were a total of 107 live births to HIV-positive women who delivered in August; and 92 in September. Over the two months, of the 24 women tested, 11 women (or 50%) tested positive for HIV in August; and in September, of the 25 women tested for HIV, 9 women (26%) were HIV positive: but only 13 women (12%) in August and 14 (15%) in September were apparently issued NVP. In the figures available, only 2% of women in August (n = 107) and 9% of women in September (n = 92) were initiated onto ART. The figures, however, seldom tally. For example, in August, the number of women choosing to exclusively breastfeed plus formula feed equals 109 while the total number of HIV-positive women giving birth during August is 107 (two more than the total). This suggests double counting, incorrect counting and generally poor record keeping.

Overall, in terms of data availability and quality, one data capturer said that 70% of the data were simply not recorded at all (Eastern Cape Province, October 2008). Referring to this province, another key informant observed: "Data quality is very poor across the province". "Statistics in the nevirapine register are accurate" but "some statistics are double-counted" and "they carry over figures from the previous month". Health personnel fill in information, but they are not working from a common definition of an indicator (Eastern Cape hospital, October 2008). Ultimately, the actual performance of the PMTCT and ART programme in the Eastern Cape facilities appears largely unknown.