Non-adherence to the single dose nevirapine regimen for the prevention of mother-to-child transmission of HIV in Bindura town, Zimbabwe: a cross-sectional analytic study

Zimbabwe has the 4^th highest rate of HIV infection in the world, with an estimated HIV prevalence of 15.6% in adults aged 15-49 years [1, 2]. The HIV pandemic has increasingly affected children, resulting in the reversal of the gains that had been made in the reduction of infant and childhood morbidity and mortality in the country.

Mother-to-child transmission of HIV (MTCT) during pregnancy, delivery, or breastfeeding accounts for more than 90% of the HIV infections in children below 15 years of age [3, 4]. Without interventions, 20-45% of babies born to HIV positive women become vertically infected with HIV [5]. Studies have shown that short-course, easy-to-use, and affordable antiretroviral regimens can significantly reduce MTCT of HIV. One such regimen is the single-dose nevirapine (SD-NVP) regimen, in which the mother ingests a dose of nevirapine (NVP) at the onset of labor (at least 2 hrs before delivery) and another dose is given to the baby within 72 hrs of birth [4, 6‐8].

In Bindura district the PMTCT programme was started in 2003 at Bindura Hospital. In 2004 seven additional satellite comprehensive PMTCT clinics were set up in the district to increase the coverage of the services. Despite these efforts, the uptake of the PMTCT interventions remained unsatisfactory. During the period January to May 2007 1161 mothers delivered at Bindura Hospital. Of all these mothers, only 248 (21%) were tested for HIV before giving birth, the rest (913) had unknown HIV status and thus could not benefit from the PMTCT programme. Out of the 248 mothers who were tested for HIV before delivery, 65 (26.2%) were HIV positive. Of these 65, only 43 (66.2%) swallowed the recommended nevirapine doses before giving birth [9].

This then raised questions on why some pregnant mothers took the bold step of being tested for HIV but would not complete the recommended protocols for them to prevent the transmission of HIV to their babies.

In Zimbabwe the SD-NVP regimen remains the recommended prophylaxis for PMTCT. The government is planning to adopt more efficacious regimens, but the problems affecting the current SD-NVP regimen with regard to access, uptake and adherence, may still affect these future regimens. Identifying these problems can assist in planning the rolling out of the new PMTCT regimens. In this study we therefore set out to determine the extent of non-adherence and to identify factors associated with the non-adherence to the SD-NVP regimen in Bindura.

An analytic cross-sectional study was conducted in the four health facilities that provide comprehensive PMTCT and maternity services in Bindura town. The study population included mothers-baby pairs registered on the PMTCT programme and attended routine six-week postnatal care in these health facilities. We reviewed the health facility PMTCT registers and all registered HIV positive mothers (and their babies) who visited these heath facilities for postnatal care between March and July 2008 were enrolled into the study. We excluded mothers who were on antiretroviral therapy before giving birth. The minimum sample size was calculated as 197 mothers-baby pairs, using the formulae: [10], assuming an error risk (z_α) of 1.96, with a precision (Δ) of ± 5% and expecting that 15.1% (p) of the mothers would be non-adherent (as reported by Albreicht et al in the study: Predictors of non-adherence to single dose nevirapine for PMTCT in Lusaka Zambia) [11].

The mothers were interviewed using a structured questionnaire to collect information on socio-demographic, socio-economic, socio-cultural variables, knowledge on PMTCT as well as perceptions on health services being provided.

Ethical approval to conduct the study was granted by institutional review boards of the Provincial Medical Directorate (PMD) for Mashonaland Central Province, the Bindura Municipality Directorate of Health Services and the Health Studies Office of the Ministry of Health. The aim of the study was explained to all the potential participants and written consent was obtained before the interviews. Confidentiality was assured and maintained throughout the study and no identifying information was collected.

Statistical analyses were conducted using the Epi-info statistical software (CDC-Atlanta, 2007) [12]. We first performed descriptive analyses, looking at the extent of non-adherence to NVP and comparing the distribution of various characteristics among the non-adherent and adherent mother-baby-pairs. Non-adherence was defined as lack of ingestion or delayed ingestion of NVP by either the mother or the baby (delayed ingestion being taken as less than 2 hours before delivery for the mother and more than 72 hours after birth for the baby). We then performed bivariate analyses of the possible associations between non-adherence and the variables under investigation. In determining these associations prevalence odds ratios (PORs) and their 95% confidence intervals (95%CI) were calculated. We then conducted stratified analysis to assess for possible effect modification (interaction) and confounding. Lastly we performed logistic regression analyses where all independent variables that were significant at the 0.25 level (χ², p ≥ 0.25) in bivariate analyses were included in the model [13], to come up with adjusted PORs and 95%CIs.

We interviewed 212 mothers between March and July 2008. Of the 273 PMTCT mothers who were due to attend the six-week postnatal care in the participating health institutions during the study period, 229 (83.9%) attended the postpartum visit. Seventeen mothers were excluded for failing to meet the inclusion criteria, 14 had been on antiretroviral therapy before giving birth while the other three indicated that they were unwilling to respond to questions. The majority of the participating mothers, 131 (61.8%), were aged between 21 and 30 years, with a median age of 27.0 yrs (Q1 = 23; Q3 = 32). The mothers were largely married (formally or informally), 170 (80.2%) and most, 150 (70.8%), had attained at least secondary education. Of the 212 mothers only 63 (29.7%) were gainfully employed (Table 1).

Non-adherence to the SD-NVP regimen was high in Bindura, compared to some previous studies [11, 14] though it was comparable to one Kenyan study where they found non-adherence rates of up-to 55% [15]. The high non-adherence in our study was probably because our definition of non-adherence included the ingestion of a NVP dose outside the recommended times. Some previous studies defined non-adherence only as the failure to ingest NVP [11, 14]. Successful PMTCT prophylaxis also depends on the timing of the NVP ingestion. If the maternal NVP dose is ingested less than 2 hours before delivery the concentration in the blood may not reach enough therapeutic levels before giving birth and if the infant NVP dose is swallowed after 72 hours of birth the baby may not get the full benefits of the drug [16, 17]. It is therefore important to consider the timing of the NVP intake when looking at adherence to the NVP regimen. While most previous studies looked at adherence only to the maternal NVP dose, our study defined non adherence looking at both the maternal and the infant doses of NVP. The infant dose of NVP is equally important for the regimen to be effective and thus the prophylaxis cannot be considered complete without administration of the infant NVP dose. A few studies have also looked at both maternal and infant adherence to PMTCT prophylaxis [11, 18].

Of concern was that most babies born at home did not receive their NVP doses or received them too late. We attributed this to cultural practices in the studied population; newborn babies are confined within the house until the shedding of the umbilical cord stump, which usually takes about a week. The low self-disclosure of HIV results by the HIV positive mother also makes it difficult for her to convince the partner or the relatives of the need to take the baby to the clinic for the NVP dose before the week lapses. Other studies have also reported an association between home deliveries and babies missing their NVP doses [19, 20].

Non-adherence to the maternal dose of NVP was more than twice higher among mothers who had no secondary education. Maternal education enhances communication between the mother and healthcare providers and also improves retention of provided information, leading to better implementation of recommended interventions. Education also empowers the woman to have autonomy in making important decisions without relying on other people. Similar findings have been reported in previous studies [11, 19].

NVP non-adherence was four times higher among mothers who attended less than three antenatal sessions during pregnancy. Attending more ANC sessions offers more opportunities for healthcare providers to reinforce issues such as the importance of taking the NVP at the recommended times, the importance of delivering institutionally, the need to take the baby to the clinic in the event that birth takes place at home, and others which have a bearing on adherence to NVP prophylaxis. Investigators in a similar study in Rwanda also found an association between non-adherence to PMTCT protocols and attending few ANC sessions [19].

Maternal non-adherence to NVP was almost three times higher in multi-parous mothers. Such women usually experience precipitate labour pains which are often detected late, resulting in delayed ingestion of prophylactic doses. Similar findings were reported in Zambia [11].

Giving the mothers NVP tablets to take home during ANC improved adherence. The woman who gets NVP in advance can ingest the pill even if she gives birth at home. Also among the mothers who deliver in health institutions, those who get the NVP in advance are more likely to ingest the tablets on time than those who wait to take the NVP doses on arrival to the hospital labour ward.

There are some limitations in this study that are worth noting. Our measurement of non-adherence relied on accurate reporting by the mother and could have been affected by recall bias since some of the mothers delivered more than 6 weeks prior to the interviews, and their records did not indicate the timing of NVP ingestion. It is also possible that some women who failed to ingest their NVP doses could have falsely claimed otherwise, making the results to underestimate the prevalence of non-adherence. Lastly, this study only looked at mothers who reported for postnatal care, there is the possibility that those women who did not attend postnatal care could be more non-adherent, possibly underestimating the true non-adherence.

Non-adherence to the SD-NVP regimen was high in Bindura. Giving the HIV positive mother a nevirapine dose to take home during antenatal care, giving birth in health institutions and self disclosure of HIV results by the mother to the partner were associated with reduced nevirapine non-adherence. Nurses working in the PMTCT clinics should ensure that HIV positive mothers are given nevirapine tablets at first contact, regardless of the stage of the pregnancy. Health promotion should be intensified at the community level to encourage ANC attendance and institutional deliveries. The government should consider adopting a policy of dispensing infant NVP doses during ANC, as is done for the maternal dose, and the possibility of incorporating traditional birth attendants into the PMTCT programme, so as to increase access to the ARV prophylaxis.