Home visits by community health workers in rural South Africa have a limited, but important impact on maternal and child health in the first two years of life

Background: More than 50% of Africa’s population lives in rural areas, which have few professional health workers. South Africa has adopted task shifting health care to Community Health Workers (CHWs) to achieve the Sustainable Development Goals, but little is known about CHWs’ efficacy in rural areas. Methods: In this longitudinal prospective cohort study, almost all mothers giving birth (N = 470) in the Zithulele Hospital catchment area of the OR Tambo District were recruited and repeatedly assessed for 2 years after birth with 84.7-96% follow-up rates. During the cohort assessment we found that some mothers had received standard antenatal and HIV care (SC) (n = 313 mothers), while others had received SC, supplemented with home-visiting by CHWs before and after birth (HV) (n = 157 mothers, 37 CHWs). These visits were unrelated to the cohort study. Multiple linear and logistic regressions evaluated maternal comorbidities, maternal caretaking, and child development outcomes over time. Results: Compared to mothers receiving SC, mothers who also received home visits by CHWs were more likely to attend the recommended four antenatal care visits, to exclusively breastfeed at 3 months, and were less likely to consult traditional healers at 3 months. Mothers in both groups were equally likely to secure the child grant, and infant growth and achievement of developmental milestones were similar over the first 2 years of life. Conclusion: CHW home visits resulted in better maternal caretaking, but did not have direct benefits for infants in the domains assessed. The South African Government is planning broad implementation of CHW programmes, and this study examines a comprehensive, home-visiting model in a rural region. This longitudinal prospective observational cohort study was conducted with approval of the Institutional Review Board of the Stellenbosch University (N12/08/046) and permission for the recruitment to occur in government health facilities was granted by the Eastern Cape Department of Health. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. This article does not contain any studies with animals performed by any of the authors. Zithulele Hospital lies within the King Sabata Dalindyebo (KSD) Sub-district of the OR Tambo District of the Eastern Cape Province of South Africa. The Eastern Cape Province has some of the worst child health indicators in South Africa. The OR Tambo district, in particular, is one of the most under-developed and impoverished districts in the country [25]. Most households lack electricity, running water, and access to toilet facilities. The unemployment rate is well over 70% and the average annual income per household is R14 600 ($1518 in 2013) [26]. The hospital has a catchment area of nearly 1300 km2 and serves a population of 130,000–150,000 [26]. From January through April 2013, a consecutive series of mothers giving birth at Zithulele Hospital and its 10 closest clinics, as well as mothers who had a home birth in the areas covered by the ten clinics, were approached to participate in a birth cohort study, the Zithulele Births Follow up Study (ZiBFUS). Mothers who gave birth at the hospital but lived outside this geographical catchment area were excluded. From 493 live births, 470 mothers provided written voluntary informed consent (4.7% refusal rate). In the case where the mother was less than 18 years old (16.4%), written consent was obtained from the adolescent mother and one of her parents/guardians. Written informed consent was obtained from all individual participants included in the study. In the cohort of 470 mother-infant pairs, we found that mothers living in certain villages in the hospital catchment area had received home visits by CHW of a non-profit organisation (NPO), the Philani Nutrition Trust, during the antenatal period and during their baby’s first 2 years of life. Mothers visited by these CHWs were labelled as being part of the home visiting (HV) group, while the others were designated as the standard care (SC) group. These home visits were unrelated and not part of the longitudinal prospective cohort study. The Philani Maternal, Child Health and Nutrition Trust is a NPO operating in the catchment area of Zithulele Hospital in the Eastern Cape since 2010. Philani Mentor Mothers (CHWs) are recruited from the areas where they live and are trained to conduct home visits in their communities. The training includes a four-week course on maternal and child health, HIV and TB, followed by 2 weeks of in-the-field training and shadowing of experienced CHWs. There is a strong emphasis on careful recruitment, ongoing training, supportive in-the-field supervision and accountability [18]. CHW are assigned a catchment area and systematically conduct home visits in their designated area/village. CHW offer to weigh children to make sure they are growing well, as a way to gain entry into the household. While there, they provide health education about nutrition, breastfeeding, HIV, PMTCT, TB and the importance of antenatal care and find out if any of the women in the household are pregnant. Some pregnant mothers are found by word of mouth in the community. A file is opened for all pregnant mothers and for all children who are under-weight for age, stunted or wasted, and these households are paid careful attention to by the CHWs and regularly visited. The Philani CHW programme schedule requires that mothers are visited every month early in pregnancy, twice in the eighth month of pregnancy, weekly in the ninth month of pregnancy, within 2–3 days of arriving home after birth, weekly in the first month after birth and then every 2 weeks up until the infant is 1 year old and monthly after this. All mothers in both the Standard Care Group and the Home Visiting Group were recruited to the study in exactly the same way – 1-2 days post birth if delivering in hospital and within 2 weeks if delivering at home – by a research team completely independent of the CHW programme. 77% of mothers recruited gave birth at the hospital, 9% at a community health centre, 4% on the way to a health facility and 10% at home. To identify women who gave birth at home or in one of the feeder clinics/community health centers, the research team relied on nurses at the clinic to identify these cases and offered nurses a small airtime reimbursement to contact the field workers with the woman’s telephone number. The majority of home births were referred to us by nurses at the time of the mother’s first visit to the clinic, usually within a few days after birth. Field workers (known as data collectors) would then visit mothers at home. We believe that virtually all home births during the recruitment period were identified in this way because of the strong incentive for mothers to visit a clinic or hospital soon after birth to acquire a Road-to-Health card (RtHC) for their baby. Not only is the RtHC a type of health ‘passport’ and an extremely important health record, but it is also used to apply for a child support grant ($25 per month in 2013) from the South African Social Security Agency (SASSA). Mothers were assigned to the Standard Care group (SC) or the Home visiting group (HV) at analysis, based on whether or not they indicated that they had been visited by a Philani Mentor Mother (CHW) antenatally or after delivery. All women recruited to the study have access to free primary health care, which includes antenatal care and HIV care, most often provided at clinics. At the time of the study, “Option A” Prevention of Mother-to-child-transmission (PMTCT) was in place, in which mothers living with HIV that have a CD 4 count of less than 350 are started on lifelong highly active antiretroviral therapy (HAART) and those with a CD4 of over 350 receive Zidovudine (AZT) from 14 weeks gestation and a stat dose of nevirapine (NVP) during childbirth, and infants of mothers living with HIV are given NVP syrup. All mothers living with HIV were encouraged to continue to monitor their CD4 count and to test their children for HIV at 6 weeks of age and again 6 weeks after the conclusion of breastfeeding. Mothers in the HV group lived in areas where CHWs had been active in the community for at least 2 years, as described above and indicated that they had been visited by a Philani Mentor Mother (CHW). A separate team of local women underwent extensive training as Data Collectors, with role-plays, certification, and random observations in the field over time. Interviewers recorded interviews and collected data on programmed mobile phones. When possible, information collected from the mothers was verified on mother’s antenatal clinic cards or children’s Road to Health Cards (RtHC). The data was also later reviewed with a supervisor and randomly audited. Mothers were interviewed and infants were assessed at six time points: during recruitment near the time of giving birth; at 3 months (85% follow-up, n = 390/460); at 6 months (92% follow-up, n = 420/456), at 9 months (88% follow-up, n = 410/454); at 12 months (91% follow-up, n = 411/450), and at 24 months (88% follow-up, n = 396/450). There were 22 infant deaths and all the second twins of the 9 sets of twins were not included in these analyses, to avoid double counting those families in the data set. The primary analysis compared mothers receiving HV and mothers receiving SC using mixed effects regression models, with mothers receiving SC as the reference group. Logistic mixed effects regression models were used for binary outcomes. All models were adjusted for repeated measures, where appropriate, and a random participant effect was used to control for the longitudinal nature of the assessments. For all analyses, all observations that have a missing value for any one of the variables used in the model were excluded. All analyses were carried out using IBM SPSS Statistics (Version 20, Armonk, NY: IBM Corp). At baseline, mothers reported their age, education level, employment, their household make-up, household monthly income, access to a water source, and electricity. In addition, they reported on previous pregnancies and HIV status. HIV status was confirmed in the mother’s antenatal clinic card and the RtHC and compared to the Zithulele Hospital maternity statistics. Depression was reported using the Edinburgh Postnatal Depression Scale (EPDS) [27]. We report both the mean scale score and identify mothers whose responses indicate probable depressive disorder using a cut-off of EPDS > 13 to indicate depressed mood [28]. Alcohol use prior to and during pregnancy was self-reported by the mothers. IPV was self-reported prior to pregnancy and during pregnancy. Mothers who completed at least four antenatal care visits, by report and confirmed on the antenatal card, were reported. Mothers who tested for HIV during pregnancy, by report and confirmed in antenatal clinical card, were reported. Of mothers living with HIV, we report on four PMTCT tasks: 1) taking appropriate ART before birth; 2) giving the child nevirapine as appropriate; 3) 6-weeks post-delivery PCR test done; and 4) whether PCR was done by 6-months post-delivery. PCR status was both self-reported and verified by interviewers assessing the RtHC. The number of months of exclusive breastfeeding and those completing 6 months of breastfeeding was self-reported by mothers. Immunisations were recorded at each assessment as documented in the child’s RtHC. With informed consent, photographs were also taken of the immunisation page in the RtHC to ensure accuracy. We report on whether all required immunisations were complete at each interview. Mothers that secured a child support grant at each follow-up assessment were documented. Finally, mothers self-reported visiting traditional healers in the last 3 months. Low birth weight was recorded for those infants < 2500 g. Infant growth was measured at each assessment. Children’s weight and height measures were then converted to Z-scores based on the WHO age-adjusted norms [29]. Growth is reflected in standardized scores (Z-scores) for height-for-age (HAZ), weight-for-age (WAZ), and weight-for-height (WHZ). A Z-score below − 2 SD was considered a serious health deficit: either stunting (> − 2SD for HAZ), wasting (> − 2 for WHZ) or severe underweight for age (> − 2 SD WAZ) [30]. The gross motor developmental milestones of the WHO for children at 6 (WHO1), 9 (WHO3), 12 (WHO1–5) and 24 months (WHO1–6) were administered [31, 32]. These include six milestones that are fundamental to acquiring self-sufficient locomotion: sitting without support, hands-and-knees crawling, standing with assistance, walking with assistance, standing alone, and walking alone [33] Each task was scored (0) if unable to perform, (1) if reported able, but not demonstrated, and (2) if demonstrated able to perform. A single developmental score was calculated by taking the mean of all the WHO Gross Motor Developmental milestone tasks [33].