Community health workers can improve child growth of antenatally-depressed, South African mothers: A cluster randomized controlled trial

Background: Maternal antenatal depression has long-term consequences for children’s health. We examined if home visits by community health workers (CHW) can improve growth outcomes for children of mothers who are antenatally depressed. Methods: A cluster randomized controlled trial of all pregnant, neighbourhood women in Cape Town, South Africa. Almost all pregnant women (98 %, N = 1238) were recruited and assessed during pregnancy, two weeks post-birth (92 %) and 6 months post-birth (88 %). Pregnant women were randomized to either: 1) Standard Care (SC), which provided routine antenatal care; or 2) an intervention, The Philani Intervention Program (PIP), which included SC and home visits by CHW trained as generalists (M = 11 visits). Child standardized weight, length, and weight by length over 6 months based on maternal antenatal depression and intervention condition. Results: Depressed mood was similar across the PIP and SC conditions both antenatally (16.5 % rate) and at 6 months (16.7 %). The infants of depressed pregnant women in the PIP group were similar in height (height-for-age Z scores) to the children of non-depressed mothers in both the PIP and the SC conditions, but significantly taller at 6 months of age than the infants of pregnant depressed mothers in the SC condition. The intervention did not moderate children’s growth. Depressed SC mothers tended to have infants less than two standard deviations in height on the World Health Organization’s norms at two weeks post-birth compared to infants of depressed PIP mothers and non-depressed mothers in both conditions. Conclusions: A generalist, CHW-delivered home visiting program improved infant growth, even when mothers’ depression was not reduced. Focusing on maternal caretaking of infants, even when mothers are depressed, is critical in future interventions. Trial registration: ClinicalTrials.gov registration # NCT00996528. October 15, 2009 The Institutional Review Boards of University of California Los Angeles (UCLA), Stellenbosch University, and Emory University approved the study, whose methods have previously been published [34]. Three independent teams conducted the assessment (Stellenbosch), intervention (Philani Maternal, Child Health, and Nutrition Project, hereafter referred to as Philani), and data analyses (UCLA). Cape Town contains five major peri-urban settlements (townships) with formal and informal rudimentary housing. Unemployment in Cape Town townships is estimated at between 25 and 50 % [35]. Most women live within 5 km of a prenatal clinic. In each area, there is formal housing and vast areas of informal houses (shacks). In 2009, 26 township neighborhoods were matched on size (450–600 households), density, public utilities (water, electricity, toilets), distance of primary health care, and the number of alcohol bars. UCLA randomized matched pairs of neighborhoods to either the PIP or SC condition. The minimum number of pregnant women needed per neighbourhood to achieve 80 % power to detect a standardized effect size of 0.40 set the sample size; the original size was 1238. Participant flow through each phase of the study can be seen in Fig. 1. Trial profile A recruiter obtained consent to contact and repeatedly visit all households in all neighbourhoods from May 2009 to September 2010 in order to identify all pregnant women. Only 2 % of pregnant women refused participation. All mothers had relatively easy access to antenatal and postnatal care. Treatment for depression is theoretically possible, but it is highly unlikely that health care providers would identify or refer cases of depression, unless symptoms of active psychosis were detected during a clinic visit. Because the HIV rate in this population is about 26 %, standard clinic care includes HIV testing of virtually all women and extensive treatment for Mothers Living with HIV (MLHIV), including the following: dual therapy for Prevent Mother To Child Transmission (PMTCT), referral to antiretroviral therapy (ART) for women with CD4 counts below 200 or World Health Organization Stage 4 disease, the return of PCR test results for infants by 6 weeks of age, and co-trimoxazole for exposed infants starting at 6 weeks of age [36]. Nutritional supplements were also available from clinics during this study period. Philani is a non-governmental organization (NGO) that has been operating in Cape Town townships for more than 30 years. Local township women with good social skills and who were mothers themselves were recruited and trained as CHW to visit the homes of pregnant women and intervene to reduce alcohol misuse, increase adherence to perinatal HIV regimens, and boost child nutrition. CHW’s were trained in: 1) cognitive-behavioural approaches to establishing healthy routines and to problem-solving around goal setting, choices, triggers, and shaping of desirable behaviours; 2) key information about general maternal and child health, techniques for framing each health issue that is a risk (nutrition, alcohol, and HIV), and strategies for applying the health information in families’ daily lives; and 3) coping with their own life challenges. Trainers actively rehearsed and observed videotapes about how to solve challenging life situations, build engaging relationships, and keep responsibility for change with the mother. CHW were supervised weekly (face-to-face supervision as well as via mobile phones), and randomly observed twice a month. Specific PIP content has been described in more detail elsewhere [34]. A driver transported all participants to a central assessment site, allowing interviewers to be blinded to condition. All women provided written informed consent to participate in data collection. All data were collected using mobile phones [37]. Data collectors were taught methods of assessing infants’ growth, strategies for building rapport and collecting honest answers, and how to interact in a non-judgmental manner. Before going into the field, all data collectors observed five client interviews and practiced interviewing skills. All CHW were certified by supervisors prior to becoming interviewers. Maternal mood was assessed at the baseline recruitment interview during pregnancy and at 6 months post-birth using the Edinburgh Postnatal Depression Scale (EPDS), a 10-item measure, with items rated on a scale of 0–3 for severity [38]. The EPDS is a screening tool and is not used to make a confirmatory diagnosis of clinical depression; however, for the purposes of this study, we used a cut-off score of >18 to indicate the probable presence of depression [23]. Screening tools, by definition, will not have the sensitivity or specificity of a clinically validated tool such as the Standard Clinical Interview [39]. If the goal was to measure prevalence, we would have wanted to maximize sensitivity and accept some degree of false-positives. In this study however, we wanted to maximize specificity to reduce false positives, so we used the highest recommended cut-point to increase the chances that a screen positive was a definite case, likely to represent cases of severe depression [40]. Weight and length was measured at two weeks post birth and at six-months post-birth. Weight was assessed on scales calibrated monthly. Length was measured supine, using a roller meter with infant feet at zero position and bringing the headboard down to the crown of the head. Standardized z-scores (height-for-age, weight-for-age, height-for-weight-for age) were generated using new World Health Organization (WHO) child growth standards for infants under 24 months of age and calculated from WHO Anthro-2005 software. Infants were then categorized as underweight if they had a weight-for-age z-score (WAZ) of < −2 and stunted if they had a length-for-age z-score (HAZ) of < −2 [41]. Newborn infants were also provided with a “Road to Health” card on which birth weight, length at birth and delivery date were recorded. These measures were highly consistent with the interviewers’ measure (r >0.9) and, therefore, we only examine the results using the interviewer collected data. Before analyzing infant outcomes, we checked for confounding variables and found no significant differences in baseline demographic characteristics within or across study conditions among the full sample or among mothers depressed antenatally. Using SAS PROC MIXED (version 9°2; SAS Institute Inc., Cary, North Carolina, USA), we modelled infant growth z-scores longitudinally using an unstructured time trend in hierarchical linear regressions that included two random effects: one to model the correlation of repeated measures within a participant and another to model the correlation between participants clustered within the same neighborhood. Explanatory variables included indicators for intervention (PIP; 1 = intervention, 0 = SC) and antenatal depression (DEP; 1 = depressed, 0 = not depressed), time (TIME; 0 = birth, 1 = post-birth, 2 = six months), and the two- and three-way interactions of PIP, DEP, and TIME. Intervention as a moderator of the depression effect on the change in z-score over time (PIP*DEP*TIME) was the effect of interest, and we considered a 2-sided p-value < 0.05 to be significant. Using SAS PROC GLIMMIX, we also modelled the moderating effect of intervention on infant growth (z-score ≥ −2) cross-sectionally at each assessment period using logistic random effects regressions controlling for neighbourhood clustering. Explanatory variables included indicators for intervention (PIP; 1 = intervention, 0 = SC), antenatal depression (DEP; 1 = depressed, 0 = not depressed, and their interaction. Intervention’s moderation of depression’s impact on infant growth (PIP*DEP) was the effect of interest, and we set the criteria for a 2-sided p-value <0.05 to be significant.