Sawubona MAMA : using mHealth to improve maternal, neonatal and child health outcomes in South Africa

Abstract

BACKGROUND: At the end of the first decade of the 2000’s, South Africa had poor, and worsening, maternal and neonatal health outcomes. In parallel, the use of mobile phone technology to support health care (mHealth) increased dramatically. Maternal mHealth has potential to support maternal care; studies of its effects have been limited, but the results have been encouraging. The Mobile Alliance for Maternal Action (MAMA) project launched a maternal mHealth project in South Africa aiming to improve maternal, neonatal, and child health outcomes using informational SMS/text messages.

AIM: To assess the informative maternal health SMS intervention, MAMA SMS, in Johannesburg, South Africa by comparing maternal and infant health outcomes of users with non-users, conducting a cost-effective analysis on the intervention, reporting on feedback given during focus group discussions attended by users, and conducting a comparison study on users of MomConnect, the maternal mHealth programme that replaced MAMA SMS.

METHODS: Sub-study I was cluster observational study of mother-infant-pairs who were followed from pregnancy to one year postnatal to monitor maternal health outcomes. The intervention arm which received the SMS intervention included 87 pairs while the control arm included 90 pairs. Sub-study II was a cohort study of HIV-positive women and their infants who were followed from pregnancy until one year postnatal, to monitor maternal health and HIV-related outcomes. The intervention arm, which received the SMS intervention included 235 mother-infant-pairs, while the control arm included 586 pairs. Sub-study III was a cost-effectiveness analysis study measuring costs at the societal level, along with measured changes in health care utilisation and health outcomes measured in sub-study I. Data were entered into the Lives Saved Tool and used to forecast lives saved and DALYs averted of gradually scaling up program activities to reach 60% of pregnant women across Gauteng province within 5 years. Sub-study IV was a qualitative study based on data from four FGDs with MAMA SMS participants who had been receiving messages for at least four months. 15 females and three males participated with ages ranging from 20 to 36 years. Sub-study V used the same methods as sub-study I to investigate the MomConnect maternal mHealth intervention implemented by the National Department of Health and based on MAMA SMS. The intervention arm included 115 participants while the control arm had 37 participants.

RESULTS: Intervention participants were more likely to attend all recommended antenatal and postnatal visits including all recommended first year vaccinations (RR: 1.71, 95% CI: 1.30-2.23) and were more likely to attend at least the recommended four antenatal visits (RR: 1.59, 95% CI: 1.23-2.04) (sub-study I). HIV-positive intervention participants also attended more ANC visits (5.16 vs. 3.95, p < 0.01) and were more likely to attend at least the recommended four ANC visits (relative risk (RR): 1.41, 95% confidence interval (CI): 1.15–1.72). Birth outcomes of intervention participants improved as they had an increased chance of a normal vaginal delivery (RR: 1.10, 95% CI: 1.02–1.19) and a lower risk of delivering a low-birth weight infant (<2500 g) (RR: 0.14, 95% CI: 0.02–1.07). In the intervention group, there was a trend towards higher attendance to infant polymerase chain reaction (PCR) testing within six weeks after birth (81.3% vs. 75.4%, p = 0.06) (sub-study II). Incremental costs per DALY averted from a societal perspective ranged from $1,985 USD in the first year of implementation to $200 USD in the 5th year. At a willingness to pay threshold of $2,000 USD, the intervention had a 40% probability of being cost effective in year 1 versus 100% in years 2 through 5 (sub-study III). Focus group participant feedback regarding the health system was mixed, with some participants having positive experiences, and a number of participants sharing negative experiences such as long waiting times, understaffed clinics and poor service. They reported that the messages were timely, written clearly and seemed supportive. Most participants reported regularly sharing the messages with both friends and family (sub-study IV). MomConnect recipients showed no differences in health outcomes measured, including complete maternal health coverage, number of ANC visits attended, likelihood of low-birthweight infants, or EPI coverage compared to the control arm. The control arm had higher than expected baseline coverage for all outcomes measured (sub-study V).

CONCLUSIONS: The results from the clinical studies show an improvement in achieving complete maternal-infant continuum of care, provide evidence of a positive impact of informative maternal mHealth messaging sent to pregnant women and new mothers, particularly if the baseline or starting health outcomes are sub-optimal (sub-studies I, II and V). The cost effectiveness evaluation findings suggest that SMS-based maternal health information messages delivered to pregnant women may be a cost-effective strategy for bolstering ANC and childhood immunizations, even at very small margins of coverage increases. Primary data obtained prospectively as part of more rigorous study designs are needed to validate modelled results (sub-study III). By providing timely and relevant information to pregnant women and new parents, contextually relevant maternal mHealth interventions could play a cost-effective part in improving maternal and child health outcomes and quality of care across the globe (sub-study IV).