|Home | About | Journals | Submit | Contact Us | Français|
Infants of women with lower education levels are at higher risk for perinatal mortality.
We explored the impact of training birth attendants and pregnant women in the Essential Newborn Care (ENC) program on fresh stillbirths (FSB) and early (7-day) neonatal deaths (END) by maternal education level in developing countries.
A train-the-trainer model was used with local instructors in rural communities in six countries (Argentina, Democratic Republic of Congo, Guatemala, India, Pakistan, and Zambia). Data were collected using a pre-post active baseline controlled study design.
A total of 57,643 infant/mothers were enrolled. The follow up rate at 7 days of age was 99.2%. The risk for FSB and END was higher for mothers with 0–7th grade education vs. ≥8th grade during both the pre-ENC and post-ENC periods in unadjusted models and in models adjusted for confounding. The effect of ENC differed as a function of maternal education for FSB (interaction p=0.041) without evidence that the effect of ENC differed as a function of maternal education for END. The model-based estimate of FSB risk was reduced among mothers with 0–7 years of education (19.7/1000 live births pre-ENC, CI 16.3, 23.0 vs. 12.2/1000 live births post-ENC, CI 16.3, 23.0, p < 0.001) but was not significantly different for mother with ≥8th grade, respectively.
A low level of maternal education was associated with higher risk for FSB and END. ENC training was more effective in reducing FSB among low education mothers.
An estimated 6.3 million deaths occur annually worldwide in children aged < 5 years and of these, 44% occur during the first 4 weeks after birth, mostly during the first 7 days after birth1. Causes of early neonatal mortality (sometimes classified as fresh stillbirths) include birth asphyxia, complications of low birth weight/prematurity, and infection2. More than 50% of the under-5 and neonatal deaths occur in South Asia and Africa1. Many countries in these areas are on the priority list for the Millennium Development Goal 4 (MDG4) which aims at reducing childhood mortality by two-thirds between 1990 and 20153. In these priority countries, 50% or less of adult females are literate4 and the poverty levels are high with most living on less than USD 1.00 per day5.
Studies have suggested that maternal education is a predictor of neonatal outcomes with newborns of more educated women having better survival 6,7. In a recent study, brief antenatal education increased mothers’ understanding of basic newborn care8. Mothers retained this knowledge until the early post-partum period and during early infancy. In a previous study by us, World Health Organization Essential Newborn Care (ENC) training of college-educated midwives working in low-risk health centers in large urban settings to practice and teach mothers basic care reduced early neonatal mortality in Zambia, and the reduction was largest among newborns of mothers with low education9. These results suggest that women with lower education levels may benefit most from a basic education intervention and led to the current secondary analysis of the First Breath Study10 as the latter study was multicountry conducted in rural areas and included a diverse group of birth attendants and health care workers. We explored the role of maternal education and birth attendant training on birth outcomes. In the First Breath study, we reported that a postnatal intervention based on ENC reduced stillbirths. Because it is difficult to distinguish between fresh stillbirth and death soon after birth, many experts recommend analyzing FSB and END rates in low-resource settings. Thus we hypothesized that training of birth attendants in a simplified ENC program was associated with a larger reduction in FSBs and ENDs as well as the combination of the two denoted as perinatal mortality (PNM) among women with lower compared to higher education levels.
This study is a secondary analysis of active baseline design pre- and post-ENC part of the First Breath Study10. The study was conducted in 96 rural communities at seven sites of the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Global Network for Women’s and Children’s Health Research involving six countries (Argentina, Democratic Republic of Congo, Guatemala, India – 2 sites, Pakistan, and Zambia) between March 2005 through February 2007. Rural and geographically distinct communities with at least 300 births per year were selected. Most communities had high rates of home births assisted by traditional birth attendants. The sample size was detrmined based on enough power to identify a mortality effort10. A pre- and post-training of birth attendants in ENC was done. The elements of ENC included neonatal resuscitation, thermoregulation, early and exclusive breast feeding, kangaroo (skin-to-skin) mother care, and care of the small baby. The birth attendants taught the mothers essential newborn care concepts. The study included collection of baseline data (pre-ENC) followed by data collection after training (post-ENC).
Maternal and neonatal data were collected for all births (including FSB) with a birth weight of 1500 g or more. Fetuses and neonates with birth weight of less than 1500g were excluded from the analyses because advanced medical care for very low birth weight infants was not available in most of the participating communities. In cases in which the birth weight was not measured, infants were included if the weight was estimated by the birth attendant to be 1500 g or more.
Data edits, including inter- and intra-form consistency checks, were performed upon data entry locally and by the data coordinating center (RTI International, Research Triangle Park, NC, USA). Multivariable logistic and proportional odds regression models with generalized estimating equations adjusting for cluster were used to determine differences in maternal and neonatal characteristics between the pre-ENC training and post-ENC training for mothers with 0–7 years of education and mothers with ≥8 years of education. Logistic models were used for binary data, proportional odds models with cumulative logit were used for ordered multinomial variables, and proportional odds models with generalized logit were used for non-ordered multinomial variables. Estimates of risk and associated adjusted relative risk (RR) and 95% confidence interval (CI) from generalized estimating equation extensions of robust Poisson regression models adjusting for clustering are reported for FSB and END as a function of ENC training period, maternal education, the interaction of ENC and education, controlling for cluster prenatal care, birth attendant type, gestational age, and birth weight. The data were analyzed using SAS software version 9.2 (Cary, NC).
A total of 57,819 pregnant women (pre-ENC 22,745 and post-ENC 35,074) were screened and 57,642 (99.7%) were enrolled (pre-ENC 22,625 and post-ENC 35,017) as shown in Table 1. Most of the women had 0–7 years of education (86.3%) and were 20–35 years old (81.2%). About 90% of the women received prenatal care and 65.7% of the women delivered at home. The follow up rate at 7 days of age was 99.2% and did not vary by group.
During both the pre- and post-ENC periods, women with 0–7 years of education were older, had a lower rate of antenatal care, and delivered more often at home with traditional birth attendants rather than in facilities with physicians compared to women with ≥ 8 years of education (Table 2). During the post-ENC training period, the crude rate of FSB was lower than during the pre-ENC period among mothers with 0–7 years of education (FSB was 21.6/1000 live births pre-ENC vs. 13.5/1000 live births post-ENC training, p < 0.001) but was not significantly different for mothers with a ≥8 years of education (FSB was 6.4/1000 live births pre-ENC vs. 7.2/1000 live births post-ENC training, p=0.70). END did not differ between the two periods for mothers with 0–7 or ≥ 8 years of education. FSB and END were higher for the mothers with 0–7 vs. ≥ 8 years of education during both the pre-ENC and post-ENC periods.
We evaluated the interaction between years of maternal education and ENC with and without adjusting for potentially confounding variables (birth weight and gestational age), prenatal care, and birth attendant type. Detailed results from these models are shown in Table 3. The models without adjusting for covariates suggested that the effect of ENC differed as a function of maternal education for FSB (interaction p=0.041) but showed no evidence that the effect of ENC differed as a function of maternal education for END (p=0.47 for the interaction). The unadjusted relative risk of FSB post-ENC compared to pre-ENC was 0.62 (95% CI=[0.49, 0.78]) for mothers with a 0–7 years of education while the unadjusted relative risk of FSB post-ENC compared to pre-ENC was 1.12 (95% CI=0.70, 18.2) for mothers with a ≥ 8 years of education. The adjusted results are included in Table 3. While the p-value for the interaction is attenuated somewhat for the adjusted models, the magnitudes of the adjusted relative risk estimates are quite comparable, suggesting that this effect is not mediated through either birth attendant or prenatal care.
Because the model for END showed no evidence of an interaction between education and ENC, we fit reduced models that included main effects for ENC and maternal education both with and without adjustment for the potential confounders (Table 3). The unadjusted models for END did not show an effect of ENC (RR=0.97, 95% CI 0.81, 1.18, p=07.9) but showed a strong effect of maternal education averaged across both periods (RR 1.33, 95% CI 1.10, 1.66, p=0.0038). The change in the magnitude of the relative risk for both ENC and maternal education was minimal after adjustment for the potential confounders, prenatal care, and birth attendant, either individually or in combination.
The current analysis shows that the impact of reduced FSB rates was larger among mothers who had little or no formal education (0–7 years) than among those with more education. In spite of the reduction in FSB rates following ENC training, the risk for FSB remained significantly higher among those with the lower education level compared to those with higher education level. The reason for the effect on FSB rather than on END may be due to the common misclassification in low-resource settings of identifying deaths soon after birth as FSB rather than END10. Thus, the reductions of FSBs can be in large part interpreted as a reduction of deaths at birth. It is likely that access to better trained birth attendants for the less educated mothers led to the reduction in stillbirths.
The strengths of this study include the use of a large multi-country community-based data with high (99%) completeness of outcome information assessed prospectively by birth attendants trained in distinction between causes and timing on perinatal deaths. Even though demographic and medical data were collected, data on many socioeconomic variables which may be associated with adverse perinatal outcomes were not collected. Nonetheless, maternal education is a pragmatic proxy of adverse perinatal outcomes that may be used to identify women and settings in which perinatal education programs may be most effective.
Maternal education has been observed to be inversely associated with perinatal, neonatal, and/or infant mortality in both low- and middle- 6,7,8,11–14 as well as high-resource settings 15. In a large study of data in Indonesia spanning over a decade, a lower neonatal mortality was observed in mother with higher education levels14. In the United States, infant mortality declines were found to be associated with higher levels of maternal education rather than to birth weight16 which is one of the best predictors on infant mortality. In Nepal, mortality at 5 years was 93 per 1000 live births among children born to mothers with no education and more than seven times higher than that of children born to a mother with a school certificate (13 per1000 live birth)17. Even high under-five mortality has been associated in part with mothers having no or only primary education18. Improved maternal education has been associated with reduced perinatal and neonatal mortality 19,20. Mothers with more advanced education are more likely to use good neonatal practices such as clean cord care and early breastfeeding 21. Limited education may be a marker of poor health seeking behaviour and lifestyle as well as social disadvantage 22, 23, 24. A brief maternal education program increased mothers’ basic newborn care knowledge8. The current study contributes new information as brief essential newborn care training program directed at birth attendants and mothers resulted in a large reduction in adverse perinatal outcomes even after adjustment for baseline risks. In a more recent cohort of the Global Network, maternal mortality was higher in mothers with less education25.
Because women in many rural communities have limited or no access to health care facilities but rely on home birth with midwives and traditional birth attendants, efforts have been made to improve outcomes through community-based interventions. In a study in Iraq, having the delivery supervised by a traditional birth attendant was associated with young maternal age, low education and being poor 18. In a cluster randomised controlled trial of training traditional birth attendants in the Safe Motherhood modules, there was a 30% reduction in perinatal mortality 23. In the First Breath study on which the current study is based, ENC training reduced stillbirth rates in community-based deliveries 10. A simplified educational program for birth attendants and mothers based on ENC reduced neonatal mortality in a large study in first level facilities in Zambia26 but the benefits were more pronounced in the infants of mothers with less education9. Simplified ENC programs (Essential Care for Every Baby and Essential Care for Small Babies) http://www.helpingbabiesbreathe.org/ (accessed May 18, 2016) programs27 which also include teaching the mothers are being implemented worldwide.
Although community enhancement is one approach, improving mothers’ level of education and socio-economic status is fundamental to long-term improvements in perinatal and childhood survival 24. Expansion of maternal health services utilization may be accelerated with primary maternal education 28. Efforts to reduce maternal and newborn health must go hand in hand with improving women’s education. Training in ENC is very cost effective so it should be provided even in low-resource settings29. It has been estimated that training in neonatal care may save between 244,000 to 1 million lives per year30.
In summary, low levels of maternal education is a risk for increased perinatal mortality. Maternal education was significantly associated with higher risk for FSB and END. ENC training of birth attendants reduced the FSB risk among mothers with 0–7 years of education but not in mothers with higher education. Focusing perinatal education training programs on birth attendants who deliver women with low levels of education may have a large impact on perinatal outcomes and increase equity.
Dr. Carlo is on the board of directors of Mednax. The authors have no other conflict of interest to report.