Structural and functional network characteristics and facility delivery among women in rural Ghana

Social support receipt as a predictor of mortality: A cohort study in rural South Africa

The mechanisms connecting various types of social support to mortality have been well-studied in high-income countries. However, less is known about how these relationships function in different socioeconomic contexts. We examined how four domains of social support-emotional, physical, financial, and informational-impact mortality within a sample of older adults living in a rural and resource-constrained setting. Using baseline survey and longitudinal mortality data from HAALSI, we assessed how social support affects mortality in a cohort of 5059 individuals aged 40 years or older in rural Mpumalanga, South Africa. Social support was captured as the self-reported frequency with which close social contacts offered emotional, physical, financial, and informational support to respondents, standardized across the sample to increase interpretability. We used Cox proportional hazard models to evaluate how each support type affected mortality controlling for potential confounders, and assessed potential effect-modification by age and sex. Each of the four support domains had small positive associations with mortality, ranging from a hazard ratio per standard deviation of support of 1.04 [95% CI: 0.95,1.13] for financial support to 1.09 [95% CI: 0.99,1.18] for informational support. Associations were often stronger for females and younger individuals. We find baseline social support to be positively associated with mortality in rural South Africa. Possible explanations include that insufficient social support is not a strong driver of mortality risk in this setting, or that social support does not reach some necessary threshold to buffer against mortality. Additionally, it is possible that the social support measure did not capture more relevant aspects of support, or that our social support measures captured prior morbidity that attracted support before the study began. We highlight approaches to evaluate some of these hypotheses in future research.

A scoping review of the residual barriers to skilled birth attendance in Ghana: A conceptual framework and a fish bone analysis

The achievement of the Sustainable Development Goals (SDGs) targets 3.1, 3.2 and 3.3.1 is strongly dependent on the effective utilization of skilled birth services. Despite advancements made in Skilled Birth Attendance (SBA) in Ghana, there are still instances of unassisted childbirths taking place. The aim of this study was to explore the residual barriers of SBA such as community- and health system-related factors affecting SBA in Ghana and to identify strategies for addressing them. An electronic search was done using PubMed, Popline, Science direct, BioMed Central, Scopus and Google scholar for peer reviewed articles as well as grey articles from other relevant sources, published between 200 and 2022 on community- and health system related factors influencing SBA in Ghana. Out of the 89 articles retrieved for full screening, a total of 52 peer-reviewed articles and 1 grey article were selected for the final review. The study revealed that cultural practices (community factors), low quality of service delivery due to the inappropriate behaviors, lack of competency of skilled birth attendants (SBAs) as well as the inefficient distribution of SBAs contribute to ineffective uptake of SBA (health system factors). Also, indirect costs are associated with the utilization of skilled delivery care even with the existence of 'free' delivery care policy under the national health insurance (policy factor). For Ghana to achieve the SDGs above and improve SBA, it is essential to enhance the quality of skilled delivery care by addressing the attitude and competencies of skilled birth professionals, while plans are put in place to expand and develop the Community-based Health Planning and Services (CHPS) strategy to help address the access barriers to SBA. More so, the 'free' delivery care policy should absorb all the costs associated with skilled delivery for pregnant women as it is intended for.

Predicting facility-based delivery in Zanzibar: The vulnerability of machine learning algorithms to adversarial attacks

Background

Community health worker (CHW)-led maternal health programs have contributed to increased facility-based deliveries and decreased maternal mortality in sub-Saharan Africa. The recent adoption of mobile devices in these programs provides an opportunity for real-time implementation of machine learning predictive models to identify women most at risk for home-based delivery. However, it is possible that falsified data could be entered into the model to get a specific prediction result - known as an "adversarial attack". The goal of this paper is to evaluate the algorithm's vulnerability to adversarial attacks.

Methods

The dataset used in this research is from the Uzazi Salama ("Safer Deliveries") program, which operated between 2016 and 2019 in Zanzibar. We used LASSO regularized logistic regression to develop the prediction model. We used "One-At-a-Time (OAT)" adversarial attacks across four different types of input variables: binary - access to electricity at home, categorical - previous delivery location, ordinal - educational level, and continuous - gestational age. We evaluated the percent of predicted classifications that change due to these adversarial attacks.

Results

Manipulating input variables affected prediction results. The variable with the greatest vulnerability was previous delivery location, with 55.65% of predicted classifications changing when applying adversarial attacks from previously delivered at a facility to previously delivered at home, and 37.63% of predicted classifications changing when applying adversarial attacks from previously delivered at home to previously delivered at a facility.

Conclusion

This paper investigates the vulnerability of an algorithm to predict facility-based delivery when facing adversarial attacks. By understanding the effect of adversarial attacks, programs can implement data monitoring strategies to assess for and deter these manipulations. Ensuring fidelity in algorithm deployment secures that CHWs target those women who are actually at high risk of delivering at home.

Increasing health facility childbirth in Ghana: the role of network and community norms

BACKGROUND

Reducing pregnancy-related deaths in Sub-Saharan Africa through increases in health facility births may be achieved by promoting community norms and network norms favoring health facility births. However, the process of how both norms shift attitudes and actions towards facility delivery is little studied. We examined the association of network and community norms with facility birth, following a quality improvement intervention to improve facility births in Ghana.

METHODS

A 2015 mixed methods evaluation of a Maternal and Newborn Health Referral (MNHR) project in Ghana included a cross-sectional survey of women (N = 508), aged 15-49 years; in-depth interviews (IDIs) with mothers (n = 40), husbands (n = 20) and healthcare improvement collaborative leaders (n = 8); and focus group discussions (FGDs) with mothers-in-law (n = 4) and collaborative members (n = 7). Multivariable logistic regression was used to examine the association of network and community norms with facility birth. Thematic analysis of the qualitative data was conducted to explain this relationship.

RESULTS

The network norm of perceived family approval of facility delivery (AOR: 5.54, CI: 1.65-18.57) and the community norm of perceived number of women in the community that deliver in a facility (AOR: 3.00, CI: 1.66-5.43) were independently associated with facility delivery. In qualitative IDIs and FGDs both norms were also collectively perceived as influencing facility delivery. However, network norms were more influential in women's utilization of facility-based pregnancy-related care. Healthcare improvement collaboratives were important in swaying both network and community norms toward facility-based delivery by offering pregnancy-related health information, antenatal care, and support for facility delivery.

CONCLUSION

Quality improvement initiatives impact both community and network norms. To be most impactful in advancing facility-based pregnancy-related care, these initiatives should focus on highlighting the shifting trend toward facility delivery in rural communicates and promoting support for facility delivery among women's personal networks.

Machine learning for maternal health: Predicting delivery location in a community health worker program in Zanzibar

Background

Maternal and neonatal health outcomes in low- and middle-income countries (LMICs) have improved over the last two decades. However, many pregnant women still deliver at home, which increases the health risks for both the mother and the child. Community health worker programs have been broadly employed in LMICs to connect women to antenatal care and delivery locations. More recently, employment of digital tools in maternal health programs have resulted in better care delivery and served as a routine mode of data collection. Despite the availability of rich, patient-level data within these digital tools, there has been limited utilization of this type of data to inform program delivery in LMICs.

Methods

We use program data from 38,787 women enrolled in Safer Deliveries, a community health worker program in Zanzibar, to build a generalizable prediction model that accurately predicts whether a newly enrolled pregnant woman will deliver in a health facility. We use information collected during the enrollment visit, including demographic data, health characteristics and current pregnancy information. We apply four machine learning methods: logistic regression, LASSO regularized logistic regression, random forest and an artificial neural network; and three sampling techniques to address the imbalanced data: undersampling of facility deliveries, oversampling of home deliveries and addition of synthetic home deliveries using SMOTE.

Results

Our models correctly predicted the delivery location for 68%–77% of the women in the test set, with slightly higher accuracy when predicting facility delivery versus home delivery. A random forest model with a balanced training set created using undersampling of existing facility deliveries accurately identified 74.4% of women delivering at home.

Conclusions

This model can provide a “real-time” prediction of the delivery location for new maternal health program enrollees and may enable early provision of extra support for individuals at risk of not delivering in a health facility, which has potential to improve health outcomes for both mothers and their newborns. The framework presented here is applicable in other contexts and the selection of input features can easily be adapted to match data availability and other outcomes, both within and beyond maternal health.

Maternal social support and health facility delivery in Southwest Ethiopia

BACKGROUND

Maternal mortality continues to decrease in the world but remain the most important health problems in low-income countries. Although evidence indicates that social support is an important factor influencing health facility delivery, it has not been extensively studied in Ethiopia. Therefore, this study aimed to assess the effect of maternal social support and related factors on health facility delivery in southwest Ethiopia.

METHODS

A cross-sectional survey data on 3304 women aged 15-47 years in three districts of Ethiopia, were analyzed. Using multivariable logistic regression, we assessed the association between health facility birth, social support, and socio-demography variables. Adjusted odds ratios with 95% confidence intervals were used to identify statistically significant associations at 5% alpha level.

RESULT

Overall, 46.9% of women delivered at health facility in their last pregnancy. Average travel time from closest health facility (AOR: 1.51, 95% CI 1.21 to 2.90), mean perception score of health facility use (AOR: 1.83, 95% CI 1.44 to 2.33), involvement in final decision to identify their place of childbirth (AOR: 2.12, 95% CI 1.73 to 2.58) had significantly higher odds of health facility childbirth. From social support variables, women who perceived there were family members and husband to help them during childbirth (AOR: 3.62, 95% CI 2.74 to 4.79), women who received continuous support (AOR: 1.97, 95% CI 1.20 to 3.23), women with companions for facility visits (AOR: 1.63, 95% CI 1.34 to 2.00) and women who received support from friends (AOR: 1.62, 95% CI 1.16 to 3.23) had significantly higher odds of health facility childbirth.

CONCLUSIONS

Social support was critical to enhance health facility delivery, especially if women's close ties help facility delivery. An intervention to increase facility delivery uptake should target not only the women's general social supports, but also continuous support during childbirth from close ties including family members and close friends as these are influential in place of childbirth. Also actions that increase women's healthcare decision could be effective in improving health facility delivery.

Identifying Programmatic Factors that Increase Likelihood of Health Facility Delivery: Results from a Community Health Worker Program in Zanzibar

INTRODUCTION

Community health worker (CHW) interventions have been utilized to address barriers that prevent pregnant women from delivering in health facilities in low- and middle-income countries (LMICs). The objective of this research was to assess the programmatic factors that increase the likelihood of health facility delivery within a large digital health-supported CHW program in Zanzibar, Tanzania.

METHODS

This study included 36,693 women who were enrolled in the Safer Deliveries program with a live birth between January 1, 2017 and July 31, 2019. We assessed whether long-term enrollment, recency of CHW pregnancy visit prior to delivery, and number of routine home pregnancy visits were associated with an increased likelihood of health facility delivery compared to home delivery. We used Chi-squared tests to assess bivariate relationships and performed logistic regression analyses to assess the association between each programmatic variable and health facility delivery, adjusting for relevant confounders.

RESULTS

We found that long-term enrollment was significantly associated with increased likelihood of health facility delivery, with the strongest relationship among women with a previous home delivery (OR = 1.4, 95%CI [1.0,1.7]). Among first-time mothers, two or more pregnancy visits by a CHW was positively associated with health facility delivery (OR = 1.8, 95%CI [1.2, 2.7]). Recent pregnancy visit by a CHW was positively associated with health facility delivery, but was not significant at the α = 0.05 level.

DISCUSSION

In this program, we found evidence that at least two routine home pregnancy visits, longer length of enrollment in the program, and recency of home visit to the delivery date were strategies to increase health facility delivery rates among enrolled mothers. Maternal and child health programs should undertake similar evaluations to improve program delivery.

Trends in socio-economic, sex and geographic disparities in childhood underweight in Mauritania: evidence from Multiple Indicator Cluster Surveys (2007-2015)

Background

Underweight is one of the largest contributors to child morbidity and mortality and is considered to be the largest contributor to the global burden of diseases in low-and middle-income countries. In Mauritania, where one-fifth of children are underweight, there is a dearth of evidence on socio-economic, sex and geographic disparities in childhood underweight. As a result, this study aimed at investigating the socio-economic, sex and geographic disparities in childhood underweight in Mauritania.

Methods

Using the World Health Organization's (WHO) Health Equity Assessment Toolkit (HEAT) software, data from the Mauritania Multiple Indicator Cluster Surveys (MICSs) conducted between 2007 and 2015 were analysed. Childhood underweight was disaggregated by five equity stratifiers: education, wealth, residence, region and sex. In addition, absolute and relative inequality measures, namely difference (D), population attributable risk (PAR), ratio (R) and population attributable fraction (PAF) were calculated to understand inequalities from wider perspectives. Corresponding 95% confidence intervals (CIs) were computed to measure statistical significance.

Results

Substantial absolute and relative socio-economic, sex and geographic disparities in underweight were observed from 2007 to 2015. Children from the poorest households (PAR=−12.66 [95% CI −14.15 to −11.16]), those whose mothers were uneducated (PAF=−9.11 [95% CI −13.41 to −4.81]), those whose mothers were rural residents (R=1.52 [95% CI 1.37 to 1.68]), residents of HodhCharghy (PAF=−66.51 [95% CI −79.25 to −53.76]) and males (D=4.30 [95% CI 2.09 to 6.52]) experienced a higher burden of underweight. Education-related disparities decreased from 2007 to 2015. The urban–rural gap in underweight similarly decreased over time with the different measures showing slightly different reductions. Wealth-driven disparities decreased marginally from 2011 to 2015. The sex-based and regional disparities increased, at least on average, over the 8-y intersurvey period.

Conclusions

The burden of underweight was significantly higher among children from disadvantaged subpopulations, those with uneducated and poorest/poor mothers, those living in rural areas and those living in HodhCharghy. Special nutrition intervention and efforts focused on these deprived subpopulations are required to reduce childhood morbidity and mortality associated with underweight and help achieve the Sustainable Development Goals.