Machine learning algorithms to predict healthcare-seeking behaviors of mothers for acute respiratory infections and their determinants among children under five in sub-Saharan Africa.

Background: Acute respiratory infections (ARIs) are the leading cause of death in children under the age of 5 globally. Maternal healthcare-seeking behavior may help minimize mortality associated with ARIs since they make decisions about the kind and frequency of healthcare services for their children. Therefore, this study aimed to predict the absence of maternal healthcare-seeking behavior and identify its associated factors among children under the age 5 in sub-Saharan Africa (SSA) using machine learning models. Methods: The sub-Saharan African countries' demographic health survey was the source of the dataset. We used a weighted sample of 16,832 under-five children in this study. The data were processed using Python (version 3.9), and machine learning models such as extreme gradient boosting (XGB), random forest, decision tree, logistic regression, and Naïve Bayes were applied. In this study, we used evaluation metrics, including the AUC ROC curve, accuracy, precision, recall, and F-measure, to assess the performance of the predictive models. Result: In this study, a weighted sample of 16,832 under-five children was used in the final analysis. Among the proposed machine learning models, the random forest (RF) was the best-predicted model with an accuracy of 88.89%, a precision of 89.5%, an F-measure of 83%, an AUC ROC curve of 95.8%, and a recall of 77.6% in predicting the absence of mothers' healthcare-seeking behavior for ARIs. The accuracy for Naïve Bayes was the lowest (66.41%) when compared to other proposed models. No media exposure, living in rural areas, not breastfeeding, poor wealth status, home delivery, no ANC visit, no maternal education, mothers' age group of 35-49 years, and distance to health facilities were significant predictors for the absence of mothers' healthcare-seeking behaviors for ARIs. On the other hand, undernourished children with stunting, underweight, and wasting status, diarrhea, birth size, married women, being a male or female sex child, and having a maternal occupation were significantly associated with good maternal healthcare-seeking behaviors for ARIs among under-five children. Conclusion: The RF model provides greater predictive power for estimating mothers' healthcare-seeking behaviors based on ARI risk factors. Machine learning could help achieve early prediction and intervention in children with high-risk ARIs. This leads to a recommendation for policy direction to reduce child mortality due to ARIs in sub-Saharan countries.

Predicting the individualized risk of human immunodeficiency virus infection among sexually active women in Ethiopia using a nomogram: prediction model development and validation.

Introduction: Women are more vulnerable to HIV infection due to biological and socioeconomic reasons. Developing a predictive model for these vulnerable populations to estimate individualized risk for HIV infection is relevant for targeted preventive interventions. The objective of the study was to develop and validate a risk prediction model that allows easy estimations of HIV infection risk among sexually active women in Ethiopia. Methods: Data from the 2016 Ethiopian Demographic and Health Survey, which comprised 10,253 representative sexually active women, were used for model development. Variables were selected using the least absolute shrinkage and selection operator (LASSO). Variables selected by LASSO were incorporated into the multivariable mixed-effect logistic regression model. Based on the multivariable model, an easy-to-use nomogram was developed to facilitate its applicability. The performance of the nomogram was evaluated using discrimination and calibration abilities, Brier score, sensitivity, and specificity. Internal validation was carried out using the bootstrapping method. Results: The model selected seven predictors of HIV infection, namely, age, education, marital status, sex of the household head, age at first sex, multiple sexual partners during their lifetime, and residence. The nomogram had a discriminatory power of 89.7% (95% CI: 88.0, 91.5) and a calibration p-value of 0.536. In addition, the sensitivity and specificity of the nomogram were 74.1% (95% CI: 68.4, 79.2) and 80.9% (95% CI: 80.2, 81.7), respectively. The internally validated model had a discriminatory ability of 89.4% (95% CI: 87.7, 91.1) and a calibration p-value of 0.195. Sensitivity and specificity after validation were 72.9% (95% CI: 67.2, 78.2) and 80.1% (95% CI: 79.3, 80.9), respectively. Conclusion: A new prediction model that quantifies the individualized risk of HIV infection has been developed in the form of a nomogram and internally validated. It has very good discriminatory power and good calibration ability. This model can facilitate the identification of sexually active women at high risk of HIV infection for targeted preventive measures.

Spatial variation and geographical weighted regression analysis to explore open defecation practice and its determinants among households in Ethiopia.

BACKGROUND: In Ethiopia, recent evidence revealed that over a quarter (27%) of households (HHs) defecated openly in bush or fields, which play a central role as the source of many water-borne infectious diseases, including cholera. Ethiopia is not on the best track to achieve the SDG of being open-defecation-free by 2030. Therefore, this study aimed to explore the spatial variation and geographical inequalities of open defecation (OD) among HHs in Ethiopia. METHODS: This was a country-wide community-based cross-sectional study among a weighted sample of 8663 HHs in Ethiopia. The global spatial autocorrelation was explored using the global Moran's-I, and the local spatial autocorrelation was presented by Anselin Local Moran's-I to evaluate the spatial patterns of OD practice in Ethiopia. Hot spot and cold spot areas of OD were detected using ArcGIS 10.8. The most likely high and low rates of clusters with OD were explored using SaTScan 10.1. Geographical weighted regression analysis (GWR) was fitted to explore the geographically varying coefficients of factors associated with OD. RESULTS: The prevalence of OD in Ethiopia was 27.10% (95% CI: 22.85-31.79). It was clustered across enumeration areas (Global Moran's I = 0.45, Z-score = 9.88, P-value ≤ 0.001). Anselin Local Moran's I analysis showed that there was high-high clustering of OD at Tigray, Afar, Northern Amhara, Somali, and Gambela regions, while low-low clustering of OD was observed at Addis Ababa, Dire-Dawa, Harari, SNNPR, and Southwest Oromia. Hotspot areas of OD were detected in the Tigray, Afar, eastern Amhara, Gambela, and Somali regions. Tigray, Afar, northern Amhara, eastern Oromia, and Somali regions were explored as having high rates of OD. The GWR model explained 75.20% of the geographical variation of OD among HHs in Ethiopia. It revealed that as the coefficients of being rural residents, female HH heads, having no educational attainment, having no radio, and being the poorest HHs increased, the prevalence of OD also increased. CONCLUSION: The prevalence of OD in Ethiopia was higher than the pooled prevalence in sub-Saharan Africa. Tigray, Afar, northern Amhara, eastern Oromia, and Somali regions had high rates of OD. Rural residents, being female HH heads, HHs with no educational attainment, HHs with no radio, and the poorest HHs were spatially varying determinants that affected OD. Therefore, the government of Ethiopia and stakeholders need to design interventions in hot spots and high-risk clusters. The program managers should plan interventions and strategies like encouraging health extension programs, which aid in facilitating basic sanitation facilities in rural areas and the poorest HHs, including female HHs, as well as community mobilization with awareness creation, especially for those who are uneducated and who do not have radios.

Prevalence of Hepatitis B virus infection and its determinants among pregnant women in East Africa: Systematic review and Meta-analysis.

INTRODUCTION: Hepatitis B virus (HBV) is one of the major public health problems globally and needs an urgent response. It is one of the most responsible causes of mortality among the five hepatitis viruses, and it affects almost every class of individuals. Different studies were conducted on the prevalence of HBV among pregnant women in East African countries, but none of them showed the pooled prevalence of HBV among the pregnant women. Thus, the main objective of this study was to determine the pooled prevalence and its determinants among pregnant women in East Africa. METHODS: We searched studies using PubMed, Scopus, Embase, ScienceDirect, Google Scholar and grey literature that were published between January 01/2020 to January 30/2024. The studies were assessed using the Newcastle Ottawa Scale (NOS) quality assessment scale. The random-effect (DerSimonian) model was used to determine the pooled prevalence and associated factors of HBV among pregnant women. Heterogeneity were assessed by I2 statistic, sub-group analysis, and sensitivity analysis. Publication bias was assessed by Egger test, and the analysis was done using STATA version 17. RESULT: A total of 45 studies with 35639 pregnant women were included in this systematic review and meta-analysis. The overall pooled prevalence of HBV among pregnant women in East Africa was 6.0% (95% CI: 6.0%-7.0%, I2 = 89.7%). The highest prevalence of 8% ((95% CI: 6%, 10%), I2 = 91.08%) was seen in 2021, and the lowest prevalence 5% ((95% CI: 4%, 6%) I2 = 52.52%) was observed in 2022. A pooled meta-analysis showed that history of surgical procedure (OR = 2.14 (95% CI: 1.27, 3.61)), having multiple sexual partners (OR = 3.87 (95% CI: 2.52, 5.95), history of body tattooing (OR = 2.55 (95% CI: 1.62, 4.01)), history of tooth extraction (OR = 2.09 (95% CI: 1.29, 3.39)), abortion history(OR = 2.20(95% CI: 1.38, 3.50)), history of sharing sharp material (OR = 1.88 (95% CI: 1.07, 3.31)), blood transfusion (OR = 2.41 (95% CI: 1.62, 3.57)), family history of HBV (OR = 4.87 (95% CI: 2.95, 8.05)) and history needle injury (OR = 2.62 (95% CI: 1.20, 5.72)) were significant risk factors associated with HBV infection among pregnant women. CONCLUSIONS: The pooled prevalence of HBV infection among pregnant women in East Africa was an intermediate level and different across countries ranging from 1.5% to 22.2%. The result of this pooled prevalence was an indication of the need for screening, prevention, and control of HBV infection among pregnant women in the region. Therefore, early identification of risk factors, awareness creation on the mode of transmission HBV and implementation of preventive measures are essential in reducing the burden of HBV infection among pregnant women.

Geospatial variation of exclusive breastfeeding and its determinants among mothers of infants under 6 months in Ethiopia: spatial and geographical weighted regression analysis.

BACKGROUND: Exclusive breastfeeding (EBF) is a major public health problem in Ethiopia. However, the spatial variation of EBF and the associated factors have not been studied as much as we have searched. This study aimed at assessing geospatial variation and the predictors of EBF using geographically weighted regression. METHODS: A cross-sectional study was conducted using the 2019 Mini-Ethiopian Demographic and Health Survey data set. The study used a total weighted sample of 548 infants. Hotspot spatial analysis showed the hotspot and cold spot areas of EBF. The spatial distribution of EBF was interpolated for the target population using spatial interpolation analysis. SaTScan V.9.6 software was used to detect significant clusters. Ordinary least squares regression analysis identified significant spatial predictors. In geographically weighted regression analysis, the effect of predictor variables on the spatial variation of EBF was detected using local coefficients. RESULTS: The weighted prevalence of EBF in Ethiopia was 58.97% (95% CI 52.67% to 64.99%), and its spatial distribution was found to be clustered (global Moran's I=0.56, p<0.001). Significant hotspot areas were located in Amhara, Tigray, Southern Nations, Nationalities, and Peoples' Region, and Somali regions, while significant cold spots were located in Dire Dawa, Addis Ababa and Oromia regions. Kulldorff's SaTScan V.9.6 was used to detect significant clusters of EBF using a 50% maximum cluster size per population. The geographically weighted regression model explained 35.75% of the spatial variation in EBF. The proportions of households with middle wealth index and married women were significant spatial predictors of EBF. CONCLUSION: Middle wealth index and married women were significant spatial predictors of EBF. Our detailed map of EBF hotspot areas will help policymakers and health programmers encourage the practice of EBF in hotspot areas and set national and regional programmes focused on improving EBF in cold spots by considering significant predictor variables.

Spatial variation and associated factors of inadequate counselling regarding pregnancy danger signs during antenatal care visits among pregnant women in Ethiopia: a Geographically Weighted Regression Model.

INTRODUCTION: Inadequate counselling of pregnant women regarding pregnancy danger signs contributes to a delay in deciding to seek care, which causes up to 77% of all maternal deaths in developing countries. However, its spatial variation and region-specific predictors have not been studied in Ethiopia. Hence, the current study aimed to model its predictors using geographically weighted regression analysis. METHODS: The 2019 Ethiopian Mini Demographic and Health Survey data were used. A total weighted sample of 2922 women from 283 clusters was included in the final analysis. The analysis was performed using ArcGIS Pro, STATA V.14.2 and SaTScan V.10.1 software. The spatial variation of inadequate counselling was examined using hotspot analysis. Ordinary least squares regression was used to identify factors for geographical variations. Geographically weighted regression was used to explore the spatial heterogeneity of selected variables to predict inadequate counselling. RESULTS: Significant hotspots of inadequate counselling regarding pregnancy danger signs were found in Gambella region, the border between Amhara and Afar regions, Somali region and parts of Oromia region. Antenatal care provided by health extension workers, late first antenatal care initiation and antenatal care follow-up at health centres were spatially varying predictors. The geographically weighted regression model explained about 66% of the variation in the model. CONCLUSION: Inadequate counselling service regarding pregnancy danger signs in Ethiopia varies across regions and there exists within country inequality in the service provision and utilisation. Prioritisation and extra efforts should be made by concerned actors for those underprivileged areas and communities (as shown in the maps), and health extension workers, as they are found in the study.