Geographical clustering and geographically weighted regression analysis of home delivery and its determinants in developing regions of Ethiopia: a spatial analysis

Emerging Themes in Epidemiology, Volume 19, No. 1, Year 2022

Interpretation

Background: Nearly three-fourths of pregnant women in Ethiopia give birth at home. However, the spatial pattern and spatial variables linked to home delivery in developing regions of Ethiopia have not yet been discovered. Thus, this study aimed to explore the geographical variation of home delivery and its determinants among women living in emerging (Afar, Somali, Gambella, and Benishangul-Gumuz) regions of Ethiopia, using geographically weighted regression analysis. Methods: Data were retrieved from the Demographic and Health Survey program’s official database (http://dhsprogram.com). In this study, a sample of 441 reproductive-age women in Ethiopia’s four emerging regions was used. Global and local statistical analyses and mapping were performed using ArcGIS version 10.6. A Bernoulli model was applied to analyze the purely spatial cluster discovery of home delivery. GWR version 4 was used to model spatial regression analysis. Results: The prevalence of home delivery in the emerging regions of Ethiopia was 76.9% (95% CI: 72.7%, 80.6%) and the spatial distribution of home delivery was clustered with global Moran’s I = 0.245. Getis-Ord analysis detected high-home birth practice among women in western parts of the Benishangul Gumz region, the Eastern part of the Gambela region, and the Southern and Central parts of the Afar region. Non-attendance of antenatal care, living in a male-headed household, perception of distance to a health facility as a big problem, residing in a rural area, and having a husband with no education significantly influenced home delivery in geographically weighted regression analysis. Conclusions: More than three-fourths of mothers in the developing regions of Ethiopia gave birth at home, where high-risk locations have been identified and the spatial distribution has been clustered. Thus, strengthening programs targeted to improve antenatal care service utilization and women’s empowerment is important in reducing home birth practice in the study area. Besides, supporting the existing health extension programs on community-based health education through home-to-home visits is also crucial in reaching women residing in rural settings. The study was conducted in emerging regions (Afar, Somali, Gambella, and Benishangul-Gumuz regions) of Ethiopia. These regions are found mainly in lowland parts of the country and their main lifestyle depends on animal livestock and farming. The societies that exist in these areas are nomadic ethnic groups and highly moveable which are not suitable for the existing health system of the country [27, 34, 35]. As a result, these regions were not well realizing most of the health and development-related indicators compared to other developed regions of the country [36]. Besides, in these regions, maternal health care (antenatal care, skilled delivery care, postnatal care, and contraceptive) utilizations are influenced by socio-cultural and religious barriers [27, 30, 37, 38]. The data for this analysis were retrieved from the Demographic and Health Survey (DHS) program’s official database website (http://dhsprogram.com), which was collected from January 18, 2016, to June 27, 2016. A total of 441(weighted sample) women living in four emerging (Afar, Somali, Gambella, and Benishangul-Gumuz) regions of Ethiopia who had at least one live birth in the 5 years preceding the survey were included in this analysis [13]. The outcome variable for this study was home delivery which was dichotomized into “Yes = 1 (for women whose last childbirth occurred at home) and No = 0 (for women whose last childbirth took place at health facilities)”. The independent variables were the sex of household head, age of respondent, marital status, birth order, women’s education level, husband’s education level, wealth index, respondent’s occupation, husband’s occupation, religion, exposure to mass media, antenatal care visit, type of residence, and distance to the health facility. Sample allocation in the Ethiopian Demographic and Health Survey (EDHS) to different regions of the country as well as urban and rural areas was not proportional. Thus, this study applied sample weights to estimate proportions and frequencies to adjust disproportionate sampling and non-response. A full clarification of the weighting procedure was explained in the 2016 EDHS report [13]. The data cleaning was executed using Stata version 16.0 and MS-excel 2019. The spatial autocorrelation (Global Moran’s I) statistic was held to assess the pattern of home delivery whether it was dispersed, clustered, or randomly distributed in the study areas. Local Moran’s I measure positively correlated (High-High and Low-Low) clusters and outliers (High-Low: a higher value is surrounded primarily by lower values, and Low–High: a lower value is surrounded primarily by higher values). The detail about its statistical determination of cluster outlier is found in this literature [39, 40]. Gettis-Ord Gi* statistics were calculated to measure how spatial autocorrelation differs through the study location by computing Gi* statistics for each area. Z-score was calculated to ensure the statistical significance of clustering and the p-value was calculated. To determine the statistical significance of clustering, Gi Z-score was calculated. A positive z-score > 1.96 with significant p-values denotes hot-spot, while negative Z-score < − 1.96 with significant p-values denotes cold-spot [41, 42]. Spatial regression was done using both local and global analysis techniques [43–45]. Therefore, a first global geographical regression model was applied, and then a local geographical analysis to ensure the variability of coefficients across each cluster [46–48]. Then, the six assumptions recommended for spatial regression were checked with the respective tests [49, 50]. Koenker Bp test was also executed to check whether the model underwent fitted geographically weighted regression (GWR) or not. GWR was executed using GWR version 4. Variables with a p-value less than 0.05 were selected as the determinants of home delivery and described based on their coefficients. The data access was obtained from the Demographic and Health Survey (DHS) website (http://www.measuredhs.com) after getting registered and permission was got. The retrieved data were used for this registered research only. The data were treated as confidential and no determination was made to identify any household or individual respondent.

AI Digest

Study Justification:

– The study aimed to explore the geographical variation of home delivery and its determinants in emerging regions of Ethiopia.
– The prevalence of home delivery in these regions was high, and the spatial pattern and variables associated with home delivery were not well understood.
– Understanding the spatial distribution and determinants of home delivery is crucial for developing targeted interventions and improving maternal health outcomes in these regions.

Highlights:

– The prevalence of home delivery in the emerging regions of Ethiopia was found to be 76.9%.
– The spatial distribution of home delivery was clustered, indicating high-risk locations.
– Factors such as non-attendance of antenatal care, living in a male-headed household, perception of distance to a health facility as a big problem, residing in a rural area, and having a husband with no education significantly influenced home delivery.
– Strengthening programs targeted at improving antenatal care service utilization, women’s empowerment, and community-based health education through home-to-home visits are important recommendations to reduce home birth practice in the study area.

Recommendations:

– Strengthen programs to improve antenatal care service utilization in the emerging regions of Ethiopia.
– Implement interventions to empower women and address socio-cultural and religious barriers to maternal health care.
– Support existing health extension programs on community-based health education through home-to-home visits, particularly in rural settings.

Key Role Players:

– Ministry of Health: Responsible for developing and implementing maternal health programs and policies.
– Regional Health Bureaus: Responsible for coordinating and implementing health interventions at the regional level.
– Non-Governmental Organizations (NGOs): Involved in providing support and resources for maternal health programs.
– Community Health Workers: Play a crucial role in delivering health education and services at the community level.
– Local Leaders and Traditional Birth Attendants: Engage in community mobilization and can contribute to improving maternal health outcomes.

Cost Items for Planning Recommendations:

– Training and capacity building for healthcare providers and community health workers.
– Development and dissemination of educational materials and resources.
– Operational costs for community-based health education programs.
– Monitoring and evaluation of program implementation.
– Research and data collection to monitor progress and inform future interventions.

Strength of Evidence

The strength of evidence for this abstract is 7 out of 10.
The evidence in the abstract is moderately strong. The study provides a clear description of the methodology and data sources used. The use of geographically weighted regression analysis adds to the strength of the study. However, the sample size of 441 women may be considered relatively small, and the study only focuses on four specific regions in Ethiopia. To improve the evidence, increasing the sample size and including a broader range of regions in Ethiopia would enhance the generalizability of the findings.

Abstract

The study was conducted in emerging regions (Afar, Somali, Gambella, and Benishangul-Gumuz regions) of Ethiopia. These regions are found mainly in lowland parts of the country and their main lifestyle depends on animal livestock and farming. The societies that exist in these areas are nomadic ethnic groups and highly moveable which are not suitable for the existing health system of the country [27, 34, 35]. As a result, these regions were not well realizing most of the health and development-related indicators compared to other developed regions of the country [36]. Besides, in these regions, maternal health care (antenatal care, skilled delivery care, postnatal care, and contraceptive) utilizations are influenced by socio-cultural and religious barriers [27, 30, 37, 38]. The data for this analysis were retrieved from the Demographic and Health Survey (DHS) program’s official database website (http://dhsprogram.com), which was collected from January 18, 2016, to June 27, 2016. A total of 441(weighted sample) women living in four emerging (Afar, Somali, Gambella, and Benishangul-Gumuz) regions of Ethiopia who had at least one live birth in the 5 years preceding the survey were included in this analysis [13]. The outcome variable for this study was home delivery which was dichotomized into “Yes = 1 (for women whose last childbirth occurred at home) and No = 0 (for women whose last childbirth took place at health facilities)”. The independent variables were the sex of household head, age of respondent, marital status, birth order, women’s education level, husband’s education level, wealth index, respondent’s occupation, husband’s occupation, religion, exposure to mass media, antenatal care visit, type of residence, and distance to the health facility. Sample allocation in the Ethiopian Demographic and Health Survey (EDHS) to different regions of the country as well as urban and rural areas was not proportional. Thus, this study applied sample weights to estimate proportions and frequencies to adjust disproportionate sampling and non-response. A full clarification of the weighting procedure was explained in the 2016 EDHS report [13]. The data cleaning was executed using Stata version 16.0 and MS-excel 2019. The spatial autocorrelation (Global Moran’s I) statistic was held to assess the pattern of home delivery whether it was dispersed, clustered, or randomly distributed in the study areas. Local Moran’s I measure positively correlated (High-High and Low-Low) clusters and outliers (High-Low: a higher value is surrounded primarily by lower values, and Low–High: a lower value is surrounded primarily by higher values). The detail about its statistical determination of cluster outlier is found in this literature [39, 40]. Gettis-Ord Gi* statistics were calculated to measure how spatial autocorrelation differs through the study location by computing Gi* statistics for each area. Z-score was calculated to ensure the statistical significance of clustering and the p-value was calculated. To determine the statistical significance of clustering, Gi Z-score was calculated. A positive z-score > 1.96 with significant p-values denotes hot-spot, while negative Z-score < − 1.96 with significant p-values denotes cold-spot [41, 42]. Spatial regression was done using both local and global analysis techniques [43–45]. Therefore, a first global geographical regression model was applied, and then a local geographical analysis to ensure the variability of coefficients across each cluster [46–48]. Then, the six assumptions recommended for spatial regression were checked with the respective tests [49, 50]. Koenker Bp test was also executed to check whether the model underwent fitted geographically weighted regression (GWR) or not. GWR was executed using GWR version 4. Variables with a p-value less than 0.05 were selected as the determinants of home delivery and described based on their coefficients. The data access was obtained from the Demographic and Health Survey (DHS) website (http://www.measuredhs.com) after getting registered and permission was got. The retrieved data were used for this registered research only. The data were treated as confidential and no determination was made to identify any household or individual respondent.

Materials

N/A

Innovations

Based on the information provided, here are some potential innovations that could improve access to maternal health:

1. Mobile Health (mHealth) Solutions: Develop and implement mobile health applications that provide pregnant women in remote areas with access to information, resources, and support related to maternal health. These applications can provide guidance on antenatal care, nutrition, and safe delivery practices, as well as connect women with healthcare providers through telemedicine.

2. Community-Based Health Education: Strengthen existing health extension programs and promote community-based health education through home-to-home visits. Trained healthcare workers can visit pregnant women in rural settings, providing them with essential information on maternal health, encouraging antenatal care visits, and addressing any concerns or misconceptions.

3. Transportation Solutions: Improve transportation infrastructure and services in remote areas to ensure that pregnant women have access to healthcare facilities. This could involve establishing transportation networks, providing ambulances or other means of transportation, and addressing geographical barriers that hinder access to maternal health services.

4. Empowerment Programs: Implement programs that empower women in emerging regions of Ethiopia, such as promoting education and vocational training opportunities. By empowering women and increasing their decision-making power, they are more likely to seek and utilize maternal health services.

5. Public-Private Partnerships: Foster collaborations between the public and private sectors to improve access to maternal health services. This could involve partnering with private healthcare providers to establish clinics or mobile health units in underserved areas, or engaging private companies to provide transportation solutions for pregnant women.

6. Targeted Antenatal Care Interventions: Develop targeted interventions to address specific barriers to antenatal care utilization in emerging regions. This could include addressing cultural and religious beliefs, providing incentives for antenatal care visits, and ensuring that healthcare facilities are culturally sensitive and respectful of women’s needs.

7. Data-Driven Approaches: Utilize data analysis and spatial mapping techniques, similar to the study mentioned, to identify high-risk areas and determine the factors influencing home delivery. This information can guide the development and implementation of targeted interventions to improve access to maternal health services in these areas.

It is important to note that these recommendations are based on the information provided and may need to be further tailored and evaluated in the specific context of emerging regions in Ethiopia.

AI Innovations Description

The study titled “Geographical clustering and geographically weighted regression analysis of home delivery and its determinants in developing regions of Ethiopia: a spatial analysis” provides valuable insights into improving access to maternal health in emerging regions of Ethiopia. The study aimed to explore the geographical variation of home delivery and its determinants in these regions using geographically weighted regression analysis.

The key findings of the study are as follows:

1. Prevalence of home delivery: The study found that more than three-fourths (76.9%) of mothers in the emerging regions of Ethiopia gave birth at home.

2. Spatial distribution: The spatial distribution of home delivery was clustered, indicating certain high-risk locations. These locations were identified as the western parts of the Benishangul Gumz region, the eastern part of the Gambela region, and the southern and central parts of the Afar region.

3. Determinants of home delivery: Several factors were found to significantly influence home delivery in the geographically weighted regression analysis. These factors include non-attendance of antenatal care, living in a male-headed household, perception of distance to a health facility as a big problem, residing in a rural area, and having a husband with no education.

Based on these findings, the study suggests the following recommendations to improve access to maternal health in the emerging regions of Ethiopia:

1. Strengthen antenatal care services: Programs targeted at improving antenatal care service utilization should be strengthened. This can include increasing awareness about the importance of antenatal care, providing accessible and quality antenatal care services, and addressing barriers that prevent women from attending antenatal care visits.

2. Women’s empowerment: Programs aimed at empowering women should be supported. This can include initiatives that promote education for women, enhance their decision-making power regarding healthcare choices, and improve their overall status in society.

3. Community-based health education: Existing health extension programs should be supported in providing community-based health education through home-to-home visits. This can help reach women residing in rural settings who may face additional barriers to accessing maternal health services.

By implementing these recommendations, it is expected that the practice of home delivery can be reduced, leading to improved access to maternal health services in the emerging regions of Ethiopia.

AI Innovations Methodology

Based on the provided description, here are some potential recommendations for improving access to maternal health in the emerging regions of Ethiopia:

1. Strengthening Antenatal Care (ANC) Services: Enhance the availability and quality of ANC services in the emerging regions. This can include increasing the number of health facilities offering ANC, training healthcare providers in ANC best practices, and promoting community awareness about the importance of ANC.

2. Mobile Health (mHealth) Interventions: Utilize mobile technology to provide maternal health information, reminders, and support to pregnant women in the emerging regions. This can include sending SMS messages with health tips, appointment reminders, and emergency contact information.

3. Community-Based Health Education: Implement community-based health education programs that focus on raising awareness about maternal health, promoting healthy behaviors during pregnancy, and addressing cultural and religious barriers to accessing maternal health services.

4. Improving Transportation Infrastructure: Invest in improving transportation infrastructure, such as roads and transportation networks, to ensure that pregnant women in the emerging regions have easier access to healthcare facilities for delivery and emergency obstetric care.

To simulate the impact of these recommendations on improving access to maternal health, a methodology could include the following steps:

1. Data Collection: Gather data on key variables related to maternal health access, such as the number of health facilities, ANC coverage rates, home delivery rates, and transportation infrastructure in the emerging regions of Ethiopia.

2. Baseline Assessment: Analyze the current state of maternal health access in the regions, including identifying areas with high rates of home delivery and determining the factors influencing these rates.

3. Intervention Design: Develop a simulation model that incorporates the recommended interventions, taking into account factors such as the population size, geographical distribution, and existing healthcare infrastructure in the regions.

4. Data Analysis: Use the simulation model to analyze the potential impact of the interventions on improving access to maternal health. This can include estimating changes in ANC coverage rates, reduction in home delivery rates, and improvements in transportation access.

5. Sensitivity Analysis: Conduct sensitivity analysis to assess the robustness of the simulation results and identify key factors that may influence the effectiveness of the interventions.

6. Policy Recommendations: Based on the simulation results, provide evidence-based policy recommendations for implementing the interventions and improving access to maternal health in the emerging regions of Ethiopia.

It is important to note that the specific methodology for simulating the impact of these recommendations may vary depending on the available data, resources, and expertise.

Author

Dima Health

Statistics:

Citations: 1

Identifiers:

DOI: 10.1186/s12982-022-00117-8

Research Areas:

Community Interventions, Health System and Policy, Maternal Access, Maternal and Child Health, Quality of Care, Sexual and Reproductive Health, Social Determinants

Study Countries:

Ethiopia

Study Design:

Cross Sectional Study

Study Approach:

Quantitative

Participants Gender:

Male & Female