A multilevel analysis of the predictors of health facility delivery in Ghana: Evidence from the 2014 Demographic and Health Survey
Open Access
- 13 March 2024
- journal article
- research article
- Published by Public Library of Science (PLoS) in PLOS Global Public Health
- Vol. 4 (3) , e0001254
Abstract
Health facility delivery has the potential to improve birth and general health outcomes for both newborns and mothers. Regrettably, not all mothers, especially in low-and-middle income countries like Ghana deliver at health facilities, and mostly under unhygienic conditions. Using data from the 2014 Ghana Demographic and Health Survey, we fitted both weighted single-level and random intercept multilevel binary logistic regression models to analyse predictors of a health facility delivery among mothers aged 15–49 years and to quantify unobserved household and community differences in the likelihood of health facility delivery. We analysed data on 4202 mothers residing in 3936 households and 427 communities. Of the 4202 mothers who delivered, 3031 (75.3%—weighted and 72.1%—unweighted) delivered at the health facility. Substantial unobserved household only (Median Odds Ratio (MOR) = 5.1) and household conditional on community (MOR = 4.7) level differences in the likelihood of health facility delivery were found. Mothers aged 25–34 (aOR = 1.4, 95%CI: 1.0–2.1) and 35–44 (aOR = 2.9, 95%CI: 1.7–4.8), mothers with at least a secondary education (aOR = 2.7, 95%CI: 1.7–4.1), with health insurance coverage (aOR = 1.6, 95%CI: 1.2–2.2) and from richer/richest households (aOR = 8.3, 95%CI: 3.6–19.1) and with piped water (aOR = 1.5, 95%CI: 1.1–2.1) had increased odds of health facility delivery. Mothers residing in rural areas (aOR = 0.3, 95%CI: 0.2–0.5) and with no religion (aOR = 0.5, 95%CI: 0.3–1.0) and traditional religion (aOR = 0.2, 95%CI: 0.1–0.6), who reported not wanting to go to health facilities alone as a big problem (aOR = 0.5, 95%CI: 0.3–0.8) and having a parity of 2 (aOR = 0.4, 95%CI: 0.3–0.7), 3 (aOR = 0.3, 95%CI: 0.2–0.6) and ≥4 (aOR = 0.3, 95%CI: 0.1–0.5) had reduced odds of health facility delivery. Our predictive model showed outstanding predictive power of 96%. The study highlights the need for improved healthcare seeking behaviours, maternal education and household wealth, and bridge the urban-rural gaps to improve maternal and newborn health outcomes.Keywords
This publication has 41 references indexed in Scilit:
- Sustainable Development Goals and the Ongoing Process of Reducing Maternal MortalityJournal of Obstetric, Gynecologic, & Neonatal Nursing, 2017
- The impact of household wealth on child survival in GhanaJournal of Health, Population and Nutrition, 2016
- Factors Influencing Health Facility Delivery in Predominantly Rural Communities across the Three Ecological Zones in Ghana: A Cross-Sectional StudyPLOS ONE, 2016
- Understanding the determinants of under-five child mortality in Uganda including the estimation of unobserved household and community effects using both frequentist and Bayesian survival analysis approachesBMC Public Health, 2015
- Childhood Malnutrition and Its Determinants among Under‐Five Children in GhanaPaediatric and Perinatal Epidemiology, 2015
- Predictors of skilled attendance at delivery among antenatal clinic attendants in Ghana: a cross-sectional study of population dataBMJ Open, 2015
- Individual and contextual factors associated with childhood stunting in Nigeria: a multilevel analysisMaternal & Child Nutrition, 2011
- Appropriate Assessment of Neighborhood Effects on Individual Health: Integrating Random and Fixed Effects in Multilevel Logistic RegressionAmerican Journal of Epidemiology, 2005
- Interpreting Parameters in the Logistic Regression Model with Random EffectsBiometrics, 2000
- MIXOR: a computer program for mixed-effects ordinal regression analysisComputer Methods and Programs in Biomedicine, 1996