Statistics

Modeling Time To Neonatal Mortality at Wallaga University Comprehensive Specialized Hospital: Application of Bayesian Survival Model with INLA

2025-07-06T00:00:00Z

Modeling Time To Neonatal Mortality at Wallaga University Comprehensive Specialized Hospital: Application of Bayesian Survival Model with INLA Wakgari Garba; Geremew Muleta; Abebe Nega Background: The neonatal period is the most vulnerable time for survival in which children face the highest risk of dying in their lives. Although, better progress has been made in reducing Neonatal Mortality before 2016, Ethiopia is currently one of the top ten countries affected by NM. Therefore, this study aims to utilize Bayesian Survival Models to analyze and model the time to neonatal mortality at WUCSH. Methods: A retrospective study of was conducted among 343 neonates admitted to WUCSHfromJanuary 1, 2022 to December 30, 2023. A Bayesian survival model with INLA was used to identify the risk factors associated with time to neonatal mortality. Results: Among the 343 neonates admitted to WUCSH, 187 (54.52%) were male, and more male were died, that means 61 (17.78%). The variable residence (p = 0.0220), gestational age (p = 0.0355), neonate age (p = 0.0048), and the global test (p = 0.0042) in multivariate Cox-PH were shows a statistically significant violation of the proportional hazards assumption. In the Bayesian Log-logistic AFT model, rural residence with AFT factor ˆγ= 0.573 (-0.975,-0.137), had significantly shorter survival time. Conversely, being married was associated with longer neonatal survival with AFTfactor ˆγ = 1.817(0.078, 1.117). Conclusion: In conclusion, the findings of this study shows that residence, neonate sex, gestational age, marital status, age of neonate and birth weight are the most determinant and statistically associated with time to neonatal mortality. It is therefore awareness should be raised about the burden of these risk factors contributing to neonatal mortality

Determinants of Fertility in Ethiopia: A Count Regression Approach

2025-08-22T00:00:00Z

Determinants of Fertility in Ethiopia: A Count Regression Approach Yakob Tafesse; Abiyot Negash; Samuel Fikadu Background: One of the key demographic factors influencing acountry’s population growth is fertility. High fertility rates present persistent challenges to Ethiopia’s population growth management and development goals. Objective: The main objective of the study is to identify the determinants of fertility in Ethiopia using the EMDHS 2019. Methods: The survey collected information from a total of 9,012 women aged 15-49 years out of which 8885 women were considered in this study. From several Count regression models namely; Poisson, NB, ZIP, ZINB, HP and HNB was selected using model comparison criteria like Akaike Information Criteria and Bayesian Information Criteria. Results: Descriptive statistics reveal that 35.93% of women in the study have never given birth, with a mean fertility rate of 2.53 children per woman and a maximum of 15 births. The pattern of fertility level did not vary across the different region of Ethiopia. From several Count regression models, the ZIP regression model was found to be the most appropriate and preferred, with an AIC value of 24,290.37 and a BIC value of 25,087.53 for fitting the fertility data. The results of ZIP regression model revealed that the variables such as family size (OR= 1.1; 95% CI: 1.095, 1.107), Amhara region (OR= 2.431; 95% CI: 1.078, 5.482), Gambela region (OR= 0.169; 95% CI: 0.072, 0.399), Addis Abeba region (OR= 2.32; 95% CI: 1.005, 5.538) and Dire Dawa region (OR= 2.401; 95% CI: 1.031, 5.591), mother’s education in secondary level (OR= 1.645; 95% CI: 1.014, 2.667), higher educational level (OR= 3.569; 95% CI: 1.970, 6.465), medium wealth index (OR= 1.757; 95% CI: 1.146, 2.694), age of household head (OR= 0.762; 95% CI: 0.725, 0.8.2), and mother’s marital status of women’s category other (OR= 57.314; 95% CI: 40.437, 81.235) were all found to be statistically significant at the 5% level of significance in fertility level Conclusion: In this study, the highest fertility level was observed in Somali regions, with no variation across Ethiopia. Based on different model comparison techniques, ZIP regression model was found to be the most appropriate to fit the fertility level data. Key determinants of fertility included family size, region, education, wealth index, and marital status.

Spatial Patterns and Determinants of Under-Five Mortality in East Africa: ASpatial and Multilevel Levels Analysis

2023-11-27T00:00:00Z

Spatial Patterns and Determinants of Under-Five Mortality in East Africa: ASpatial and Multilevel Levels Analysis Chaltu Diyesa; Akalu Banbate; Demeke Kifle Background: Sub-Saharan Africa is still the region having the highest burden of under-five mortality rate in the world. Of 5 million under-five deaths in 2021, more than 80 percent of under-five death reported from Sub-Saharan Africa and Southern Asia. Thus, the aim of this study was to assess the spatial patterns and identify determinant factors of under-five mortality among selected East Africa countries. Methods: A study was conducted using data from the Demographic and Health Survey (DHS) of 9 East African countries between 2016 and 2022. The study included a weighted sample of 115,335 live births within the 5 years prior to the survey for analysis. Due to the hierarchical nature of the DHS data, multilevel logistic regression models were used and model comparison was done using Likelihood ratio, AIC, and BIC. Bivariate analysis identified variables with p<0.2, which were considered for multivariable analysis. Among several appropriate models, the one that considered individual and community-level factors was found to be the most suitable for analyzing the data in this study.The sig nificant predictors of under-five mortality were determined in the multilevel logistic regression analysis, and this was reported using the adjusted odds ratio (AOR) along with a 95% Confidence Interval (CI) and with a p-value <0.05. XY coordinate data was also taken from the selected enumeration areas. A total of 6,736 clusters were included in this study to analyze the spatial patterns of under-five mortality. Results: This study revealed that the spatial distribution of under-five mortality was non-random in the region with Moran’s index 0.552 (P-value<0.001), with high-risk areas identified in Burundi, Madagascar, Uganda, and Ethiopia. Multiple births (AOR = 5.73, 95% CI: 5.01, 6.36), being male child (AOR = 1.29, 95% CI: 1.22, 1.37), born from mother who had no formal education(AOR = 1.57, 95% CI: 1.28,1.91), while children born to mothers with primary education (AOR = 1.17, 95% CI: 1.05-1.31) and children born into family sizes of more than four (AOR = 2.60, 95% CI: 2.41, 2.81) were significantly associated with an increased odds of under-5 mortality. Whereas being ≥ 5th birth order (AOR=0.72, 95% CI:0.64-0.82), health facility delivery (AOR = 0.83, 95% CI: 0.77, 0.89), born from mother who use contraceptive method(AOR=0.59,95% CI:0.5,0.71), children born from employed mother (AOR=0.90, 95% 0.84 , 0.97), and children delivered by Cesarean section (AOR =0.81, 95% CI: 0.72, 0.91) were significantly associated with a lower risk of under-5 mortality. Conclusion: Under-five mortality in East Africa is spatially clustered, with high-risk areas identified in Burundi, Madagascar, Uganda, and Ethiopia. Effective interventions are possible, as evidenced by the lower risk levels in Kenya and Rwanda. Findings suggest that family planning, increased access to education, and safe delivery practices are essential for reducing under-five mortality rates

Bayesian Method to Examine the Factors for Birth Defects in Children Admitted to CURE Ethiopia Children’s Hospital: Application of a Bivariate Multinomial Regression model

2023-11-18T00:00:00Z

Bayesian Method to Examine the Factors for Birth Defects in Children Admitted to CURE Ethiopia Children’s Hospital: Application of a Bivariate Multinomial Regression model Abenezer Yohannes; Tadele Akeba Diriba; Tokuma Wayessa Background: Cleft Lip (CL) and Cleft Palate (CP) are holes or splits inside the upper lip and roof of the mouth (palate). These are the most prevalent congenital abnormalities of the head and neck and mainly occur when the facial structure of a growing child does not fully close. This study aimed to examine the determinants of cleft lip and cleft palate in children admitted at CURE Ethiopia Children’s Hospital. Methods: In this study, the data were collected from 544 children and related fami lies cleft lip and palate patients using the cross-sectional study design. A bivariate multinomial regression model was employed to examine the determinants of cleft lip and cleft palate. Parameter estimation and inference association to the model were performed based on the Bayesian method. Results: Among the study participants, unilateral, bilateral, and median cleft lip types of birth defects were observed on 360 (66.17%), 118 (21.69%) and 66 (12.13%) children, respectively. Also, 143(26.28 %) , 83 (15.25%) and 318 (58.45%) children were with muscular/soft, bony/hard and both (soft and hard) parts of the cleft palate, respectively. The results suggests that mothers consumption of alcohol for unilateral (OR =1.127; 95% CI: 1.066, 5.546) was significant predictors for cleft lip. On the other hand, mothers consumption of alcohol for bony/hard part (OR = 1.701; 95% CI: 1.286, 5.680) was significant predictors for cleft Palate at 5% level of confidence. The choice of prior distributions significantly impact the posterior distributions. Conclusion: In this study, alcohol consumption during pregnancy, maternal folic acid deficiency, history of birth defects, poor antenatal care service, and other variables were the risk factors for the prevalence of cleft lip and cleft palate in children. There fore, in order to decrease the occurrence of these defects, special attention should be paid to specific population groups by raising awareness of the risk factors.