Department of Biostatistics

Permanent URI for this collectionhttps://repository.mu.edu.et/handle/123456789/157

Browse

Start date: 2024Subject: search.filters.subject.Tigray

Search Results

Now showing 1 - 4 of 4
  • Item
    Assessment of Preterm Birth and Predictive Factors at Ayder Comprehensive Specialized Hospital, Northern Ethiopia: An Ordinal Logistic Regression Analysis
    (Mekelle University, 2025-08-28) DANIEL ABATE
    Background: Preterm birth is a major public health concern due to its important impact on infant mortality and morbidity. Previous studies conducted in Ethiopia have examined the prevalence and risk factors of preterm birth, using a binary outcome of preterm birth, without considering the severity of preterm birth. Objective: -To assess Predictive Factors of Preterm Birth severity at Ayder Comprehensive Specialized Hospital, North Ethiopia from 2018 to 2020. Methods: A facility-based retrospective cross-sectional study was conducted among 2082 preterm and term neonates from February 2018 to May 2020 at Ayder Comprehensive Specialized Hospital. The minimum sample size was 538. All preterm and term neonates were included in this study. Ordinal logistic regression with partial proportional odd model (PPOM) was used to determine predictors of preterm. Parallel line assumption was tested using Brant test. Odd Ratio with 95% confidence interval was used to assess the strength of association between independent and dependent variables. Result- The overall prevalence of preterm was found to be 36.7% (95%C. I:34.67, 38.86). Being having congenital malformation the odd very preterm versus (moderate preterm, late preterm and term) increased by OR= 2.295(95% C.I :1.566,3.363) times. Being having multiple gestation the odd very preterm versus (moderate preterm, late preterm and term) increased by OR= 2.319(95% C.I:1.526,3.524) times, Being having history of preterm birth the odd very preterm versus (moderate preterm, late preterm and term) increased by OR= 10.03(95% C.I:6.803,14.788). Being having hypertension the odd having higher preterm level increased by OR= 3.835(95% C.I:2.036,7.226). Being having ANC visit the odd of having higher preterm level decreased by 91.8% OR= 0.182(95% C.I:0.039,0.841). Conclusions and Recommendations- in the PPOM, the variables congenital malformation, multiple gestation, history of preterm birth, hypertension, and malaria infection, had a positive significant association with the odd of preterm birth, whereas ANC visit had a negative significant effect. In order to decrease the probability of preterm birth, every mother should prevent chronic disease by changing life style. Health professional should provide health education, early screening of chronic disease and aware mothers to have appropriate ANC follow up during prenatal period.
  • Item
    Predicting birth asphyxia in newborns via supervised machine learning: a cross sectional study in Tigray, Ethiopia 2025
    (Mekelle University, 2025-07-22) GOITOM YEMANE
    Background: Birth asphyxia, a critical condition characterized by insufficient oxygen supply to a newborn before, during, or after birth, is the second leading cause of neonatal mortality in Ethiopia. It contributes substantially to preventable neonatal morbidity and long-term neurodevelopmental impairment. The burden is especially high in low-resource regions like Tigray, where healthcare systems have been severely impacted by conflict and limited infrastructure. Early and accurate prediction of at-risk newborns is essential, and supervised machine learning (ML) offers a powerful data-driven solution to support clinical decision-making. Objective: To predict birth asphyxia in newborns using supervised machine learning: a cross sectional study in Tigray, Ethiopia (2025). Methods: An institution-based prospective study was conducted among 1014 mothers and their newborns who delivered at four selected hospitals in Tigray (Ayder, Mekelle, Quiha, and Wukro) between February 25 and April 10, 2025. A convenience sampling technique was used to recruit eligible participants. The dataset underwent thorough preprocessing, including handling missing values, one-hot encoding, normalization, hybrid feature selection approach, and class balancing. Seven ML models—logistic regression, support vector machine, decision tree, random forest (RF), naive bayes, k-nearest neighbors, and extreme gradient boosting were trained and evaluated. The data were split into 80% for training and 20% for testing, with model performance assessed using accuracy, sensitivity, specificity, F1-score, and area under the receiver operating characteristic curve (AUC) with 95% confidence intervals. Shapley Additive Explanations was employed for model interpretability, validated across cross-validation folds. Results: Of the 1014 neonates included, 195 (19.2%) were diagnosed with birth asphyxia based on APGAR scores and physician confirmation. The random forest classifier achieved the best performance, with an AUC of 0.99 (95% CI: 0.98–1.00) and Brier score of 0.0099 (95% CI: 0.008–0.012). SHAP analysis identified fetal heart rate (38.6%), birth weight (11.2%), mal-presentation (8.1%), hypothermia (7.7%), referral status (7.5%), and prolonged labor (6.5%) are collectively contributing 79.6% to the model’s predictive capacity, consistent across folds (standard deviation of SHAP values <0.02). Conclusion: The RF model demonstrated excellent performance in predicting birth asphyxia and offered strong interpretability. Nearly 80% of the model's predictive power was explained by a small number of clinically actionable variables. These findings support the integration of interpretable machine learning tools into routine labor management to reduce birth asphyxia. Future external validation and deployment as a web-based tool are planned.
  • Item
    TIME TO END STAGE OF RENAL DISEASE ASSOCIATED WITH SERUM CREATININE AMONG ADULT CHRONIC KIDNEY PATIENTS IN AYDER COMPREHENSIVE SPECIALIZED HOSPITAL: JOINT AND COPULA MODELING
    (Mekelle University, 2024-12-20) MENGSTU SURAFEL TEKULU
    Background- Chronic kidney disease (CKD) is a global public health issue and around 13.4% of people had CKD. Out of these, 14.5 million people have end stage of renal disease (ESRD) and 7.083 million needed renal replacement therapy. Paired kidney shared the same gene and comes from the same person. In Ethiopia, the incidence of CKD is estimated to be 21.71%. The war between Ethiopia federal and Tigray government leads to 70–80% of health facility dysfunctional. Objective The main objective was to determine time to ESRD associated with longitudinal serum creatinine, joint time to right and left kidney failure among adult CKD patients, and determinant factors in Ayder Comprehensive Specialized Hospital from 2019 to 2024. Methods - A retrospective cohort study conducted among 408 adult CKD patients in Ayder Comprehensive Specialized Hospital from 2019 to 2024. CKD patients without ESRD at the start of follow-up included in the study and data collected from medical card. Pretest, supervision and training used to assess data completeness. Joint and Copula Cox-proportional hazard model (CoxPH) used. Archimedean Copula model fitted to predict joint time to fail both kidney and interpretation done using hazard ratio (HR) and 95% confidence interval (C.I). Result- In the joint Cox-PH model variables with HR and 95% C.I interpreted. A unit increased in hemoglobin level the hazard of ESRD decreased by 11 %, HR= 0.89(0.7865,0.9935). A unit increased in urea the hazard of ESRD increased by HR =1.01(1.0065,1.0135) times, as estimate glomerular filtration rate increased, hazard of ESRD decreased by 2%, HR= 0.98(0.9621,0.9978). As longitudinal serum creatinine increased, hazard of ESRD increased by HR= 4.31(3.5760,5.0641) and being hepatitis, hazard of ESRD increased by HR= 2.61(1.8936,3.3264) times. In the Copula model, 86% of time to ESRD due to dependency of right and left kidney. Conclusion and Recommendation - The variables hemoglobin level and glomerular filtration rate value had a negative association, whereas urea level, longitudinal serum creatinine value and hepatitis had a significant positive effect on time to ESRD at the 5% level of significance. These factors increase the incidence of ESRD among CKD. To decrease the progression, health professional should provide early screening of chronic disease and strictly follow for laboratorial abnormality, every patient should follow appropriately for medical service and life style modification to prevent for chronic disease.
  • Item
    PREDICTORS OF TIME TO RECOVERY FROM MODERATE ACUTE MALNUTRITION AMONG 6-59 MONTHS OLD CHILDREN IN INTERNALLY DISPLACED PERSONS SITES, NORTHWEST TIGRAY, ETHIOPIA, 2024, A RETROSPECTIVE FOLLOW UP
    (Mekelle University, 2024-10-20) EFREM SHUSHAY BERHE
    Background: Moderate acute malnutrition is one of the acute malnutrition categories diagnosed with an anthropometric measurement of weight for height [-3, -2) Z-score standard deviation or/and mid upper arm circumference (12.5-11.5] cm and the child should be with-out nutritional edema. Despite the existence of targeted supplementary feeding programs, the prevalence of moderate acute malnutrition was seriously high (15.1%) and the studies conducted previously used an out dated criteria to assess their recovery status from moderate acute malnutrition. Objective: The aim of this study was to determine the recovery time and its predictors among 6- 59 Months old children with moderate acute malnutrition enrolled to targeted supplementary feeding program in internally displaced persons sites of Northwest Tigray, Ethiopia, 2024. Methods: An institutional based retrospective cohort study was conducted among 452 children with moderate acute malnutrition selected using lottery method of simple random sampling with proportional allocation of the study participants to the selected fourteen sites. Data was collected using kobo tool box and imported to Stata version 17. Variables with p-value <0.25 at the restricted mean survival time uni-variable analysis, with 95% confidence interval were considered as important variables. The recovery status of the children was measured using mid upper arm circumference with measurement of >= 12.5 cm for two consecutive visits. Result: About 244 (53.08%) were female and 265(58.6%) with an age category of 24-59 months old. The overall restricted mean survival time was 15.16 weeks and recovery rate 68.36% with a truncation time of 16 weeks. Admission mid upper arm circumference category with restricted mean survival time difference of 5.47 (95% CI 3.53:8.48), ready to use supplementary food sharing status 2.07 (95% CI 1.39:3.08), and follow up status 0.57 (95% CI 0.42:0.76) were Significant predictors of time to recovery from moderate acute under nutrition. Conclusion: The study found an overall restricted mean survival time of 15.16 weeks and a recovery rate was below the minimum acceptable international standard. Recommendation: Strategies that enhance early detection should be implemented to get the child with moderate acute malnutrition at early stage. Tracking lost to follow-ups are critical, alongside counseling caregivers to treat food as medicine.