Tag Archives: Odds ratio

Perinatal Mortality and Associated Obstetric Risk Factors in Urban Delta State, Nigeria; Rural-Urban Differences (Published)

The Perinatal Mortality which is the number of stillbirths and early neonatal deaths per 1000 deliveries was discussed in this paper, with associated obstetric risk factors of which twenty seven risk factors were considered. Simple descriptive analysis, Independent t-tests, Time series analysis and Logistic Regression techniques was applied on a 10 year study from which 9018 deliveries resulting in 344 perinatal death were obtained. Perinatal Mortality Rate (PNMR), Still Birth Rate (SBR) and, Early Neonatal Death Rate (ENDR) were 38.15/1000, 26.17/1000, 11.98/1000 deliveries respectively. A linear trend in Perinatal Mortality given as ŷt = 91.1973 + 8.15788*t was obtained. The null hypothesis of no significant difference between Perinatal mortality in the urban region and Perinatal mortality in the rural region of Delta State was rejected at 5% alpha level (t(18) = -4.336, p-value = 0.000). Ante partum Hemorrhage, Hypertensive Disorders, Abruptio Placentae, e.t.c. were found significantly causing 22.45%, 22.67%, 21.5%, of perinatal death cases in the state respectively. Hence these factors were considered the major risk factors associated with Perinatal Mortality in Urban Delta Sate Nigeria. Also, the logistic regression model adequately fits the data at 5% alpha level (Hosmer and Lemeshow test ; χ2 = 19.9190, p-value = 0.1463) and the significant risk factors as a group were related to the likelihood of Perinatal Mortality in Urban Delta Sate (Omnibus test of model coefficient; χ2 = 566.271, p-value = 0.000). For every unit increase in Ante Partum Hemorrhage, the odds of perinatal death occurring increases by 3.001 when all other variables are controlled

Keywords: Early Neonatal Deaths, Obstetric Risks, Odds ratio, Perinatal Mortality, Still Births

Application of Logistic Regression Model to Admission Decision of Foundation Programme at University Of Lagos. (Published)

This paper considers the application of logistic regression model to admission process from foundation to degree of university of Lagos. The choice of this model becomes imperative as a result of dichotomous relationship existing in the model (either recommended for the admission or not). 395 students of foundation programme was selected from faculty of business administration. Statistical package for social scientist (SPSS) was used for the analysis. The results show that type of secondary school attended, mode of school fees payment at first registration, sponsor and first semester grade point average contribute significantly to the chance of gaining admission to the degree programme of the institution. The fitness of the model was assessed using Hosmer and Lemeshow test, split-sample approach and other supplementary indices to validate the model. The fitted model indicated that fitted binary logistic regression model could be used to predict the future admission process.

Keywords: Cross-validation, Direct entry admission, Foundation programme Likelihood ratio, Logistic regression model, Odds ratio, Roc curve.