Non-medical determinants of caesarean deliveries using logistic regression

Published: 22 October 2019| Version 1 | DOI: 10.17632/mz3c929xwr.1
Contributors:
Daniel Abaye,
Ernest Yeboah Boateng

Description

The non-medical factors that influence expectant mothers to odecide for caesarean deliveries in Ghana were examined. Data on 395 expectant mothers across the ten regions of Ghana, located in urban, semi-rural and rural areas, and spanned a period of five years (2012 to 2016) were obtained from the Ghana Health Service. In fitting the logistic regression model, data on 355 expectant mothers (i.e. 89.9% of the data) was assigned to the analysis sample while 40 (i.e. 10.1%) was assigned to the hold-out sample. The hold-out sample together with other statistical measures of overall model fit, pseudo R^2 measures and classification accuracy were used to validate the results obtained from the analysis sample. Significance was tested at P=0.05. Determinants including, educational level of expectant mother, parity of expectant mother, baby’s birth weight, previous caesarean delivery, location of expectant mother, age of expectant mother and, period within the year of childbirth had a significant effect on caesarean delivery.

Files

Steps to reproduce

Data for this study was secondary data obtained from the Statistics and Information Department of the Ghana Health Service. The data consist of the demographic and socio-economic records of 395 expectant mothers from the ten regions of Ghana located in urban, semi-urban and rural areas. The expectant mothers had visited the antenatal and postnatal clinics during the years 2012 to 2016. The secondary data comprise of vital maternal data of interest including, educational level of expectant mother, parity of the expectant mother, birth weight of baby, insurance status of expectant mother, marital status, religion, previous caesarean delivery by expectant mother, ethnic group of expectant mother, age of expectant mother, location of expectant mother, type of health facility delivery took place, period within the year of childbirth, occupation of expectant mother and that of her partner. The Logistic Regression Model The response variable, caesarean delivery, is a binary variable, i.e., whether the child birth was through caesarean or not. Therefore, the logistic regression is a suitable technique to use because it is developed to predict a binary dependent variable as a function of the predictor variables. The logistic regression model is widely used in studies where the dependent variable is binary. The logit, in this model, is the likelihood ratio that the dependent variable, non-caesarean delivery, is one (1) as opposed to zero (0), caesarean delivery. The probability, P, of non-caesarean delivery is given by; ln[P(Y)/(1- P(Y) )]= β_0+ β_1 X_1+ β_2 X_2+β_3 X_3+β_4 X_4+ ... + β_K X_K (1) ln[(P(Non-Caesarean Delivery))/(1-P(Non-Caesarean Delivery))] = β_0+ β_1 Educational Level+ β_2 Parity+ β_3 Baby^' sBirth Weight+ β_4 Age of Expectant Mother + β_5 Loaction of Expectant Mother + β_6 Previous Caesarean Delivery + β_7 Period of Delivery (2) Where, ln[(P(Y))/(1-P(Y))] is the log odds of Caesarean Delivery (3) 〖 β〗_0,β_1,β_3,β_4,…,β_K are the regression (model) coefficients Y is the dichotomous outcome which represents caesarean delivery (whether the child birth was through caesarean or not). X_1,X_2,…,X_K are the predictor (independent) variables which are educational level of expectant mother, parity of the expectant mother, birth weight of baby, previous caesarean delivery by expectant mother, age of expectant mother, location of expectant mother, and period within the year of childbirth.