Background: Women are more liable to die during or following delivery than during pregnancy but use of both delivery services and post-partum care is low. Objective: To find out the prevalence and predictors of institutional delivery in Wolaita Sodo (Sodo) town, southern Ethiopia. Methods: A cross-sectional study was used to look at 844 women who had given birth in the previous five years in Sodo town. The study employed a multistage-sampling scheme. Codes were given for all identified women in selected kebeles (neighbourhoods) and a simple randomsampling technique was used after generating random numbers using the Statistical Package for Social Sciences (SPSS). SPSS was then used to carry out binary- and multiple logistic regressions. A 95% CI for the odds ratio was applied to judge the presence of relationships between variables. Results: The prevalence of institutional delivery-service utilisation in Sodo town was 62.2%. Husband educational status, parity, number of antenatal clinic visits, perceived quality of care and knowledge regarding pregnancy danger signs were independent predictors of utilisation of institutional delivery services. Conclusion: Institutional delivery service utilisation in Sodo town was much higher than the national figure. Findings in this study showed that promotion of antenatal care, involvement of men in maternal healthcare, provision of health education regarding the danger signs of pregnancy and improvement of service quality are recommended in order to sustain or even improve the current level of utilisation in the town.
The source population was all women of child-bearing age who had given birth during the previous five years in Sodo town. The study population included selected women of reproductive age group who had given birth at least once during the previous five years. The number of study participants was estimated by applying a single population proportion formula with the following assumptions: α = the risk of rejecting the null hypothesis (0.05), d = degree of precision or margin of error, Z = the standard score corresponding to a 95% confidence interval and p = 50% (the proportion of institutional deliveries).8 The final sample size was obtained after adding 10% to compensate for possible non-response and multiplying the result by a factor of two for design effect, resulting in a sample size of 844. The study employed a multistage sampling scheme. First, approximately 40% (four out of 11 kebeles) of the town were selected randomly to conduct a census for the identification of mothers who met the study criteria. This generated a sample frame comprising 1573 women. The probability of being included in the sample was equal for all the women in the sample frame since random numbers generated using SPSS version 16 were applied in order to identify the study participants. The women who were included in the sample were interviewed at home, using a questionnaire administered in Wolaitigna (the local language). A cross-sectional study design was used in Sodo town, which is 327 km south of Addis Ababa, the capital city. In the town there are two hospitals (private and government), three health centres, seven private clinics, three pharmacies and eight drug vendors. A pre-tested interviewer-administered structured questionnaire adapted from the Ethiopian Demography and Health Survey (EDHS) tool and related theses was used.4, 9, 10 Scales combining multiple items were used in order to measure perceived access to and quality of care at the nearest health facility. Data were checked for completeness and then coded and entered in to EpiData version 3.5.3. The data were then exported into SPSS version 16 for further analysis. Variables with significant association during bivariate analyses were entered into multiple logistic-regression models so as to control for possible confounding effects of the variables and to identify the independent predictors of institutional delivery. The strength of association was estimated by calculating the odds ratios (OR) with 95% confidence intervals (CI). A p-value of < 0.05 was taken as being of statistical significance for all analyses.
N/A