Statistical analysis of time to recovery from vesico vaginal obstetric fistula: a case study at metu hamlin fistula center, metu, south west Ethiopia

Abstract

This study was done to identify factors affecting recovery time of vesico vaginal obstetric fistula patients. A sample of 206 patients was taken from a hospital records at Metu Hamlin Fistula Center enrolled from November 2010 to June 2014 G.C. Objective: The objective of this paper is to model time to recovery from VVOF at Metu Hamlin Fistula Center and to compare the performance of LASSO method of variable selection with step wise selection methods. Methods: Least Absolute Shrinkage and Selection Operator (LASSO) method of variable selection and stepwise variable selection method, Kaplan-Meier estimation method, Cox proportional hazard model and parametric regression model were applied. Results: The Cox PH analysis using LASSO variable selection method identified that survival of the patients was significantly related with age at first marriage, weight, antenatal care ,educational status, marital status, duration of labor, duration of incontinence, place of delivery, mode of delivery, width of fistula and status of urethra. But the covariate like current age of a patient, height, parity, accompanying person, fetal outcome, and bladder size were not statistically significant at 5% significance level. The result from Weibull regression analysis showed that recovery of VVOF patients was significantly related with age at first marriage, duration of incontinence, duration of labo r, place of delivery, mode of delivery, fetal outcome, width of fistula, and status of urethra. Conclusion: The finding of this study showed that out of 206 VVOF patients 76.2% of them were physically cured while the rest 23.8% were censored. The Cox PH model provides better predictions to the survival probability of VVOF patients. It is advisable to make an intervention on the identified risk factors.