Zero-inflated negative binomial modeling, efficiency for analysis of length of maternity hospitalization

Abstract

Background: Mothers’ delivery is one of the most common hospitalization factors throughout the world and it’s modeling can explain distribution and effective factors on rising and decreasing of it. The objective of the present study was a suitable modeling for mother hospitalization time and comparing it with different models. Materials & Methods: Present study is an observational and cross-sectional study with randomized sampled of 1600 mothers’ refered to Arak university treatment centers in the first seamester in 2004 for delivery. The following parameters were registered: hospitalization time as dependent variable, mother’s age and its square, mother job, having abnormal child, ordinal pregnancy or delivery and its square, number of abortions and its square, number of present children and its square, mothers’ residency, type of delivery, twice and triplets all were considered as independent variables. For analysis of data, advanced recent methods of countable data modeling were used. We also introduced an innovative method of analysis. Results: The results of modeling of mothers’ hospitalization time showed negative binomial model was a suitable model because of unequal variance and means of dependent variables for explanation of mothers’ hospitalization time, having abnormal child, type of delivery (NVD, C&S) and twice delivery all were significant variables in this model. More specific models (Zero-truncated Poisson and negative binomial), showed to be more suitable for age and its square, having abnormal child, type of delivery, twice delivery and triplet delivery which were all significant variables in determining of mothers’ hospitalization time rates. Conclusion: In this article, with a simple change of mothers hospitalization time, a suitable statistical model to explain them and modeling of these times were achieved. The suggested model could included more variables than conventional because of its higher specificity.