Introduction: Patient hospitalization rate (P.H.R) is an important issue in health decision making because its increasing can imposes abundant expenditure on each patient or therapeutic or health systems like assurance companies. Moreover, it can increase the danger of nosocomial infections. So, description, examination and modeling of P.H.R. can be as a guideline for above subjects and as an introduction would lead one to study the issue with insight.
Materials and Methods: For determining of P.H.R., two samples of size 1428 and 816 of patients were chosen systematic randomly from Vali-Asr and Ta'min-Ejtemai hospitals, respectively during the year of 2002. Hospitalization rate, age, sex, location of birth place and residence place of patients recorded. Mixed distributions modeling and Poisson regression models as well as ordinary regression methods were applied to analyze and modeling the data.
Results: Means of P.H.R of patients at Vali-Asr hospital was 4.1 ±5.4 days. Two component Poisson was a relatively appropriate model was a very relevant model for analyzing of their P.H.R. By using mixed Poisson regression method a significant relationship among patient's age, marriage status and location of life with P.H.R was obtained. Findings in Ta'min-Ejtemai showed that the mean of P.H.R was 4.7±3.2 days. One component Poisson model was appropriate model for data distribution. A significant relationship among age, sex, location of life with P.H.R in this hospital was seen.
Conclusion: Our findings showed that P.H.R in Vali-Asr hospital follows two component Poisson model. For Vali-Asr data, the normal model was not appropriate even with changing the statistics variables, but for P.H.R data of Ta'min-Ejtemai, changing the logarithmic variable could transform data into approximate of normal model. Finally, comparison between mixed distribution Poisson model and acquired regression with simple linear regression in Vali-Asr hospital presented different results, but in Ta'min-Ejtemai hospital had similar results.