Determination of Factors Affecting Length of Stay With Multinomial Logistic Regression in Turkey

Abstract

Length of stay (LOS) has important implications in various aspects of health services, can vary according to a wide range of factors. It is noticed that LOS has been neglected mostly in both theoratical studies and practice of health care management in Turkey. The main purpose of this study is to identify factors related to LOS in Turkey. A retrospective analysis of 2.255.836 patients hospitalized to private, university, foundation university and other (municipality, association and foreigners/minority hospitals) hospitals which have an agreement with Social Security Institution (SSI) in Turkey, from January 1, 2010, until the December 31, 2010, was examined. Patient's data were taken from MEDULA (National Electronic Invoice System) and SPSS 18.0 was used to perform statistical analysis. In this study t-test, one way anova and multinomial logistic regression are used to determine variables that may affect to LOS. The average LOS of patients was 3,93 days (SD = 5,882). LOS showed a statistically significant difference according to all independent variables used in the DETERMINATION OF FACTORS AFFECTING LENGTH OF STAY WITH MULTINOMIAL LOGISTIC REGRESSION IN TURKEY75 study (age, gender, disease class, type of hospitalization, presence of comorbidity, type and number of surgery, season of hospitalization, hospital ownership/bed capacity/ geographical region/residential area/type of service). According to the results of the multinomial lojistic regression analysis, LOS was negatively affected in terms of gender, presence of comorbidity, geographical region of hospital and was positively affected in terms of age, season of hospitalization, hospital bed capacity/ ownership/type of service/residential area.