Predictors of in-hospital mortality in nonsurgical departments: a multivariable regression analysis of 2855029 hospitalizations.

Introduction: In-hospital mortality is an important outcome of hospital admission. Objectives: To identify predictors that are independently associated with non-surgical units’ in-hospital mortality. Patients and methods: The National Health Fund database provided data from 2014 on 2,855,029 hospitalizations of adults, unrelated to surgical procedures. Analyzed mortality predictors were: the patients’ age and gender, diagnosis-related group category assigned to the hospitalization, length of the hospitalization, type of hospital and admission, day of week and month of admission. Results: The mean in-hospital mortality rate was 4.1%. Odds for in-hospital death increased with patients’ age. The female gender was associated with a lower odds of death than the male gender. Among the diagnosis-related groups assigned to the hospitalizations, the highest mortality was found for vascular diseases (11.95%). Considering the length of hospitalization, it was found that the lowest mortality occurred during 5–7-day hospitalizations (2.63%). Compared with clinical hospital, the odds of death was 1.31-fold higher for regional hospitals, 1.35-fold higher for private hospitals and 1.48-fold higher for county/town hospitals; 92% of all in-hospital deaths occurred after urgent and emergency admissions. Hospital admissions at weekends or other non-working days (bank holidays) were statistically significant predictors of in-hospital death. There were differences in mortality between particular months, but there was no seasonal relationship. Conclusions: Age, male gender, emergency admission, admission during the weekend or on another non- working day (bank holiday) and hospitalization in a county/town, private or regional hospital (vs teaching hospital) are factors associated with higher mortality in non-surgical units.