Abstract
Background/Objectives: Hospital length of stay (LOS) is widely analyzed and serves as a benchmark for assessing changes during hospitalization. This study introduced a method to estimate patients' LOS and highlighted the variations in LOS among individuals with or without multiple chronic conditions (MCCs) and across different levels of disease severity, using data from the 2016 National Inpatient Sample in the United States. Methods: To analyze the factors influencing LOS, a multinomial logistic regression model was employed, demonstrating its effectiveness in estimating and predicting expected LOS. Factors such as demographic characteristics, MCCs, and disease severity were strongly linked to LOS. Results: The overall prevalence of MCCs exceeded 66%, rising to over 90% among elderly patients and more than 88% among those with severe diseases. LOS distribution was primarily concentrated within the first month following admission: over 13% of patients were discharged within a day, over 85% within a week, and more than 99% within a month. Multinomial logistic regression analysis showed that LOS was significantly influenced by age, disease severity, and the presence of MCCs. Older patients, especially those with MCCs, had significantly longer LOSs compared to younger patients without MCCs. Conclusions: LOS tended to increase with age and higher disease severity, particularly in patients with MCCs. Multinomial logistic regression revealed that patients over 65 and those with high disease severity (severity score 4) had significantly longer LOS. Shorter LOS was more frequent among patients under 65 years old, those without MCC, and those with low disease severity, whereas longer LOS was commonly observed in patients with MCCs or high disease severity.