Investigating Predictors of Length of Stay and Mortality in the ICU

Introduction: As the need for intensive care units (ICUs) rises due to an aging population and increasing disease burden, optimizing outcomes like length of stay (LOS) and mortality is essential for improving care quality and resource utilization. Current predictive tools like APACHE and SAPS often overestimate mortality and lack population-specific calibration. Here we discuss the development of an ICU registry to conduct a longitudinal analysis on factors associated with longer ICU LOS and increased mortality.


Methods: We developed a longitudinal EHR-derived registry capturing adult ICU admissions across eight hospitals in a large hospital system. Data spanned from 2016–2025. Data included demographics, comorbidities, labs, insurance, social determinants (via Area Deprivation Index), and clinical outcomes. Multivariable regression models identified predictors of ICU LOS, mortality, and discharge disposition.


Results: The dataset included 157,188 ICU encounters. Median age was 66.5 years; 44.9% were female. Renal disease, COPD, high WBC/BUN, and low albumin/RBC were associated with high ICU LOS. Age ≥80 was associated with the shortest LOS but highest mortality. African American and Asian patients had longer LOS but lower mortality. Self-pay patients had shorter LOS but higher mortality, while those with private insurance had longer stays and lower mortality. Patients with metastatic tumors had a shorter LOS but markedly higher mortality (OR 3.72). ICU LOS was strongly linked to mortality risk, with the fourth quartile showing the highest mortality odds. Patients from high socioeconomic deprivation areas had higher LOS and mortality, particularly in the 50–59 and ≥80 age groups.


Conclusions: Our findings highlight demographic, clinical, and socioeconomic contributors to ICU LOS and mortality. While LOS can reflect illness severity or care intensity, shorter LOS in older or high-risk patients may be driven by early mortality. These results emphasize the need for individualized ICU care planning and integration of social determinants to address disparities in outcomes.