Related Experiment Videos
Outcome prediction model for very elderly critically ill patients.
D M Nierman1, C B Schechter, L M Cannon
1Department of Medicine, The Mount Sinai Medical Center, New York, NY, USA.
Critical Care Medicine
|October 6, 2001
Summary
This study developed a prediction model for very elderly critically ill patients, identifying factors for hospital discharge or death. The model aids in discussing patient outcomes and planning care.
Area of Science:
- Critical care medicine
- Geriatric medicine
- Health outcomes research
Background:
- Outcomes for critically ill elderly patients (≥85 years) are complex, including home discharge, skilled nursing facility discharge, or death.
- Factors influencing these distinct outcomes remain largely unknown.
Purpose of the Study:
- To develop a predictive model for three distinct hospital outcomes in critically ill elderly patients.
- To identify key factors associated with patient disposition after intensive care unit (ICU) admission.
Main Methods:
- Retrospective chart review and analysis of 455 patients aged 85 years or older admitted to ICUs in 1996-1997.
- Ordinal logistic regression was used to develop and validate a predictive model.
Main Results:
- The final model incorporated age, gender, baseline support, ICU type, heart rate, mechanical ventilation, vasopressor use, pulmonary artery catheter use, and development of respiratory, neurologic, hematologic failure, or sepsis.
- The model demonstrated good calibration and high discriminant index when validated on independent data.
Conclusions:
- A validated mathematical model can predict the likelihood of home discharge, skilled nursing facility discharge, or death for very elderly critically ill patients.
- These predictions can inform discussions about prognosis and goals of care, and assist in discharge planning.