Index for predicting mortality in elderly surgical patients.

Area of Science:

• Geriatric Surgery
• Surgical Outcomes Research
• Clinical Prediction Models

Background:

• Elderly patients (over 65) undergoing surgery require accurate mortality prediction.
• Existing indices may not be suitable for this demographic or treatment spectrum.

Purpose of the Study:

• To validate a mortality index specifically designed for elderly patients (over 65) receiving surgical or conservative management.
• To assess the index's predictive accuracy in a contemporary patient cohort.

Main Methods:

• A multivariate logistic regression model was used to develop the index from 11 significant predictors identified in a historical cohort.
• The index was validated using data from a separate cohort of comparable elderly patients treated between 1999-2001.

Main Results:

• The validation cohort showed similar variable distributions and mortality rates (9.6%) compared to the derivation cohort (9.3%).
• Calibration and discrimination tests confirmed no significant difference between predicted and actual deaths.
• Receiver-operator characteristic (ROC) plots demonstrated good predictive discrimination in both datasets.

Conclusions:

• The developed mortality index is validated for predicting outcomes in elderly patients (over 65) undergoing operative or conservative treatment.
• This tool can assist clinicians in counseling patients and their families about treatment outcomes.