Usefulness of Risk Models to Predict Hospital Length of Stay After Cardiac Surgery

pp 185-191

Abstract

Background

A rational administration of cardiac surgery care resources requires prediction and controlling of in-hospital length of stay. The close relation between preoperative risk and postoperative length of stay, needs a risk model able to predict these time.

Objective

To validate two risk indexes to compare mortality and overall hospital length of stay after cardiac surgery.

Method

Postoperative mortality and length of stay of 320patients undergoing cardiac surgery, were compared with two risk indexes (PACCN and Parsonnet).

Results

No statistical differences were found between ob-served and expected mortality with both indexes. In-hospital length of stay was 7 days (median) for aPACCN score of 6 or less, and 12 days for a greater score. Length of stay with Parsonnet index was 7days when risk score was inferior to15, 8 days when values were between 15 and 19, and 10 days for greater scores.

Conclusions

Both risk indexes adequately predict observed mortality in this study. On the other hand, these models could be used to predict postoperative length of stay.