A comparative analysis of predictive models of morbidity in intensive care unit after cardiac surgery – Part II: an illustrative example
2007

Comparative Analysis of Predictive Models for Morbidity After Heart Surgery

Sample size: 1090 publication 10 minutes Evidence: moderate

Author Information

Author(s): Gabriele Cevenini, Emanuela Barbini, Sabino Scolletta, Bonizella Biagioli, Pierpaolo Giomarelli, Paolo Barbini

Primary Institution: Department of Surgery and Bioengineering, University of Siena, Siena, Italy

Hypothesis

How do different predictive models compare in estimating morbidity in ICU patients after cardiac surgery?

Conclusion

The Bayes quadratic model, using only three predictors, provided the best performance in predicting morbidity after heart surgery.

Supporting Evidence

  • All models showed acceptable discrimination capacities.
  • The Bayes quadratic model exhibited the best generalization ability.
  • Poor calibration was observed in scoring systems and artificial neural networks.

Takeaway

This study looked at different ways to predict if heart surgery patients would have problems afterward, and found that a simple model using just three factors worked best.

Methodology

Eight predictive models were developed and tested using preoperative, intraoperative, and postoperative data from 1090 patients.

Potential Biases

Potential overfitting in models with many predictors, particularly artificial neural networks.

Limitations

The models were not tested on independent data from different ICUs, which may limit generalizability.

Participant Demographics

Adult patients who underwent coronary artery bypass grafting.

Statistical Information

P-Value

p<0.05

Confidence Interval

95% CI for AUC ranged from 0.722 to 0.831.

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1186/1472-6947-7-36

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication