Artikel
Vision Survival after Open Globe Injury Predicted by Classification and Regression Tree Analysis
Suche in Medline nach
Autoren
Veröffentlicht: | 18. Juni 2008 |
---|
Gliederung
Text
Objective: To assist ophthalmologists in treating ocular trauma patients, this study develops and validates a prognostic model to predict vision survival after open globe injury.
Design/Methods: A retrospective cohort review of 221 patients who presented to the Wilmer Ophthalmological Institute with open globe injuries from January 1, 2001, to December 31, 2004, was conducted. Binary recursive partitioning was used to construct a classification tree to predict two visual outcomes: 1) vision survival (range from 20/20 to light perception vision) or 2) no vision (includes no light perception, enucleation, and evisceration outcomes). Patients with open globe injuries from January 1, 2001, to December 31, 2004, were used to grow the tree and constitute the “training” sample. A second independent sample of patient eyes seen from January 1, 2005, to October 15, 2005, were used to validate the classification tree.
Results: A classification tree was constructed using 214 open globe injuries. Out of 14 predictors determined to be associated with a no vision outcome in univariate analysis, presence rAPD and poor initial visual acuity were the most predictive of complete loss of vision; presence of lid laceration and posterior wound location also predicted poor outcome. In an independent cohort of 51 eyes, the tree had 85.7% sensitivity to correctly predict no vision, and 91.9% specificity, to correctly predict vision survival.
Conclusions: The classification tree constructed in this study demonstrated excellent predictive accuracy and should be useful in counseling patients and making clinical decisions regarding open globe injury management.