Visual outcome after open globe injury: a comparison of two prognostic models--the Ocular Trauma Score and the Classification and Regression Tree.

AIMS: To compare the ocular trauma score (OTS) and the classification and regression tree (CART) as prognostic models of visual outcome after open globe injury.
METHODS: A retrospective review of 100 consecutive open globe injuries at the Sunderland Eye Infirmary was conducted from January 1999 to December 2007. Univariate chi-square analysis was used to identify the variables associated with visual outcome. We compared the CART and OTS predictions with the actual visual outcomes and calculated the sensitivity and specificity of each model.
RESULTS: The variables most predictive of visual loss were an RAPD, poor initial vision, lid laceration, posterior wound, and globe rupture. The sensitivity to predict visual survival (LP or better) was 97.4% for OTS and 93.5% for CART. The specificity to predict no vision (NPL or enucleation) was 100% for OTS and 73.9% for CART. The sensitivity to predict minimal-to-severe visual loss (3/60 or better) was 90.9% for OTS and 85.7% for CART. The specificity to predict profound visual loss (worse than 3/60) was 100% for OTS and 81.8% for CART.
CONCLUSIONS: We identified several factors that can help in deciding on the prognostic value of primary globe repair. Both the OTS and CART had high predictive accuracy, but the OTS had higher prognostic accuracy and could be used in counselling patients and in management decision-making.