Predicting visual outcome after open globe injury using classification and regression tree model: the Moradabad ocular trauma study.

Abstract

OBJECTIVE\nThis study was conducted to identify factors associated with visual outcome in patients with open globe injuries (OGIs).\n\n\nDESIGN\nRetrospective case series of OGIs\xa0presenting to a tertiary eye care institute in North India from October 2009 to December 2016.\n\n\nMETHODS\nA total of 157 patients with open globe injury have been included in the study. Multivariate analysis to ascertain the effects of different identified variables on the likelihood of poor visual outcome was done using binomial logistic regression. Visual survival (counting fingers or better) versus minimal/no vision (hand motion, light perception, and no light perception) was predicted using the classification and regression tree (CART) model. Main outcome measures were visual outcomes, risk factors, and rates of postoperative complications.\n\n\nRESULTS\nUnivariate analysis determined 9 predictors associated with poor visual outcome. Out of these, presence of relative afferent pupillary defect (RAPD), poor presenting visual acuity, presence of adnexal injuries, and location of injuries were the most significant predictors of vision loss. Absence of RAPD led to 79% chance of vision survival. Sixty-eight percent of patients with RAPD and initial visual acuity (VA) of less than 6/60 resulted in poor vision.\n\n\nCONCLUSION\nThe CART model is useful in predicting final VA based on some prognostic factors present initially.