Prognosis of Open Globe Injuries at a Tertiary Referral Center: The Modified Florida Ocular Trauma Score

Abstract

To analyze vision outcomes after open globe injury and propose modifications to the ocular trauma score to offer more specific vision prognoses. Validity and reliability analysis. Patients presenting to the University of Florida with a new open globe injury from October 2015 to January 2021 with subsequent follow-up were included in the study. Demographics, ophthalmic history, trauma details, timeline, imaging, operative findings, and ocular examinations were collected from the medical record. Z tests, χ2 test, Fisher exact test, receiver operating characteristic curve, and ordinal correlation were used. A weighted logistic model was optimized to predict vision outcomes. Measured outcomes included the best-corrected visual acuity, Ocular Trauma Score category, and performance of vision prognosis scores. A total of 162 eyes were identified from chart review. Eighty percent of the Ocular Trauma Score categories were accurate. Only the absence of orbital fractures was associated with a significant weight in the logistic model, which produced more accurate prognoses for 59 patients, and less accurate prognoses for 30 patients compared to the Ocular Trauma Score. Kendall Tau-B was 0.639 for the logistic model and 0.582 for the Ocular Trauma Score. The Ocular Trauma Score accurately estimates vision prognosis after open globe injury. We propose inclusion of orbital fracture status in our Modified Florida Ocular Trauma Score. This addended score is more correlated with final vision outcome and provides more specific prognoses for severe open globe injuries. Prospective, multicenter validation is needed to refine and confirm the use of this new scoring system.