Surgical Outcomes and Quality Assessment of Trabeculectomy: Leveraging Electronic Health Records for Clinical Data Visualization

摘要

Supplemental Digital Content is available in the text. Précis: This electronic medical record-based study demonstrates 1 approach to monitor clinical outcomes for individual physicians performing trabeculectomy. Purpose: The purpose of this study was to develop an interactive data visualization platform for evaluating trabeculectomy outcomes and monitoring clinical care quality. Patients and Methods: This is a retrospective electronic health record study in a glaucoma division of an academic medical center. We included 633 patients who underwent trabeculectomy between July 2016 and December 2018. Data were collected for intraocular pressure (IOP), vision loss, number of follow-up visits, and return to the operating room within the initial 3 months. Regression analyses were used to assess factors associated with these study outcomes. Risk of IOP ≤5?mm?Hg was assessed using survival analysis. An interactive data visualization platform was created. Results: A total of 429 trabeculectomy and 204 combined surgeries were analyzed, with preoperative IOP of 21.9±9.4 and 19.4±6.9?mm?Hg, respectively. Mean pressure declined by 53% and 41% ( P =0.001) by 3 months. Risk of pressure ≤5?mm?Hg was higher in white patients and those who underwent trabeculectomy alone. At 3 months, patients on average had 5 clinic visits; 5.5% required additional surgical procedures, and 18% experienced loss of at least 3 lines of best-recorded (not best-corrected) vision. Returning to the operating room was significantly associated with an increased number of clinic visits (coefficient=3.94) and higher odds of vision loss (odds ratio=11.12). Conclusions: Over the initial 3 months after trabeculectomy, additional surgeries are rare, while a significant proportion of patients lost vision, and this was strongly associated with a need to return to the operating room. Interactive electronic health record data visualization can be a useful tool for monitoring clinical outcomes.