Ophthalmic Surgery Outcomes Database Project

A pilot project was developed to begin data collection and make recommendations regarding ophthalmic surgical procedures in the Veterans Health Administration. Only cataract surgeries were included because these represented approximately 85% of all ophthalmic surgical procedures in the Veterans Health Administration system. Five Veterans Health Administration pilot sites were selected to participate. The pilot sites had to meet the following criteria: (1) perform a minimum of 500 ophthalmic surgical cases per year; (2) have an identifiable Clinical Champion for the project; and (3) have the ability to identify and employ a nurse reviewer for the project. Registered nurse reviewers were selected to collect preoperative, intraoperative, and postoperative data from the Veterans Affairs electronic medical record and also collect preoperative (within 30-90 days of surgical procedure) and postoperative (30-90 days post procedure) National Eye Institute Visual Function Questionnaire (NEI-VFQ) forms from patients who were undergoing cataract surgery.

Development of a data collection tool known as the Risk Assessment Form (RAF) was initiated by Clinical Champion physicians at the participating pilot sites. It was decided to include data from all phases of the surgical experience. In all, approximately 150 data elements were chosen for inclusion in the original RAF. These data elements were defined by using Veterans Affairs Surgical Quality Improvement Project definitions when possible and establishing new definitions as needed when prior definitions were not available. Use of the RAF enabled the 5 pilot sites to standardize electronic templates that include preoperative assessment, intraoperative reporting, and postoperative follow-up to provide a viable framework for data gathering. Surgeons entered all relevant surgical elements in an operative note template immediately following each surgical case.

A centralized Microsoft Access database was developed and nurse reviewers began administering preoperative and postoperative NEI-VFQs and manually entering collected preoperative and postoperative clinical data on each patient, as well as the surgeons’ operative notes, and transferring these data in masked form to the database in July 2009. Data from each of the pilot sites were validated and compiled for statistical purposes.

Records for eligible patients were abstracted by the nurse reviewers and entered into individual Microsoft Access databases via a graphical user interface. The Ophthalmic Surgery Outcomes Database Project employed a statistician to help with the data management portion of the project. After records accumulated for a period of time, the Access databases were provided to the statistician for data validation and statistical analysis. The statistician first merged the individual Access databases into a central staging Microsoft SQL Server database, and then ran data validation and cleaning code using the R statistical software package. Next, the statistician discussed validation issues with the nurse reviewers and necessary corrections to the database were made. Finally, clean records were stored in a Microsoft SQL Server database for statistical analysis.

Pertaining to this retrospective analysis, the following data elements were analyzed: alpha-blockers (both selective and nonselective), IFIS, intraoperative iris trauma, intraoperative iris prolapse, posterior capsular tear, anterior capsule tear, intraoperative vitreous prolapse, and use of pupillary expansion devices. The primary endpoint of this study was to determine the prevalence of IFIS and association with the use of alpha-antagonists. Secondary endpoints included analysis of the use of pupillary expansion devices and occurrence of complications in patients experiencing IFIS and taking alpha-antagonists. The decision of determining IFIS was left to the individual surgeon, although flaccid iris that surges with normal intraoperative fluidics and progressive pupil constriction were used in the determination. All cases of IFIS were documented by the surgeon immediately after the surgical case. Iris prolapse was listed as a separate variable that was documented. Determination of use of a selective or nonselective alpha-antagonist was done by analyzing the Veterans Affairs electronic medical record, looking for selective alpha1a-antagonists (tamsulosin and alfuzosin) and nonselective alpha1-antagonists (doxazosin, prazosin, terazosin), in which the patient was currently taking the drug or had taken it in the past year. This was a dichotomous variable (ie, yes or no) and did not differentiate between medications within a specific class.

Data Analysis

Data were analyzed using Microsoft Excel 2010 (Version 14.0; Microsoft Corporation, Redmond, Washington, USA), EpiInfo (Version 7.1.0.6, 2012; Centers for Disease Control and Prevention, Atlanta, Georgia, USA), and JMP Pro software (Version 10.0, SAS Institute, Cary, North Carolina, USA). Statistical evaluation included χ ² testing, Fisher exact test, and Mantel-Haenszel method or nominal logistic regression to estimate odds ratios (OR) with 95% confidence intervals (CI). P values <.05 were considered statistically significant.