Katia Noyes, PhD, MPH1, Evan Kessler, MD2, Monami Majumdar, MPH2, Ajay A Myneni, MBBS, PhD, MPH2, Steven D Schwaitzberg, MD2, Aaron B Hoffman, MD, FACS2. 1Division of Health Services Policy and Practice, Department of Epidemiology and Environmental Health, School of Public Health and Health Professions, University at Buffalo, 2Department of Surgery, Jacobs School of Medicine and Biomedical Sciences, University at Buffalo
INTRODUCTION: This study examined the quality of Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program (MSBASQIP) data to assess its usability potential, improve study methodologies based on large pre-existing datasets and data collection procedures, and advance the field of bariatric surgery. Use of large national registries and population-level datasets has become a common practice in clinical research, quality improvement and health policy. Relatively low cost and ease of access make these datasets especially attractive for trainees and small clinical practices that do not have affiliations with academic research centers. However, little is known about methodological challenges associated with the analysis of multi-center self-reported registries that could limit their impact and compromise patient safety.
METHODS AND PROCEDURES: We developed a single flat file for patient-level outcomes evaluation (n=168,093) using 5 files (Main, BMI, Readmission, Intervention and Reoperation) that constitute the 2015 MBSAQIP. Logic and validity tests included (1) individual profiles of patient BMI changes over time; (2) individual patient care pathways (chronologic record of patient admission, discharge and procedure history); and (3) correlation tests between pairs of variables associated with the same clinical encounters (emergency intervention vs. procedure type, related admission with intervention vs. planned intervention). Appropriate bivariate and multivariate analyses were performed.
RESULTS: Our analysis revealed several data consistency problems. Among 168,093 patients, only 8.6% had the same measurement unit for preoperative and postoperative weight; 10,725 cases (6.4%) had missing/zero values. The percent difference in weight between pre- and post-surgical appointments ranged from -132% to 72% (-522lb to 625lb, with 5.5% of patients having pre-surgical BMI 60 or higher) of the pre-operative weight. Hospital readmissions on “day 0” were prevalent (n=238 or 2.7%). The self-reporting of “emergency” vs. “planned” interventions did not correlate with the type of procedure and the indication.
CONCLUSIONS: Utilization of a national bariatric registry is fundamental to advancing quality of bariatric care and patient outcomes. However, MBSAQIP data inconsistencies significantly undermine the quality of the registry. Given that the dataset lacks information on where the surgery was performed, it is not possible to conclude if these inconsistencies represent data errors, patient outliers or inappropriate care. Including automated data checks and biomedical informatics oversight, standardized coding for complications, additional de-identified facility and provider information, and training and mentorship opportunities in data informatics for all researchers who get access to the data have been shown to be effective in improving data quality and minimizing patient safety concerns.
Presented at the SAGES 2017 Annual Meeting in Houston, TX.
Abstract ID: 95256
Program Number: S122
Presentation Session: Bariatric IV – Quality and Outcomes
Presentation Type: Podium