Implementation of a Standardized Tool for Root Cause Analysis Selection

This study sought to determine if a standardized root cause analysis (RCA 2 ) selection algorithm, developed by the Veterans Affairs, would select high-risk events for RCA.

Physician-entered incident reports for all surgical service admissions and perioperative visits were queried over 12 months in the DATIX Clinical Incident Management System. Independent reviewers assigned potential harm and event frequency scores using an institutional scoring system and then calculated and rounded average scores. These were classified using RCA 2 terminology (catastrophic, major, moderate, minor for harm; frequent, occasional, uncommon, remote for frequency). The scores were then evaluated with the standardized Safety Assessment Code Matrix (SAC) algorithm from the National Patient Safety Foundation's RCA 2 guidelines to determine Potential Harm Scores. The SAC combines severity and probability to determine the necessity of conducting an RCA. Catastrophic and major high-frequency events (matrix score = 3) were classified as "RCA recommended." The study then compared cases selected for RCAs using the updated RCA 2 algorithm against cases selected using a current, institutional-specific RCA selection process.

One hundred four cases were reviewed, comprising 20 catastrophic, 48 major harm, 26 moderate harm, and 10 minor harm events. After removing 9 high-variance cases, our institution's current selection process selected 18 cases for RCAs, including 6/20 catastrophic, 8/39 major harm, and 4/36 moderate/minor harm events. Only 17.3% of cases had an RCA completed, while the standardized RCA 2 algorithm recommended investigation for 56.7% of patient safety events, based on SAC Matrix scoring. Current RCA selection processes rendered 4 RCAs on low potential harm or low-frequency events, while 45 potential high-frequency, high potential harm events did not complete RCAs.

Standardizing the selection of patient safety incidents for RCA using the RCA 2 algorithm improves case identification based on the event frequency and potential harm score. Thus, this algorithm has the potential to advance patient safety.