Detecting patient safety indicators: How valid is "foreign body left during procedure" in the Veterans Health Administration?

Abstract

Background: The Agency for Healthcare Research and Quality (AHRQ) developed patient safety indicator (PSI) 5, "Foreign body left during procedure," to flag accidental foreign bodies in surgical and medical procedures. This study examined how well this indicator identifies true foreign body events in the Veterans Health Administration (VA).

Study design: This was a retrospective study within 28 selected VA hospitals from fiscal year 2003 to 2007. Trained abstractors reviewed medical charts flagged by PSI 5 and determined true foreign body cases. We calculated the positive predictive value (PPV) of this indicator and performed descriptive analyses of true positive and false positive cases.

Results: Of the 652,093 eligible cases, 93 were flagged by PSI 5 (0.14 per 1,000). Forty-two were true positives, yielding a PPV of 45% (95% CI 35% to 56%). False positives were due to a foreign body that was present on admission (57%) or coding errors (43%). True foreign bodies were associated with surgical (n = 23) and medical (n = 19) procedures. The most common type of surgical foreign body was a sponge (52%). Overall, approximately 40% of foreign bodies were related to a device failure or malfunction (30% surgical vs 53% medical foreign bodies). Postoperative complications included pain (24%), infection (12%), adhesions (5%), and bowel obstruction (5%).

Conclusions: The reported rate of foreign body events as detected by PSI 5 is low in the VA, but occurs in both surgical and medical procedures. Despite widespread implementation of surgical counts, quality improvement efforts should focus on novel ways to eliminate this "never event" from operations. Future studies are needed to better understand the preventability of medical procedure-associated foreign bodies and particularly, device failure-related foreign bodies.