Discussion Between Reviewers Does Not Improve Reliability of Peer Review of Hospital Quality

摘要: Objectives. Peer review is used to make final judgments about quality of care in many assurance activities. To overcome the low reliability peer review, discussion between several reviewers often recommended point out overlooked information or allow for reconsideration opinions and thus improve reliability. The authors assessed impact 2 on review. Methods. A group 13 board-certified physicians completed a total 741 structured implicit record reviews 95 records patients who experienced severe adverse events related laboratory abnormalities while hospital (hypokalemia, hyperkalemia, renal failure, hyponatremia, digoxin toxicity). They independently degree which each event was caused by medical leading up event. Working pairs, they then discussed differences opinion, clarified factual discrepancies, rerated record. compared measure before after discussion, within pairs reviewers, using intraclass correlation coefficient continuous ratings kappa statistic dichotomized rating. Results. assessment whether abnormality iatrogenic had 0.46 0.71 paired indicating considerably improved agreement members pair. However, across reviewer 0.36 0.40 discussion. Similarly, rating overall care, physician went from 0.35 0.58 as increased only 0.14 0.17. Even prediscussion ratings, substantially higher pair than suggesting that work learn be more consistent with other even but this consistency also did not pairs. Conclusions. When discuss are reviewing, it improves those physicians. improvement illusory, does examining were part different discussions. This finding may have implications regard how disagreements resolved consensus panels, guideline committees, literature meta-analyses.