A 100% departmental mortality review improves observed-to-expected mortality ratios and University HealthSystem...

Public reporting of mortality, Patient Safety Indicators (PSI) and hospital-acquired conditions (HACs) is the reality of quality measurement. A review of our department's data identified opportunities for improvement. We began a surgeon-led 100% review of mortality, PSIs, and HACs to improve patient care and surgeon awareness of these metrics.

From December 2012 through August 2013, there were 11,899 patients cared for on 12 surgical services. A surgeon from each service led monthly reviews of all mortality, PSIs, or HACs with central reporting of preventability and coding accuracy. We compared the University HealthSystem Consortium observed-to-expected (OE) mortality ratios (mean <1 fewer observed than expected deaths) and University HealthSystem Consortium relative rankings (lower number is better) before and after implementation. Statistical significance was p < 0.05 by Poisson regression.

Of the 11,899 patients in the study period, there were 235 deaths, 290 PSIs, and 26 HACs identified and reviewed. The most common PSIs were postoperative deep vein thrombosis/pulmonary thromboembolism (n = 75), respiratory failure (n = 61), hemorrhage/hematoma (n = 33), and accidental puncture/laceration (n = 33). Before December 20, 2012, the OE ratio for mortality was consistently >1, then fell and remained <1 during the study period (p < 0.05). The OE mortality ratio in the fourth quarter of 2012 was 1.14 and fell to 0.88, 0.91, and 0.75 in the first, second, and third quarters of calendar year 2013 (p < 0.05). The overall Inpatient Quality Indicators #90 (composite postoperative mortality rank) rankings increased from 109 of 118 in the third quarter of 2012 to 47 of 119 in the third quarter of 2013.

A surgeon-led systematic review of mortality, PSIs, and HACs improved our OE ratio and University HealthSystem Consortium postsurgical relative rankings. Surgeon engagement and ownership is critical for success.