An evidence-based approach to quality improvement in gynecologic surgery: Identifying complications with the highest impact on adverse clinical outcomes and resource utilization following hysterectomy (395)

Abstract

<b>Objectives:</b> To identify surgical complications with the highest impact on adverse clinical outcomes (end-organ dysfunction [EOD], unplanned reoperation [ROR], death) and resource utilization (prolonged length of stay [PLOS], unplanned re-admission) following hysterectomy. These results will help focus quality improvement efforts on high yield targets. <b>Methods:</b> NSQIP data for women undergoing hysterectomy from 2014-2017 were obtained and stratified by surgical approach (laparotomy vs minimally invasive surgery [MIS]) and indication (benign vs malignant). Univariate associations between postoperative complications and adverse outcomes were assessed using Chi-square and Fisher's exact tests. We calculated the population attributable fraction (PAF) to quantify the contribution of each postoperative complication on adverse outcomes. PAF estimated the proportion of adverse outcomes that could be eliminated if the corresponding complication was prevented. <b>Results:</b> We included 94,708 women with a mean age of 51.4 years undergoing hysterectomy. A benign diagnosis was made in 74.9% (<i>n</i>=70,902), the remainder had cancer. Route of surgery was open for 28.3% (<i>n</i>=26,843) and MIS for 71.7% (<i>n</i>=67,865). Results stratified by indication and surgical approach are presented below. Open hysterectomy for malignancy: Blood transfusion was the largest contributor to EOD (PAR: 23.5% [95% CI: 15.5-32.0%]), PLOS (30.7% [27.0-34.4%]), and unplanned ROR (19.0% [10.3-28.5%]). In other words, preventing blood loss requiring transfusion would be expected to prevent 23.5% of EOD. Surgical site infection (SSI) had the highest impact on unplanned re-admission (11.5% [8.8-14.1%]). The number of 30-day deaths was too low for multivariable analysis, but univariate analysis identified DVT, cardiovascular events, transfusion, anastomotic leak, and sepsis as significantly associated with an increased risk of death. MIS hysterectomy for malignancy: Anastomotic leaks were the largest contributor to EOD (7.8% [3.3-16.9%]), while sepsis had the largest impact on unplanned ROR (9.8% [5.2-17.5%]) and unplanned re-admission (5.2% [3.4-7.0%]). Cardiovascular events and transfusion were associated with an increased risk of 30-day mortality in univariate analyses. Hysterectomy for benign indication: Blood transfusion was the largest contributor to unplanned ROR and PLOS for patients undergoing benign hysterectomy, irrespective of surgical approach. For women undergoing open hysterectomy, incisional SSI had the highest impact on EOD (9.4% [4.8-16.1%]) and unplanned readmission (14.1% [11.2-16.9%]). <b>Conclusions:</b> Our analysis identifies key drivers of adverse clinical outcomes and excessive resource utilization. Reducing the need for blood transfusions and decreasing rates of surgical site infection are the two highest-priority targets for quality improvement efforts and would be expected to have a substantial impact on reducing rates of EOD, PLOS, unplanned ROR, and re-admission following hysterectomy.