Assessment of emergency general surgery care based on formally developed quality indicators.

BACKGROUND: Emergency general surgery outcomes vary widely across the United States. The utilization of quality indicators can reduce variation and assist providers in administering care aligned with established recommendations. Previous quality indicators have not focused on emergency general surgery patients. We identified indicators of high-quality emergency general surgery care and assessed patient- and hospital-level compliance with these indicators.

METHODS: We utilized a modified Delphi technique (RAND Appropriateness Methodology) to develop quality indicators. Through 2 rankings, an expert panel ranked potential quality indicators for validity. We then examined historic compliance with select quality indicators after 4 nonelective procedures (cholecystectomy, appendectomy, colectomy, small bowel resection) at 4 academic centers.

RESULTS: Of 25 indicators rated as valid, 13 addressed patient-level quality and 12 addressed hospital-level quality. Adherence with 18 indicators was assessed. Compliance with performing a cholecystectomy for acute cholecystitis within 72 hours of symptom onset ranged from 45% to 76%. Compliance with surgery start times within 3 hours from the decision to operate for uncontained perforated viscus ranged from 20% to 100%. Compliance with exploration of patients with small bowel obstructions with ischemia/impending perforation within 3 hours of the decision to operate was 0% to 88%. For 3 quality indicators (auditing 30-day unplanned readmissions/operations for patients previously managed nonoperatively, monitoring time to source control for intra-abdominal infections, and having protocols for bypass/transfer), none of the hospitals were compliant.

CONCLUSION: Developing indicators for providers to assess their performance provides a foundation for specific initiatives. Adherence to quality indicators may improve the quality of emergency general surgery care provided for which current outcomes are potentially modifiable.