Survival Before and After Direct Surgical Quality Feedback in a Population-Based Lung Cancer Cohort

Background: Surgical resection is the main curative modality for non-small cell lung cancer (NSCLC), but variation in the quality of care contributes to suboptimal survival rates. Improving surgical outcomes by eliminating quality deficits is a key strategy for improving population-level lung cancer survival. We evaluated the long-term survival effect of providing direct feedback on institutional performance in a population-based cohort. Methods: The Mid-South Quality of Surgical Resection cohort includes all NSCLC resections at 11 hospitals in four contiguous Dartmouth Hospital Referral Regions in Arkansas, Mississippi, and Tennessee. We evaluated resections from 2004 to 2013, before and after onset of a benchmarked performance feedback campaign to surgery and pathology teams in 2009. Results: We evaluated 2,206 patients: 56% preintervention (pre-era) and 44% postintervention (post-era). Preoperative positron emission tomography/computed tomography (46% vs 82%, p \u3c 0.0001), brain scans (6% vs 21%, p \u3c 0.0001), and bronchoscopy (8% vs 27%, p \u3c 0.0001) were more frequently used in the post-era. Patients had 5-year survival of 47% (44% to 50%) in the pre-era compared with 53% (50% to 56%) in the post-era (p = 0.0028). The post-era had an adjusted hazard ratio of 0.85 (95% confidence interval [CI], 0.75 to 0.97; p = 0.0158) compared with the pre-era. This differed by extent of resection (p = 0.0113): compared with the pre-era, the post-era adjusted hazard ratio was 0.49 (95% CI, 0.33 to 0.72) in pneumonectomy, 0.91 (95% CI, 0.79 to 1.05) in lobectomy/bilobectomy, and 0.85 (95% CI, 0.63 to 1.15) in segmentectomy/wedge resections. Conclusions: Overall survival after surgical resection improved significantly in a high lung cancer mortality region of the United States. Reasons may include better selection of patients for pneumonectomy and more thorough staging