Variation in Surgical Spending Among the Highest Quality Hospitals for Cancer Surgery

Abstract

Objective: This study evaluates the variation in spending by the highest-quality hospitals performing complex cancer surgery in the United States.

Summary background data: As mortality rates for high-risk cancer surgery have improved, increased attention has focused on other elements of quality, such as complications. However, high-value surgical care requires both high-quality care and cost savings. Therefore, understanding any residual cost variation among high-quality hospitals is essential to better direct efforts to achieve efficient, high-value care.

Methods: Medicare beneficiaries age 65 to 99 who underwent surgery for pancreas, esophageal, lung, rectal, and colon cancer from 2014 to 2016 were identified. The highest-quality hospitals were identified as those in the quintile with the lowest risk- and reliability-adjusted serious complication rates for each operation. Within this cohort of high-quality hospitals, 30-day total episode, index hospitalization, physician, postacute care, and readmis-sion spending were analyzed. Logistic regression models were utilized to estimate the probability of postoperative outcomes and post-discharge resource utilization.

Results: A total of 43,007 Medicare patients underwent either pancreas, esophageal, lung, rectal, or colon resection for cancer at a hospital within the highest-quality quintile. Among the highest quality hospitals, total episode spending ranged from $18,712 for colectomy to $38,054 for esophagectomy. Spending between the lowest- and highest spending hospitals varied from $1207 [confidence intervals (CI 95% ) $1195-$1220] or 6.6% of total episode spending in the lowest tertile for colectomy to $5706 (CI 95% $5,506-$5906) or 16.1% of total episode spending in the lowest tertile for esophagectomy. The largest component of variation was from postacute care spending followed by readmission. For all operations, the risk-adjusted rate of postacute care facility utilization was lower among the lowest spending hospitals compared to the highest spending hospitals. For example, for pancreas the lowest-spending hospitals on average discharged patients to a postacute care facility at a rate of 18,6% (CI 95% 16.2-20.9) compared to 31.0% (CI 95% 28.2-33.9) in the highest-spending hospitals. In all operations, the risk-adjusted readmission rate was lower among the lowest-spending hospitals compared to the highest-spending hospitals. For instance, within the esophagus cohort, the lowest-spending hospitals had an average risk-adjusted readmission rate of 17.3% compared to 29.4% in the highest spending hospitals ( P < .001).

Conclusions and relevance: Even among the highest-quality hospitals, significant cost variation persists among cancer operations. Postacute care variation, rather than residual variation in complication rates, explains the majority of this variation and represents an immediately actionable target for increased cost-efficiency.