J. Tong1, K. T. Collier1, S. Dasani1, R. Hoffman1, G. C. Karakousis1, R. R. Kelz1 1University Of Pennsylvania,Department Of Surgery,Phiadelphia, PA, USA
Background: A shift from the fee-for-service system to a value-based-purchasing model(VBP) is underway. In a GI cancer model, the study aimed to examine treatment patterns and observed costs in the in-patient setting to understand the impact that VBP may have on patients.
Methods: A retrospective cohort study was performed using inpatient claims from California and New York (2008-2011) for all patients admitted with a new diagnosis of colon, rectal, stomach, or pancreatic cancer regardless of the intended treatment plan. Patterns of treatment, surgical and non-surgical, and associated costs during the index admission for cancer were examined. Patient characteristics were compared by treatment pattern using the Chi square test. Univariate costs were compared using the Kruskal Wallis test. Log-transformed wage-adjusted total costs were compared using multivariate linear regression.
Results: Of 75,177 patients, there were 39,207(52.2%) colon cancer, 14,020(18.7%) rectal cancer, 12,341(16.4%) pancreatic cancer, and 9,609(12.8%) stomach cancer admissions. Patients admitted with a new diagnosis of colon cancer were most likely to be treated with surgery during the index admission(81.8%) compared to 61.3% of rectal cancer patients, 23.5% of pancreas cancer patients and 43.1% of stomach cancer patients(p<0.001). 34% of surgical patients underwent an operation within 24 hours of admission. The median cost per admission was $20,819(IQR 12,448-34,928). Cost was greatest for patients with stomach cancer($24,168; IQR 12,861-43,978) and least expensive for pancreas cancer patients($19,474; IQR 10,506-36,708) (p<0.001). Cost for patients who had surgery during the index admission($24,177; IQR 15,311-38,714) differed from that those who did not($15,351; IQR 8,767-27,673)( p<0.001). Among patients who underwent an operation during the index admission, the cost was significantly less for those who underwent surgery within 24 hours of admission($21,606; IQR 13,990-34,396) when compared to those that were admitted for >24 hours before surgery($27,360; IQR 17,251-43,801) (p<0.001). The direction of the relationship between cost and treatment type and timing of surgery remained the same after adjustment for potential confounders.
Conclusions: Surgical care remains the definitive therapy for solid GI cancers with variable cost depending on tumor type. Not all patients are receiving surgical treatment at the time of admission. Patients that are ready for surgery at the time of admission incur less costly care. The goal of bundled cancer care should be timely diagnosis and referral for definitive surgical care. Optimizing outpatient health status prior to admission may result in substantial cost savings. While VBP offers sensible features for cost containment for insurers and employers, the current inefficiencies in utilization of in-patient services will ultimately lead to greater costs incurred by the patient.