Abstract

Using a large data set, Lucas et al 1 studied the importance of appropriate risk adjustment when applying performance measures in payment and accountability programs. Despite limited variation in readmissions for colorectal cancer, a readmission rate of 12.3% should encourage all of us to strive to get below 10%. It is about improvement and accountability in payment and public reporting. The hierarchical logistic regression model used found variations in age, race, procedure, and comorbidities. Risk-adjusted readmission odds ratios revealed risks shared by all hospitals in these specific groups of variables. They also point to areas in need of focused improvements. The authors shed light on the concerns of raw data for readmissions, an approach that lends itself to misclassification of delivery systems and misguiding patients with the information if the analytics do not properly use risk adjustment and CIs. Lucas et al 1 emphasize the importance of hierarchical regression in thoughtful design of analytics to ensure “apples to apples” comparisons. The authors note that results from the Centers for Medicare & Medicaid Services (CMS) hierarchical model have inconsistencies in applying the results to payment programs when compared with the CMS public reporting programs. It seems that payment programs act without regard to CIs, whereas the CMS public reporting programs are respectful of their statistical effect. I take exception to the authors when they suggest a surgeon- and specialty-specific risk-adjusted readmission rate as a solution. Surgical practice profiles vary greatly even within general surgery, such as in the subspecialties of breast, bariatrics, trauma, hepatopancreaticobiliary, transplantation, pediatrics, and so forth. Patients need condition-, disease-, and procedure-specific information tailored to their informed decision needs. Focusing on the specific procedure with “fit for purpose” performance measures will drive health care closer to sensible accountability and will best propel improvement efforts.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call