Variation in Bariatric Surgery Episode Costs in the Commercially Insured: Implications for Bundled Payments in the Private Sector.

OBJECTIVE

To describe hospital-level variation in roux-en-Y gastric bypass (RYGB) and sleeve gastrectomy (SG) in Michigan.

BACKGROUND

Bariatric surgery is an increasingly prevalent elective surgical procedure that will likely be considered for future bundled payment programs, both public and private. Past research in the Medicare population found that the index hospitalization is responsible for the majority of payment variation among hospitals. However, this research largely excluded SG, now the most commonly performed bariatric surgery procedure nationally.

METHODS

We used data from a state-wide quality collaborative to calculate the average risk and price-adjusted 30-day episode payment for patients undergoing RYGB and SG procedures at Michigan hospitals between January 2009 and October 2014. We organized hospitals into quintiles and compared the variation in payments between highest and lowest-cost quintiles, and also the payment categories that drove this variation.

RESULTS

We identified 9035 patients undergoing RYGB (n = 4194) or SG (n = 4841) procedures at 31 hospitals. The average price and risk-adjusted episode payment ranged from $11,874 in the lowest hospital quintile to $13,394 in the highest quintile, representing a difference of $1519 (12.8%). Payments for the index hospitalization accounted for the largest share of total episode costs for both procedure types. Despite representing 2.7% to 6.0% of payments across quintiles, postdischarge payments explained 22.6% of hospital variation in SG. Similarly, readmissions explained 24.5% of payment variation for SG episodes, despite representing between 1.2% and 4.4% of payments.

CONCLUSIONS

Collectively, our findings suggest that there are previously underappreciated differences in episode payment variation between bariatric surgery procedures. SG may be more amenable to cost containment under bundled payment initiatives by virtue of the greater share of variation explained by readmission and postdischarge payments, components of episode payment more likely to be influenced by provider discretion.