The Anatomy of Physician Payments: Contracting Subject to Complexity

The reimbursement rates that private insurers pay to physicians are closely linked to those set by Medicare, despite the well-known limitations of Medicare's fee schedule.  We ask to what extent this relationship reflects the use of Medicare's relative price menu as a benchmark, in order to reduce transaction costs in an otherwise complex pricing environment.  We analyze 71 million claims from a large private insurer, Blue Cross/Blue Shield of Texas (BCBSTX).  These claims represent $6.3 billion in spending over three years, and allow us to carefully estimate the pervasiveness of payments linked directly to Medicare's fee schedule.

We use two empirical strategies to measure the pervasiveness of links between Medicare's reimbursement rates and payments from BCBSTX to the physicians who treat its beneficiaries.  Our first strategy exploits the fact that the payment for any service can be decomposed into its Medicare-allotted number of RVUs times a scaling of dollars per RVU, which we call the “implied conversion factor” (ICF).  If many services share a common ICF, this might plausibly reflect a contract specifying this particular ICF.  Payments that follow the Medicare benchmark, as evidenced by sharing common ICFs, comprise around three quarters of BCBSTX's claims and account for two thirds of spending.

Second, we measure how often changes in Medicare's payment rates pass through to BCBSTX's payments.  Using institutional knowledge of the precise dates on which BCBSTX implements Medicare's annual reimbursement updates, we infer the share of BCBSTX's payments linked to RBRVS.  This fine-grained timing helps eliminate confounding from long-run technological changes or active contract renegotiations.  Consistent with the previous results, we again find that around three quarters of BCBSTX's payments are linked to Medicare.

Using both strategies, we examine heterogeneity in price benchmarking across physician groups and service categories.  We find that payments to relatively large firms are less tightly linked to the RBRVS than payments to small firms.  Looking across service categories, payments are more closely linked to Medicare's relative values for services where Medicare's average-cost pricing system gets closer to the marginal costs of care.  Thus we see more benchmarking for labor-intensive services, like standard office visits, than for capital-intensive services, like diagnostic imaging.

Finally, we show that adjustments relative Medicare's price menu narrow likely gaps between marginal costs and Medicare's average-cost payments. Specifically, payments for labor-intensive services tend to be adjusted up while payments for capital-intensive services tend to be adjusted down. 

Our results suggest that physician contracts are written to manage the tension between gains from fine-tuning payments and costs from making contracts complex. The benefits of fine-tuning payments are greatest for contracts with large physician groups, which explains why such contracts deviate more often from Medicare's rates.  Since Medicare's average cost approach will have greater difficulty managing the payments for capital- than labor-intensive services, the benefits of adjustment will tend to be greater for the former.  Providers and private insurers appear to coordinate around Medicare's menu of relative payments for simplicity, but innovate when the value of doing so is likely highest.