Tiered Cost-sharing at the Clinic Level: Consumer and Provider Responses to Tiering

Many health care reform initiatives can be classified as either supply-side or demand-side initiatives. Supply-side initiatives consist largely of provider payment reform, while demand-side initiatives focus on consumer incentives. Demand-side initiatives such as reference pricing and tiered cost-sharing that result in consumers choosing one provider over another have the potential to elicit a significant response from suppliers because the entire revenue stream from the consumer is at risk.

We analyze data from a tiered cost-sharing system administered by the State Employees Group Insurance Program (SEGIP) for employees of the State of Minnesota and their dependents (127,000 covered lives). The SEGIP cost-sharing system has been in place for fifteen years and to our knowledge, provides the longest individual-level tiered cost-sharing panel dataset in the nation.

In the SEGIP system, employees and their dependents choose a primary care clinic which refers and is attributed their care from hospitals and specialists. Based on past total average annual risk-adjusted cost of care for its members, primary care clinics are placed in one of four cost-sharing tiers for the coming year. Average risk-adjusted cost of care among clinics in tier 4 (the highest cost tier) are roughly double those in the tier 1 (the lowest cost tier).

Cost-sharing differentials among the tiers are substantial. Annual family deductibles in 2016 ranged from $300 to $2,500 and office visit copays varied from $30 to $85. In 2016, 88 percent of SEGIP members were in the lower two tiers.

Our quantitative analysis estimates the effect of tier membership on consumer choice of clinics. To address the possibility of omitted variable bias, we take advantage of a unique natural experiment – SEGIP’s reassignment of clinics from tiers 3 or 4 to tier 2 in order to assure that all members have access to a tier 2 clinic. When a clinic is reassigned, nothing changes except the clinic’s tier. Because a clinic that is reassigned to a lower tier in one zip code is reassigned to the lower tier in all zip codes, we are able to isolate the “pure” effect of tiering by examining the consumer’s response in zip codes where a tier 3 or tier 4 clinic was adjusted to tier 2, but whose residents already had a tier 2 clinic available in their zip code. Our preliminary results suggest a modest pure effect of tiering on consumer choice of clinic.

Clinics can move to a lower cost tier by reducing unnecessary utilization, agreeing to a fee reduction, or participating in shared savings. In 2016, 700 of the 2,786 plan-clinic choices had voluntarily reduced their fees and 12 clinics were participating in shared savings. Our qualitative analysis built a conceptual model of clinic behavior in response to participation in a tiered insurance network. Providers were found to be sensitive to both the reputational effects of tier placement and to any real or perceived changes in patient volume.