Do Within-Practice Referrals for Shoppable Services Contribute to Spending Variation?

Patients heavily rely on clinicians’ referral recommendations when choosing other clinicians and facilities (hereafter, providers) for shoppable services—generally defined as non-urgent encounters entailing generally consistent processes. Clinicians disproportionately refer to other providers in their practice, within-practice referrals, because of greater awareness of those providers and the practice’s financial incentive to generate business. Practices may consistently negotiate relatively high or low reimbursements from insurer(s) for various services, so private insurance enrollees may spend different amounts for a set of services if they disproportionately use certain practice(s), e.g., because of within-practice referrals. This feature of price-varying service markets may explain why primary care physician (PCP) office visit price strongly predicted spending variation in a recent study: patients attributed to low- or middle-price PCPs ($60 and $65 average office visit prices) had $690 (8.5%) and $462 (5.7%) lower one-year insured plus out-of-pocket spending, respectively, than patients attributed to modestly more expensive, high-price PCPs ($86 average price). The study’s data did not identify providers or practices, so did not permit investigating within-practice referrals.

Objective:

To assess if within-practice referrals explain the correlation between referring clinician office visit price and spending.

Methods:

Using claims, referrals, and related administrative health plan data for ~100,000 adult, non-Medicare private insurance enrollees residing in the Pacific Northwest each year from 2012-2017, we will identify price shopping opportunities defined as an enrollee having an ambulatory office visit that resulted in a referral for a shoppable service—any of 114 service encounters with price estimates in the enrollees’ price transparency tool (selected specialist office visits, advanced imaging, and outpatient and inpatient procedures). We will identify within-practice referrals using practices’ tax ID numbers. We will estimate two-part regressions for utilization and spending amounts given utilization. Spending outcomes are service encounter and plan-year (calendar-year) spending, each by out-of-pocket, insured, and total amounts.

The shoppable service encounter spending model will include main effects and interactions for the referring clinician’s office visit price and the occurrence of within-practice referrals. The plan-year spending model will include main effects and interactions for average office visit price of the attributed referring clinician (for most enrollees, the clinician for >50% of office visits) and a composite measure of within-practice referral intensity weighted by the average cost of each referred service in the year. Covariates include patient demographics; insurance benefits; comorbidity; and fixed effects for individual services, year, and geographic location.

Conclusions:

Presented at conference. We hypothesize that within-practice referrals increase (decrease) service encounter spending and plan-year spending for patients of referring clinicians with high- (low-) office visit prices, which will eliminate the correlation between referring clinician office visit price and later spending. In that case, insurers might reduce spending by steering patients to ambulatory clinicians employed by less expensive practices.