Specialist Services under a Patient‐Centered Medical Home Program

Patient-centered medical homes (PCMH) have been developed by a variety of payers and while they share selected structural features, they also provide a rich set of natural experiments in primary care payment design.  In particular, PCMH models vary in how financial incentives are structured, whether payment is tied to overall financial performance or not, whether it is based on per member monthly fees, fee-for-service payments, or capitation, and whether financial risk is two-sided or one-sided. Moreover, PCMHs differ in whether physicians are provided “transparency” tools with which to manage expenditures and referrals.  While studies generally have found PCMH interventions to lower costs and reduce emergency department and inpatient care, few have examined the impact of PCMH on specialty care specifically. An evaluation of an ACO-like model, the Massachusetts Alternative Quality Contract, found reductions in outpatient care and tied these to lower spending in facility-based care, procedures, imaging, and tests.

A recent PCMH model included two key elements predicted to lower spending on specialty care. First, it rewarded primary care physicians whose attributed members achieved lower annual total spending growth than projected by the payer and who met quality targets.  This one-sided risk model rewarded practices that achieved savings and quality targets, but did not withhold payments from those that did not.  Second, it provided each practice an electronic portal or transparency tool through which physicians could monitor their overall performance and receive timely data on costly services, conditions, and patients relative to their peers.  Importantly, through this tool primary care physicians had the ability to identify the most efficient specialists in the payer’s network.

In the current study we examine the impact of this PCMH model on specialty spending through a variety of approaches. We examine selected specialties in aggregate, followed by referral-sensitive conditions, and finally so-called tracer conditions, i.e., high-cost elective conditions. Because savings in specialty care can arise from reductions in utilization, shifts to less expensive settings, or referrals to lowered priced providers we attempt to identify the source of savings in each case. 

We use a staggered difference in difference strategy with propensity score matching to compare trends for adults who were attributed to a PCMH provider to adults in a comparison group who did not participate in the PCMH program.  Moreover we test whether there were differences in performance by physician organizational type (i.e., single practice, multi-practice, or hospital-based). We discuss our results in light of recent physician payment reform proposals.