How does provider network size influence health care utilization?


Are narrow networks a good thing?  On the one hand, they may be able to reduce health insurance premiums if insurers can negotiate lower prices.  On the other hand, narrow networks means less choice of providers for patients. 

An less well-understood question is how narrow networks affect health care costs and utilization.  Empirically, broader networks have higher cost and utilization.  However, this phenomenon is likely driven by adverse selection; broader networks are more attractive to sick individuals who are likely to use more health care services. 

To better understand how network size affects health care utilization and cost, Boone (2019) creates a model with homogenous consumers, providers and insurers. Consumers are risk avers and buy insurance with a weakly positive premium.  Insurers are assumed to pay providers using a two-part tariff approach, which includes a capitation payment for each patient the provider is responsible for and a fee-for-service payment for each service provided.  Assume the capitation fee is t, the fee-for-service is  p, and the cost for the provider to provide the service is c.  Insurers who want to incentivize providers to reduce cost will set the fee-for-service price such that each service is marginally unprofitable (i.e., p) and the capitation fee t makes up the difference so that the provider is solvent at year end.  This approach, however, only works for providers with patient lists (e.g., general practitioners) whereas secondary and tertiary care providers (e.g., hospitals) will need full fee-for-service payments that are profitable individually (i.e., p>c).

Using this framework—and a large number of other
assumptions I will not get to in this summary, the authors conclude the
following.  First, increasing network
size leads to relatively more FFS and less capitation payment.  The authors claim that with a large network, capitation
contracts (low fee-for-service, high capitation) become too expensive.  Why is this? 
In large network, providers are at risk for many sick patients visiting
them and thus will demand a high capitation rate t. With few providers in an area, a provider’s patient base is more
well known and there is less risk of adverse selection. 

The authors also find that with private
contracts—i.e., individual insurer-provider contract terms are not revealed to
other providers or consumers—providers make positive profits.

Third, insurers with broad provider networks
have higher health care costs and more utilization.  A broad network is a signal of general
insurance to providers in that a higher share of compensation is likely to
occur FFS (p) rather than capitation
(t). Consumer believe these plans
have providers who are more likely to treat patients more aggressively. 

The findings are interesting but there are a
number of short-comings to the model.  Principally,
the limitation is that insurers and providers are heterogeneous.  This means that insurer or provider network
size do not affect negotiations. 
Further, there is no difference in quality across providers. 

Despite these limitations, the finding that network size not only affects negotiated prices but also health care utilization levels, is an interesting theoretical result.