Innovations in Hospital Merger Simulation

Economists Ink: A Brief Analysis of Policy and Litigation

View Argue’s Profile

View Shin’s Profile
David A. Argue is a Principal in EI’s Washington, DC office. He has analyzed competition in health care areas, including hospitals, physician practices, ambulatory surgery centers, health insurance, and pharmaceuticals. He has also taught health economics at Johns Hopkins University. Richard T. Shin is a Senior Vice President in EI’s Washington, DC office. He has experience in demand estimation and merger simulation in various industries including health care . A more detailed version of this article appears in the Fall 2009 issue of Antitrust.

Despite a lengthy history of failures to block hospital mergers in court, the antitrust agencies continue to be interested in their competitive implications. Among the insights that come from recent FTC investigations of hospital mergers is that the FTC staff has developed a new form of merger simulation model. The new model has, for the most part, not been publicly aired by the FTC staff, but its basic structure is evident. The merger simulation model closely follows the theoretical “two-stage” theory of hospital competition embraced by the FTC in its retrospective review of the Evanston Northwestern-Highland Park Hospital merger, but it uses a novel approach for conducting the simulation. If the new merger simulation approach withstands scrutiny, it will be a very powerful enforcement tool because it gives direct, though simulated, evidence and does not depend upon the often controversial issue of market definition.

In the first stage of the two-stage theoretical model used by the FTC staff, hospitals and health plans negotiate over the price at which a plan will accept a hospital into its network. Logically, hospitals and health plans arrive at prices that reflect the relative value added to the network by each hospital. In the second stage, enrollees who become ill must choose a hospital from among those in their previously chosen network, and they are assumed to choose entirely based on non-price factors (e.g., hospital’s location, services, reputation, staff physicians). Thus, while hospitals compete on price to join a network in the first stage, they compete on non-price attributes in the second stage to attract enrollees who have become ill. The key to the simulation is first to estimate consumer preferences for hospitals as revealed in their second-stage choice of hospital, then to use that information to estimate prices from the first-stage negotiation between hospitals and health plans.

In the bilateral price negotiations over the first-stage price, each hospital negotiates using bargaining power it possesses from offering incremental value to enrollees who choose the network. The amount of incremental value is affected by the alternative hospitals that are already in the network or that might be added to the network. A different type of hospital (e.g., a children’s hospital) may bring more incremental value to enrollees than would one of many community hospitals. The more readily available in-network substitutes are, the less incremental value another hospital adds, and the more constrained on price is the bargaining position of a new hospital being considered for the network.

The most interesting scenario from an antitrust perspective concerns hospitals that are similar to each other but different from other hospitals that are either in the network or are available to join the network. Because of their differentiation from the in-network hospitals, the incremental value to enrollees of including either hospital would be significant. Thus, either hospital likely is able to negotiate a price that is higher than its costs. Each one’s bargaining power is attenuated, however, by the ability of the health plan to choose its rival. A merger of these two hospitals increases the combined hospitals’ bargaining strength because it eliminates competition from a close substitute.

The incremental value a hospital brings to a network is reflected in the health plan’s willingness to pay to include that hospital. Willingness to pay (WTP) can be estimated econometrically with data on patient characteristics and hospital characteristics. To determine whether one configuration of a hospital network yields greater satisfaction to enrollees than another, WTP is calculated for each network configuration. In effect, calculating the difference in WTP is the basis for understanding how a hospital merger may enable the merged entity to acquire increased leverage over payors.

Once the incremental value of adding a hospital to a network or removing a hospital from a network (with or without replacement) is estimated, the price effect from the merger can be estimated from each payor’s claims data. Price is assumed to be a function of WTP, patient demographics, length of stay, discharge status, admit source, and other factors. The change in price due to the merger is calculated using the estimated relationship between price and WTP and the change in WTP induced by the merger.

The new approach to hospital merger simulation is conceptually compelling, but it is not without shortcomings. First, the model explicitly assumes away health plans’ use of financial incentives to steer patients among in-network hospitals. Instead, the model assumes that hospitals will either be included in or excluded from a network, and that plan members pay the same prices at each in-network hospital. Second, reflecting a problem that exists in much analysis of health care services, the simulation model does not incorporate quality improvements stemming from the merger. Third, the model does not account for the likelihood of entry or repositioning that may occur in response to an attempt to increase price or reduce quality from competitive levels. While a model need not explain everything to be useful, insofar as factors as important as quality changes or entry cannot be incorporated into the model, they must be evaluated outside of the model. Fourth, competition in the payor market is not explicitly incorporated in the simulation model. This omission is evident in the tension between the assumption that hospitals maximize profits but health plans maximize enrollees’ satisfaction given that payment from enrollees is fixed.

The FTC staff’s new hospital merger simulation model differs significantly from previous approaches. It takes advantage of the distinctive characteristics of hospital services markets in a creative approach to the long-sought goal of predicting price changes prior to a merger’s occurrence. The new simulation model has not yet undergone substantial public review and critique, but it holds promise as an innovative approach to an old problem. It remains to be seen, however, how significantly its shortcomings affect its usefulness.