In a Business Review Letter dated April 26, 2010, the U.S. Department of Justice, Antitrust Division (hereinafter, “the Division”) stated that it would not challenge a data sharing program proposed by three health care associations located in California. From the Division’s vantage point, the program posed little risk of facilitating anticompetitive conduct; rather, the most likely effect of the program would be to increase transparency about the relative costs and utilization rates of hospitals that participated in the survey.
The Proposed Data Exchange Program
The Pacific Business Group on Health (“PBGH”), California Public Employees’ Retirement System (“CalPERS”) and California Health Care Coalition (“CHCC”) proposed a data sharing program called the Hospital Value Initiative (“HVI”) to measure hospital cost of care and resource use. Through a third-party consultant, Milliman, Inc., the HVI would collect claims data for each service provided by the approximately 330 participating California hospitals from five health insurance providers comprising 67% of all health insurance enrollment in the State.
After compiling data on a hospital’s “Allowed Amount”—that is, the total payment a hospital receives from a provider and its insured—as well as the level and type of bed-days for a given service over the course of a year, Milliman would calculate Buyer Cost Index (“BCI”) and Resource-Use Efficiency (“RUE”) ratios. The All-Payor BCI would allow a hospital to gauge its Allowed Amount from all five providers and their insureds for a specific procedure against the average Allowed Amount paid to the other participating hospitals. In contrast, the Payor-Specific BCI would compare the Allowed Amount received from a specific provider and its insureds against the average Allowed Amount paid to the other hospitals participating hospitals. The RUE ratio would focus on inpatient services, comparing the number and type of bed-days for a specific procedure that the hospital performed for all providers providing such coverage against the average number of bed days for all other participating hospitals performing that service.
Using a secure Web site, Milliman would deliver the scores to hospitals, group purchasers and health insurance providers participating in the HVI, with each constituency receiving a different type of report. Each hospital could obtain All-Payor BCI and RUE reports for its own procedures in addition to a regional BCI score, which would calculate the average BCI scores for participating hospitals located in the same region. Group purchasers—employers, public agencies, unions and health and welfare funds that belong to PBGH, CalPERS and CHCC—would also have access to the All-Payor reports. Each of the five health insurance providers would be able to review its own BCI and RUE scores as well as those for each of the 330 participating hospitals. Given the time and resources necessary to produce the reports, at least 10 months would elapse between collection of the data and distribution of the results.
The HVI program imposed safeguards to prevent hospitals and health insurance providers from accessing their competitors’ data. For example, if a provider owns or acquires a hospital, then that provider could not access All-Payor reports. Similarly, each participating hospital could review its own data, but would not be allowed to see other hospital reports or any provider-specific report. The HVI also pledged to implement an antitrust compliance program.
The Division’s Antitrust Analysis
The Division’s analysis focused on whether the proposed data sharing program “is likely to facilitate express or tacit collusion resulting in increased prices or reduced quantity or quality or otherwise reduce competition among the recipients of the data.” To evaluate the potential anticompetitive effects, if any, the Division applied the criteria articulated in Statement 6 of the Department of Justice’s and Federal Trade Commission’s Statements of Antitrust Enforcement Policy in Health Care (1996). As set forth therein, an agency will first decide whether the proposed data sharing program falls within an antitrust “safety zone” that, absent extraordinary circumstances, shields the proposal from further scrutiny. If the program does not meet the safety zone criteria, then the agency will use a balancing test to ascertain whether the program generates anticompetitive effects that outweigh any procompetitive justifications. Specifically, the Division first considered whether
(i) The survey is managed by a third party;
(ii) The information provided by the survey participants is based on data more than 3 months old; and
(iii) There are at least five providers reporting data upon which each disseminated statistic is based, no provider’s data represents more than 25% on a weighted basis of that statistic and any information disseminated is sufficiently aggregated such that it would not allow recipients to identify the prices charged by any particular provider.
Although the HVI program fulfilled the first two criteria, the Division was not certain that the third provision would be satisfied with respect to hospital services for which a provider’s data, or lack thereof, disproportionately skewed the All-Payor BCI ratio.
Accordingly, the Division next considered whether the anticompetitive effects, if any, of the program outweighed its procompetitive justifications. The Division’s answer was a resounding no. The likelihood of collusion among competitors was low given that the parties would not have access to raw claims data and the reports would not divulge the prices that hospitals charged for specific services. Equally significant, the aggregate presentation of data made it virtually impossible to “reverse engineer” the BCI and RUE scores and indirectly glean information about a competitor’s prices. On the other hand, the data sharing program offered significant procompetitive benefits by facilitating more informed purchasing decisions and fostering competition among hospitals. However, the Division noted that the parties’ representations about the anticipated high volume of HVI participants was a critical assumption of its analysis, and its conclusion could change if the actual number of participants was materially lower than projected.