Independent entity’s medical billing data ruled inadmissible

Trial judges did not err in declining to admit medical billing data compiled by the independent nonprofit FAIR Health under two arguably applicable exceptions to the hearsay rule, separate District Court Appellate Division panels in the Northern and Southern districts in Massachusetts have found.

Other judges have taken an opposing view of FAIR Health data, allowing it to be considered at trial. As a result, there is no end in sight to the admissibility battles, attorneys say.

The FAIR Health database was created in response to an investigation by the New York attorney general. That investigation forced a number of the country’s largest insurance companies to abandon the use of a similar but less comprehensive database created and maintained by a wholly owned subsidiary of one of those companies. As with its predecessor, the FAIR Health database purports to identify the usual charge for similar medical services in a geographic area.

In the two recent cases, the insurer decided to pay less than the full amount that the providers had billed, deeming its payments “fair and reasonable” after consulting the database. The providers then sued to recover the balances they believed they were due.

FAIR Health states its case

In an amicus brief submitted in the 2013 Massachusetts Supreme Judicial Court case N.E. Physical Therapy Plus, Inc. v. Liberty Mutual Insurance Company, FAIR Health argued that its database passed the test for admissibility under G.L.c. 233, §79B, in matters in which the reasonableness of health care charges, or the range of usual and customary charges, are at issue.

FAIR Health touted its “sound statistical methods,” which were developed by a team of disinterested experts based at Syracuse University.

In part, FAIR Health argued that its database is too big to be unreliable. As of five years ago, it contained billing information related to more than 15 billion medical and dental procedures.

At the time, the database captured medical and dental services paid for by plans covering more than 129 million Americans, or approximately 70 percent of those covered by employer-provided private health insurance plans, according to FAIR Health.

FAIR Health noted that insurers, providers and government agencies had come to rely on “compiled” data sources such as its database “precisely because there is no equally reliable and cost-effective alternative.”

The brief continued: “If actors in the healthcare industry must and do depend on compilations of data like FAIR Health’s, then it would make little sense to exclude information from the same sources when marketplace disputes are submitted for resolution in the legal system.”

In both cases, the trial judge decided the data qualified for neither the hearsay exception for “statements of fact published for persons in a particular occupation” under G.L.c. 233, §79B, nor for business records under G.L.c. 233, §78. The Appellate Division has now validated those decisions.

The Northern District panel found that the trial judge had reasonably relied on the findings of a Salem District Court judge in a previous case.

The Salem judge wrote that FAIR Health “did not offer sufficient evidence to detail: an objective gathering of data; assurances that the data gathered was neither biased nor self-serving; the exact methodology for its calculation; nor assurance that the data gathered, in the states from which it is gathered, has created a set of facts relevant to [the case].”

In the Southern District case, the trial judge had appropriately determined that the FAIR Health data “was not relevant to the question of whether [the provider’s] charges were reasonable under G.L.c. 90, §34M, or alternatively, if marginally relevant, had the potential to mislead or confuse the jury,” the panel wrote.

The decisions are Lomibo, LLC v. Vermont Mutual Insurance Group and Patriot All Pro Physical Therapy Centers, Inc. v. Vermont Mutual Insurance Group.

Suspect sources?

It is not the first time state courts in Massachusetts have had to grapple with the reliability of a massive compilation of medical charges. In its 2013 decision in N.E. Physical Therapy Plus, Inc. v. Liberty Mutual Insurance Company, the Supreme Judicial Court identified flaws with a database maintained by Ingenix, a wholly owned subsidiary of United Health Group, one of the country’s largest health insurers.

With its decisions, the Appellate Division is implicitly recognizing that many of those flaws remain, despite the fact that FAIR Health is an independent nonprofit, said Salem lawyer Matthew T. LaMothe, who represented the plaintiff in Lomibo. LaMothe’s colleague, Brian P. McNiff, represented the plaintiff in Patriot All Pro.

In an amicus brief submitted in N.E. Physical Therapy Plus, FAIR Health said its founding principle was to “bring fairness and transparency to health insurance information.” That goal has not been realized, according to LaMothe, because FAIR Health relies on voluntary submissions from health insurers and does not include any information from auto insurers or medical providers.

“Generally, the sources of the data are suspect,” he said, adding that auditing of the data is lacking.

Rather than continue to try to rely on FAIR Health data, the Massachusetts PIP statute, G.L.c. 90, §34M, permits insurance companies to hire their own medical providers to review bills, LaMothe noted. The Legislature apparently preferred that alternative to a reliance on “hearsay a few times over” available through the database, LaMothe said.

The FAIR Health database was created in response to an investigation by the New York AG that forced a number of the country’s largest insurance companies to abandon the use of a similar database created and maintained by a wholly owned subsidiary of one of those companies.

Newton attorney Francis A. Gaimari, who represented the plaintiff in N.E. Physical Therapy Plus, said the recent Appellate Division decisions are a natural consequence of not just his case but its 2008 precursor, Davekos v. Liberty Mutual Insurance Company, decided by the Northern District Appellate Division.

Given the state of the law, Gaimari said he was surprised that some insurance companies are still fighting to introduce aggregated data on the reasonableness of medical charges.

“It’s almost like it has a cult following,” Gaimari said of the data.

But Milton attorney Jason J. Mellon, who represented Vermont Mutual in the two recent cases, said it is hardly a settled issue and pointed to decisions from 10 judges across the state finding FAIR Health data admissible after full evidentiary hearings.

For example, on May 2, Springfield District Court Judge William P. Hadley found in Valley Chiropractic & Rehabilitation, LLC v. Metropolitan Property and Casualty Insurance Company that the FAIR Health data met the requirements of G.L.c. 233, §79B, and was “distinguishable” from the Ingenix database considered in N.E. Physical Therapy.

Hadley added: “As to the question of trustworthiness, I am satisfied that Fair Health has sufficient safeguards in place to justify an imprimatur of reliability, which is demonstrated by the fact that multiple agencies from Florida to Alaska rely upon and use its data.”

But in comparing the Ingenix and FAIR Health data, Gaimari said he saw “no practical difference” in terms of transparency. He added that there also is no practical way to verify the data, even through the normal discovery process.

Boston attorney Marc L. Breakstone said he, too, thinks there are too many unknowns with FAIR Health data.

“There was no foundation provided for any indicia of reliability that would exempt the data from the usual bar on the admissibility of hearsay,” he said.

According to Gaimari, the last word on the subject of whether insurance companies can use FAIR Health data should be the Appeals Court’s 2014 decision in Hartunian v. Pilgrim Insurance Company.

In Hartunian, the court held that the insurance company’s use of a computer program to assess the reasonableness of a doctor’s charges not only did not comply with the clear requirements of G.L.c. 90, §34M, but its use as a substitute for a practitioner’s review of billing statements and underlying services provided an additional basis for an inference of the insurance company’s lack of good faith under Chapter 93A.

Rather than get into the weeds over whether the database was accurate or not, Gaimari said Hartunian stands for the proposition that it is “still not statutorily permissible” to use FAIR Health data.

He acknowledged that G.L.c. 90, §34M, enacted in 1970, pre-dates sophisticated computer-aided statistical analysis. But it is the Legislature’s job — not the courts’ — to revise the statute, he said.

Upon further review

Plaintiff Brittany Waugh sought and received physical therapy from Patriot All Pro after an Oct. 11, 2011, car accident. After treating Waugh for two months, Patriot submitted a bill to Vermont Mutual for $2,845.

Vermont Mutual paid $2,099.11 on the claim, which according to the FAIR Health database represented the 80th percentile of the cost of comparable services in the same geographic area.

Patriot sued to collect the balance of $745.89, along with costs and attorneys’ fees.

The stakes were similarly modest in the Lomibo case, in which plaintiff Lomibo LLC, which operated under the name All State Pain Treatment and Therapy Center, sought payment from Vermont Mutual of $3,628, and Vermont Mutual paid Lomibo $579.43 less after reviewing the relevant FAIR Health data.

In Patriot All Pro, the battle over the admissibility of FAIR Health data was fought in the context of a motion in limine Patriot filed, anticipating that Vermont Mutual would seek to introduce the data at trial.

In Lomibo, there was a dispute over whether the parties had stipulated to the admission of the FAIR Health data. The Appellate Division determined there had been no stipulation but rather that the trial had been conducted with the understanding that the evidence was being offered on a de bene, or provisional, basis.