For Release:
November 25, 2003

Contact: J. Robert Hunter, 703/528-0062
212/209-7666; 917/548-5263

THE MILLIMAN REPORT: FATALLY FLAWED

Milliman USA, consultants for the health care insurance industry, wrote an article in the November/December issue of Contingencies that purportedly shows that caps reduce insurance premiums. This report conflicts with a number of other studies, such as Weiss Ratings, an independent insurance-rating agency that found in June that over the last decade, states with caps saw median doctors� malpractice insurance premiums rise faster than states without caps.

Several findings in the report also conflict with American for Insurance Reform (AIR)�s own study, Stable Losses/Unstable Rates.

There are many points in the Milliman report that reflect various shortcomings in their data and analysis. Here are a few:

The Milliman figures are fatally flawed because they are not adjusted for medical inflation, and are only inconsistently adjusted for the steadily increasing number of doctors.

• In order to conclude that claims, losses, jury verdicts or any other tort system costs are having a significant impact on the overall costs for insurers - and are therefore the reason behind skyrocketing insurance rates - losses per doctor should be rising faster than medical inflation over time. Milliman�s failure to make these adjustments completely undermines its analysis. By comparison, AIR�s studies always make these required adjustments. With these adjustments made, AIR�s insurance industry data show paid losses stable since the mid-1980s.

• To show how misleading it is NOT to adjust figures for medical inflation, Milliman shows losses as up about 32 percent over the last decade in states without caps - but medical inflation is 51 percent for the same period. In other words, in real terms, the paid losses are falling in the non-capped states under the Milliman analysis.

The data sources that Milliman uses are either incomplete or unidentified.

• The source for Milliman�s claims data is the National Practitioner Data Bank, instead of more reliable insurance industry reports. It seems to be a strange choice for insurance industry consultants to skip the industry�s own data and instead use a source which the U.S. General Accounting Office, the American Medical Association, Physician Insurers Association of America, consumer advocates, and others have all complained is incomplete, inconsistent, and flawed.1 On the other hand, AIR only uses insurance industry data provided by A.M. Best Company for its analyses, which provides both claims and premium data.

• The data source used by Milliman for premiums is completely unidentified; NPDB does not collect premium data so it unknown what data Milliman is using here.

• The Milliman report claims to be comparing �capped� and �uncapped� states, but the authors never identify which states fall in each group, how it defines �cap,� and so on, providing absolutely no basis for reviewers to test the data or the analysis.

• The report fails to provide the actual data on which its charts are based, making it impossible for anyone to check for inaccuracies or inconsistencies. (Notably, AIR always supplies such data for its reports.) For example, Milliman�s charts indicate that the number of claims has been steady for a decade, but the text says �the number of claims . . . have grown since 1992, as fig. 1 clearly shows.� Without the �raw� data, the reader can�t figure out which is the mistake.

• The report relies on limited data from only Texas and Florida to find that �noneconomic damages compose over two-thirds of the total cost of claims payments.� An earlier report from the same authors, commissioned by the Florida Hospital Association, pointed out that the data for such a statement is extraordinarily limited,2 and may not be reliable even when looking just at those states. There does not seem to be any basis for believing that this questionable finding applies elsewhere, and is questionable even for those two states.

Milliman either ignores data that does not fit its conclusions, or attempts to explain them away by analysis that make no sense.

• According to Milliman, in recent years losses have been lower in capped states by about 50 percent on average � yet premiums have only been lower by about 27 percent. This suggests that insurers are "ripping off" doctors by not passing savings through to them. To avoid having to reach this conclusion, Milliman illogically asserts that if caps have not reduced claim costs and premiums, it is because other factors are responsible for keeping costs and premiums high, specifically loss adjustment expenses (�LAE�), the cost of adjusting claims. According to this logic, these factors would somehow need to rise faster, or be more powerful, in states with caps in order to offset those savings. This makes no sense. The opposite should be true. Further, the Milliman study claims that claim frequency is sharply down in capped states. This should lower, not raise, LAE costs.

• In making state comparisons, Milliman made no effort to hold equal other factors that might impact a state�s premium and loss results, like a state�s insurance regulatory law, whether a state had an elected insurance commissioner, and other differences between the states.

Despite all of these deficiencies, the Milliman report itself confirms many findings reached by consumer advocates: i.e., that the number of malpractice claims per physician have dropped steadily over the last decade, with a particularly steep drop in claims since 2001, and losses per physician dropped significantly since 2001.

The report also shows that the insurance industry raised doctors� premiums dramatically in the last two years, just while losses and claims were dropping.

###

___________________
1 See, e.g., GAO, �National Practitioner Data Bank: Major Improvements Are Needed to Enhance Data Bank�s Reliability,� Nov. 2000; Mary Jane Fisher, �GAO Report Slams National Practitioner Data Bank,� National Underwriter, Jan. 1, 2001.
2 Richard S. Biondi, Arthur Gurevitch, David S. Wolfe, Florida Hospital Association: Medical Malpractice Analysis, Nov. 7, 2002 (�Note that only about a quarter of the records in the [Florida] Archive database had payments broken out in these [economic and noneconomic] categories. . . . We assume that economic and non-economic values are initial estimates of what will ultimately be paid and not necessarily components of the actual paid loss.� �Similarly, only one third of the cases in the Texas DOI database have indemnity payments broken out into economic, non-economic, and �other� (exemplary/punitive damages and pre-judgment interest) categories.�)