Statistical Analysis

One of the cases we consulted on was a medical malpractice action brought against a hospital. The decedent in this matter had contracted a severe streptococcus infection while he was a patient. Streptococcus and similar bugs abound in hospitals. Many such hospital-borne infections are specialized for survival in the antiseptic hospital environment, where they are subject to high-powered antibiotics and develop a stubborn resistance to treatment. The hospital treated the decedent for the infection, and he apparently recovered. He was discharged, and died from the same strain of strep infection 6 months later.

The decedent's estate sued, alleging negligence on the part of the hospital for not treating his infection properly. The hospital presented testimony from a doctor who opined that, because the bacteria was so virulent, the decedent would have died within a month after his release had the original strep bacteria been the one responsible for his death: Therefore, he said, the decedent must have contracted a second infection outside the hospital.

A question for dueling MDs? Not necessarily. A call to the federal Centers for Disease Control established that 176,000 people contract that particular strain of strep infection annually. Stated statistically, Americans have a six in 10,000 risk of such infection. Thus, the chances of becoming infected twice in one year by that strain of strep equaled .0006 squared, or six in 100 million. In other words, the chances that the decedent suffered that particular infection twice were astronomical. The fatal infection, beyond any reasonable doubt, had occurred in the hospital.

Workplace Violence Suit: Statistics Backfire

Another instance where quick statistical analysis was able to cut through a fog of confusing testimony occurred in a recent case involving workplace violence. A suit was brought against an agency that supplied temporary labor to businesses. A worker supplied by the agency became violent at his workplace and the employer was sued for negligence by a worker who was injured.

A human-resources expert hired by the plaintiff produced voluminous testimony. The expert's testimony ranged far and wide, citing a huge number of possibly relevant sources. One, however, clearly zeroed in on the crux of the controversy. The expert cited a statistical study showing that the expected rate of workplace violence nationally is 13 incidents per 1000 workers per year. The human-resources expert might have been more careful in her choice of citations: The defendant agency's carefully documented records showed that the rate of workforce violence among the employer's workers was barely one per thousand workers, significantly under the national average. The inference, therefore, is that, far from being negligent, the defendant employer was exemplary in terms of averting workplace violence.

Statistics and the Med-Mal Cap Debate

The lessons are obvious: Much seemingly well-documented expert testimony, when subjected to the light of statistical scrutiny, can decompose rather quickly. However, statistical analysis can conceal as readily as it reveals. Such is the case in the ongoing debate on the use of caps on non-economic damages in medical malpractice matters. While it may be too early to tell how effective caps on non-economic damages might prove, a stampede of statistics is being unleashed by organizations and individuals on both sides of the debate, augmented by a plethora of government and independent studies.

One lively and perhaps instructive exchange of statistical gunfire erupted last year when Weiss Ratings Inc., an independent insurance-industry evaluator, published a study of the effects of caps on medical malpractice insurance premiums. The reports stated that, counter-intuitively, median annual premiums increased at a higher rate in states with caps (48%) than in states without caps (36%). The Impact of Non-Economic Damage Caps on Physician Premiums, Claims Payout Levels, and Availability of Coverage, June 2, 2003, Weiss Ratings Inc.

Not surprisingly, the Physicians Insurance Association of America (PIAA) countered with a blistering attack on the statistical methodology employed by Weiss Ratings. The PIAA's impassioned rebuttal emphatically asserts that the caps are universally effective in slowing the rate increase for medical malpractice premiums: They said that Weiss simply did not consider the correct data in the correct manner. The Weiss Ratings Report on Medical Malpractice Caps: Propagating the Myth That Non-Economic Damage Caps Don't Work, July 8, 2003.

A smoking gun of sorts uncovered a year ago added fuel to the raging debate. The document was a filing made by GE Medical Protective, the nation's largest medical-malpractice insurer, to the Texas Department of Insurance. Texas enacted caps on non-economic damages in medical malpractice 6 months prior to the filing in October 2003. The filing explained why GE protective Medical was raising its premiums 19% even though the caps were in effect. The filing stated that non-economic damages are a small percentage of total losses paid, and the cap will result in a loss saving of only 1.0%. This was a surprising admission, given the fact that GE Medical Protective had actively sought such caps because non-economic damages were “spiraling out of control.”

No less a statistical powerhouse than the US General Accounting Office also weighed in with a report in August of last year, “Medical Malpractice: Implications of Rising Premiums on Access to Health Care,” GAO-03-836. The GAO concludes that the limited available data indicate that caps slow the rate of premium growth, but that it could not determine the extent to which differences in premiums and claims payments across states were attributable to states' tort reform laws or to factors such as laws regulating rate setting, other tort reform measures beside caps, the level of competition among insurers in particular markets, and interest rates.

Using Statistics: An Inexact Science

The jury is still out on caps; the participants seem to have an unending supply of statistical ammunition to throw at one another, and a conclusion to the debate is nowhere in sight at this point. One thing is certain, however: While statistical analysis can cut through a lot of misunderstandings, it can also cloud issues to the point where reasonable conclusions are almost completely lost. The latter has been the case with the use of statistics in the med-mal cap argument. Properly employed, though, statistical analysis has a far more positive than negative value in litigation. It can make or break the case in certain situations, and can certainly help a case along in many others. Just be sure you do your homework or you may find that your statistics hurt you rather than help you.

Complex statistical evidence — based on huge samples, reams of academic and government studies and hours of testimony — has been the weapon of choice in many epic battles. The list of major cases in which both sides have deployed legions of statistical experts is impressive: Bendectin, silicone breast-implants, Agent Orange and tobacco are just several of the many substances over which massive statistical battles have been waged. Currently, both sides of the debate over whether caps on non-economic damages help reduce medical malpractice insurance premiums are trading ground-shaking volleys of weapons-grade statistical analysis.

However, in many litigated cases, a more modest use of statistical analysis can be a deadly counter to the opposing side's expert evidence. In two cases in which my firm has rendered assistance to a legal team, ill-advised testimony was refuted quickly and convincingly by holding the statements up to the cold light of probability analysis.

Statistics Prove Where Infection Was Contracted

One of the cases we consulted on was a medical malpractice action brought against a hospital. The decedent in this matter had contracted a severe streptococcus infection while he was a patient. Streptococcus and similar bugs abound in hospitals. Many such hospital-borne infections are specialized for survival in the antiseptic hospital environment, where they are subject to high-powered antibiotics and develop a stubborn resistance to treatment. The hospital treated the decedent for the infection, and he apparently recovered. He was discharged, and died from the same strain of strep infection 6 months later.

The decedent's estate sued, alleging negligence on the part of the hospital for not treating his infection properly. The hospital presented testimony from a doctor who opined that, because the bacteria was so virulent, the decedent would have died within a month after his release had the original strep bacteria been the one responsible for his death: Therefore, he said, the decedent must have contracted a second infection outside the hospital.

A question for dueling MDs? Not necessarily. A call to the federal Centers for Disease Control established that 176,000 people contract that particular strain of strep infection annually. Stated statistically, Americans have a six in 10,000 risk of such infection. Thus, the chances of becoming infected twice in one year by that strain of strep equaled .0006 squared, or six in 100 million. In other words, the chances that the decedent suffered that particular infection twice were astronomical. The fatal infection, beyond any reasonable doubt, had occurred in the hospital.

Workplace Violence Suit: Statistics Backfire

Another instance where quick statistical analysis was able to cut through a fog of confusing testimony occurred in a recent case involving workplace violence. A suit was brought against an agency that supplied temporary labor to businesses. A worker supplied by the agency became violent at his workplace and the employer was sued for negligence by a worker who was injured.

A human-resources expert hired by the plaintiff produced voluminous testimony. The expert's testimony ranged far and wide, citing a huge number of possibly relevant sources. One, however, clearly zeroed in on the crux of the controversy. The expert cited a statistical study showing that the expected rate of workplace violence nationally is 13 incidents per 1000 workers per year. The human-resources expert might have been more careful in her choice of citations: The defendant agency's carefully documented records showed that the rate of workforce violence among the employer's workers was barely one per thousand workers, significantly under the national average. The inference, therefore, is that, far from being negligent, the defendant employer was exemplary in terms of averting workplace violence.

Statistics and the Med-Mal Cap Debate

The lessons are obvious: Much seemingly well-documented expert testimony, when subjected to the light of statistical scrutiny, can decompose rather quickly. However, statistical analysis can conceal as readily as it reveals. Such is the case in the ongoing debate on the use of caps on non-economic damages in medical malpractice matters. While it may be too early to tell how effective caps on non-economic damages might prove, a stampede of statistics is being unleashed by organizations and individuals on both sides of the debate, augmented by a plethora of government and independent studies.

One lively and perhaps instructive exchange of statistical gunfire erupted last year when Weiss Ratings Inc., an independent insurance-industry evaluator, published a study of the effects of caps on medical malpractice insurance premiums. The reports stated that, counter-intuitively, median annual premiums increased at a higher rate in states with caps (48%) than in states without caps (36%). The Impact of Non-Economic Damage Caps on Physician Premiums, Claims Payout Levels, and Availability of Coverage, June 2, 2003, Weiss Ratings Inc.

Not surprisingly, the Physicians Insurance Association of America (PIAA) countered with a blistering attack on the statistical methodology employed by Weiss Ratings. The PIAA's impassioned rebuttal emphatically asserts that the caps are universally effective in slowing the rate increase for medical malpractice premiums: They said that Weiss simply did not consider the correct data in the correct manner. The Weiss Ratings Report on Medical Malpractice Caps: Propagating the Myth That Non-Economic Damage Caps Don't Work, July 8, 2003.

A smoking gun of sorts uncovered a year ago added fuel to the raging debate. The document was a filing made by GE Medical Protective, the nation's largest medical-malpractice insurer, to the Texas Department of Insurance. Texas enacted caps on non-economic damages in medical malpractice 6 months prior to the filing in October 2003. The filing explained why GE protective Medical was raising its premiums 19% even though the caps were in effect. The filing stated that non-economic damages are a small percentage of total losses paid, and the cap will result in a loss saving of only 1.0%. This was a surprising admission, given the fact that GE Medical Protective had actively sought such caps because non-economic damages were “spiraling out of control.”

No less a statistical powerhouse than the US General Accounting Office also weighed in with a report in August of last year, “Medical Malpractice: Implications of Rising Premiums on Access to Health Care,” GAO-03-836. The GAO concludes that the limited available data indicate that caps slow the rate of premium growth, but that it could not determine the extent to which differences in premiums and claims payments across states were attributable to states' tort reform laws or to factors such as laws regulating rate setting, other tort reform measures beside caps, the level of competition among insurers in particular markets, and interest rates.

Using Statistics: An Inexact Science

The jury is still out on caps; the participants seem to have an unending supply of statistical ammunition to throw at one another, and a conclusion to the debate is nowhere in sight at this point. One thing is certain, however: While statistical analysis can cut through a lot of misunderstandings, it can also cloud issues to the point where reasonable conclusions are almost completely lost. The latter has been the case with the use of statistics in the med-mal cap argument. Properly employed, though, statistical analysis has a far more positive than negative value in litigation. It can make or break the case in certain situations, and can certainly help a case along in many others. Just be sure you do your homework or you may find that your statistics hurt you rather than help you.