The growing number of Methicillin resistant Staphylococcus aureus (MRSA) infections in hospitals may in part be driven by physicians’ tendency to over-prescribe antibiotics to avoid being sued by disgruntled patients, according to a study published this past fall in the American Journal of Therapeutics.
And how did the study reach that conclusion? Here’s the abstract of the study, Relationship Between Population Density of Attorneys and Prevalence of Methicillin-Resistant Staphylococcus aureus: Is Medical-Legal Pressure on Physicians a Driving Force Behind the Development of Antibiotic Resistance?:
We performed a hypothesis-generating study evaluating the perceived threat of lawsuits among physicians and methicillin resistance in Staphylococcus aureus. We found a correlation between the prevalence of methicillin resistance among clinical S. aureus isolates and both antibiotic prescriptions per capita and density of attorneys in countries in Europe and North America. We did not find a correlation between prevalence of methicillin resistance and physician density. Further investigation is warranted to study whether physicians’ perceived fear of lawsuits, of which attorney density may be a crude surrogate marker, results in antibiotic prescription practices that contribute to the emergence of antimicrobial resistance among virulent pathogens such as S. aureus, with global implications on the ethics of the delivery of quality health care to all members of society.
Although there can be merit to a preliminary investigation of data, a "hypothesis-generating study" is, by definition, unscientific. The scientific method requires the investigator form the hypothesis before performing the experiment.
There’s good reason for that; e.g., following the method helps guard against bias, which can cause investigators to draw spurious relationships from data.
Like the spurious relationship claimed between the "density of lawyers" in a given area and MRSA.
Despite the headline-grabbing conclusion in the abstract, Drs. Sakoulas, Wormser, Visintainer, Aronow, and Nadelman didn’t actually find much evidence with which to blame lawyers for MRSA. As the study — which you have to pay to read, and thus the vast majority of people can’t or won’t read — says:
Using the nonparametric test for trend, we noted a trend of increased lawyers per capita in states with higher percentages of MRSA, but this did not achieve statistical significance (P = 0.215). When analyzed in 2 groups, states with a prevalence of .40% MRSA in ICUs tended to have more attorneys per capita than those with a prevalence of ,40% MRSA (mean 347 vs. 288 per 100,000; mean rank 23.4 vs. 16.5; P = 0.08). The difference in physician density was less significant (300 vs. 257 per 100,000; mean rank 22.1 vs. 19.0; P = 0.42).
That is to say, the authors found no relationship between lawyers per capita and MRSA.
Instead, the authors found a relationship between lawyers per capita and MRSA after they narrowed the definition of "MRSA" to mean an abnormally high prevalence of MRSA in ICUs.
Does that mean that more lawyers in a given area causes an abnormally high prevalence of MRSA in ICUs?
As every good scientist knows, correlation does not imply causation. This logical fallacy is so common, and so thoroughly disproven, that it has its own fancy Latin name: cum hoc ergo propter hoc.
The authors know that. As the study — that is, the part not available to the public — admits:
This study has a number of limitations, and therefore, its conclusions must be interpreted with caution. … The relationships that emerged should only be considered as crude, not causal, associations because practice behavior was not assessed at the individual level and adjustment for confounders or other covariates was not conducted.
So let’s recap what the "study" did:
- Disregarded the scientific method; then,
- Played around with data in the hopes of finding a link between lawyers and antibiotic-resistant bacteria; then,
- Failed to find such a link; then,
- Admittedly cherry-picked data until they could claim a weak correlation; then,
- Admittedly made numerous assumptions about the link between "density of lawyers" and physicians’ behavior; and finally,
- Admittedly did not evaluate any other variables that could explain the correlation.
Nobel-worthy this is not.
But that’s not what bugs me about the "study."
What bugs me is how supposedly objective researchers don’t even try to conceal their motives anymore:
George Sakoulas, M.D., assistant professor of medicine and lead author of the study, concluded, "The findings suggest that more research is needed to evaluate the potential impact of medical liability concerns on the medical care system. The study findings hint toward the importance of medical tort reform as a way to reduce healthcare costs and improve quality. Another way might be to foster more judicious prescription of antibiotics based on science and evidence rather than on risk aversion."
Remember: Dr. Sakoulas didn’t have any "findings [that] hint toward the importance of medical tort reform as a way to reduce healthcare costs and improve quality." He cherry-picked data until he found a weak correlation between lawyers per capita and an abnormally high prevalence of MRSA in intensive care units, a correlation he admitted could be entirely meaningless since he chose not to investigate other possible explanations.
That, however, is enough for him to jump right to "the importance of medical tort reform as a way to reduce healthcare costs and improve quality."
Worse, we already know there’s a better explanation for the weak correlation between lawyers and MRSA, an explanation that requires far fewer leaps in logic.
There’s another fancy Latin phrase scientists use when faced with competing explanations: entia non sunt multiplicanda praeter necessitatem. The principle is better known as Occam’s razor:
Occam’s razor states that the explanation of any phenomenon should make as few assumptions as possible, eliminating those that make no difference in the observable predictions of the explanatory hypothesis or theory. The principle is often expressed in Latin as the lex parsimoniae (translating to the law of parsimony, law of economy or law of succinctness). When competing hypotheses are equal in other respects, the principle recommends selection of the hypothesis that introduces the fewest assumptions and postulates the fewest entities while still sufficiently answering the question.
When faced with data suggesting a correlation between the "density of lawyers" and a high prevalence of MRSA in ICUs, there are at least two possible explanations:
- The "density of lawyers" in a given area — including lawyers who work for hospitals, lawyers who have devoted their careers to defending doctors, and lawyers who have never touched a medical malpractice file — specifically makes physicians more concerned about malpractice liability than they otherwise were, leading them to over-prescribe antibiotics; or,
- The "density of lawyers" and a high prevalence of MRSA in ICUs are both determined by one or more common factors.
The first explanation — the one adopted by Dr. Sakoulas — requires making a lot of assumptions, such as (a) doctors know how "dense" their area is for lawyers; (b) that the "density" of lawyers really is "a crude surrogate marker" for the fear of a lawsuit, even though the authors themselves admit "there was no significant relationship between attorney density and number of paid malpractice claims per 1000 physicians at the state level;" and, (c) concern about malpractice liability specifically increases the likelihood that a physician will prescribe one of the antibiotics tied to increase MRSA development, over and above the numerous other factors that encourage the physician to prescribe the antibiotic.
That’s an awful lot of assumptions.
The second explanation requires making a single assumption: that there is at least one factor that can lead to both a higher density of lawyers and a higher prevalence of MRSA in a given area.
One such factor comes to mind: the degree of urbanization.
MRSA is primarily an urban disease: it was first diagnosed in large, urban hospitals, thereafter spread faster in urban areas, and today is more prevalent in urban areas than in rural areas. The more urbanized an area is, the more prevalent MRSA will be.
Lawyers, too, are an urban phenomenon: the states with the highest "density" of lawyers (D.C., New York, Delaware, Massachusetts) are also the most urbanized, while the states with the lowest "density" of lawyers (North Dakota, Arkansas, South Dakota, Kansas, Idaho) are the least urbanized. The more urbanized an area is, the more lawyers it will likely have.
An increase in lawyers doesn’t cause an increase in MRSA; urbanization causes both. Quod erat demonstrandum.
I don’t doubt the authors — who included an infectious disease specialist and an epidemiologist — knew the relationship between a high prevalence of MRSA and urbanization. I also don’t doubt the authors suspected the relationship between lawyer density and urbanization.
They just didn’t want to let science get in the way of politics.
– – –
If you were injured by medical malpractice, contact a Philadelphia medical malpractice attorney.