Stat counter

View My Stats

Sunday, August 1, 2010

Overstating certainty

How do we know what we know? It is a simple question but not so simple to answer. Sometimes there are things that we know which require no proof. I love my wife and I know this. I am not required to prove this (except perhaps to my wife at times such as our anniversary). Whether I can prove this or not does not really affect other people.

Other beliefs have consequences which impact on one or many other people. In the world of health care, what we know dictates what we do and how we advise our patients. When someone presents with chest pain, shortness of breath, diaphoresis, tachycarida, and classic ST changes on EKG, we know they are at risk of cardiac arrest based upon hundred's of thousands of observations. Specific interventions have been tested in controlled fashions and although there are still gray areas in terms of optimum interventions, there are reasonably solid numbers in terms of risk associated with specific scenarios.

It is unfortunate that in many if not most of healthcare we see certainty where it does not exist and nowhere is this more disturbing than where health care and law intersect. There is a recent story about Dr. Marta Cohen, a pathologist in England who has taken on the study of the neuropathological findings in shaken baby syndrome. From the New Scientist:

A triad of markers

The three markers for a shaken baby diagnosis are retinal haemorrhages in both eyes; subdural haemorrhages between the fibrous dura layer that protects the brain and the brain surface beneath; and swelling of the brain. Subdural haemorrhages are said to arise from ripping and shearing of so-called bridging veins. New lines of evidence challenge this hypothesis with the discovery that subdural bleeds are much more common in babies than generally appreciated, and for a host of innocent reasons (see "Anatomy of a murder?").
Last year, Cohen and co-researcher Irene Scheimberg of Barts and the London NHS Trust examined post-mortem tissue from fetuses and newborns and found subdural haemorrhages in 16 of the 25 fetuses and 20 of the 30 newborns. They also found haemorrhages within the dural layer itself, suggesting that the bleeding started here (Pediatric and Developmental PathologyDOI: 10.2350/08/08/0509.1). The research is just the latest of many reports to show that subdural bleeds can occur without shaking (see "Anatomy of a murder?").

It seems that by calling into question the dogma accepted as fact by an army of expert witnesses who have testified in trials over many years, Dr. Cohen has raised the ire of those who have a vested interest in maintaining the aura of certainty. The General Medical Council has placed restrictions on her medical license which bar her from further testimony as an expert witness. It appears that she was consistently spoiling the party by showing up to testify in trials and unlike her adversaries claiming the that the triad of findings were specific, she had actual data to support the opposite contention.

I have seen this before. Twenty five years ago the presence of genital or peri-anal warts in an infant was declared proof positive of child abuse. However, after putting countless parents through the ringer, we discovered that..oh never mind, such warts can be transmitted through many other means. Furthermore, the actual numbers which are relevant are poorly defined. Are half of the cases of genital or peri-anal warts in children due to sexual abuse? 25%? 10%? 1%? Who knows and who knows how to know?  How many people were jailed or placed upon the sex offender list based upon the testimony of experts whose hypotheses were essentially untestable.

I have done a modest amount of expert witness work and reviewed a number of depositions. I am surprised by the culture of certainty which prevails. It is scary how many people who are certain they know, don't know that they don't know. Furthermore, they have no concept of what they need to do in order to know how they know. The casualties are people like Dr. Cohen who has the audacity to take on a contrary opinion based upon actually testing the evidence and the tools used to make these assessments of certainty. She had her license restricted based upon the fact she failed to agree with a consensus. Sounds like Galileo. Oh yes the sun does revolve around the earth.


  1. chauncey McHargue M.D.August 3, 2010 at 12:08 PM

    " We prove what we want to prove..." a quote from the French philosopher Alain who understanding the plasticity of the mind checked himself saying also that "Every idea I get I have to deny, that's my way of testing it." An anecdote from my internship when doing a peds rotation, I admitted a 12 mo old with a broken skull and bite marks and his parents were being accused of child abuse of the worst sort although this indigent couple said that the infant had fallen out of the bed that they, the infant and a toddler all slept at night. My impression was of two loving and concerned if poor parents, TV melodramas of a perfect family on every block harboring abusive adults notwithstanding. I could certainly not envision them being so malicious as to inflict bite wounds on their child—but I could certainly conceive of a jealous toddler doing so, having his place in bed with his parents displaced by the newcomer. But the pediatricians and social service workers were ready to pounce on the parents and remove the child from the home until I went down to the gift shop, got some gum and made impromptu dental impressions of the adults and the two year old and showed conclusively that it was the infant's brother who had bitten him, and therefore not unlikely pushed him out of bed in the middle of the night too. To quote Alain again, "There is nothing more dangerous than an idea when you have only one idea,"

  2. One thing I learned from this post is your gender. Certainty should invite skepticism, especially in our profession.

  3. You purport to be certain you know why the GMC made their decision.
    Do you know this?
    If so, please share it. Your blog appears to do no more than smear them for reaching a highly unusual conclusion that doesn't please you.

  4. I am basing this on articles published in the Scientist and the BMJ (BMJ 2010; 340:c1989 doi: 10.1136/bmj.c1989 (Published 13 April 2010)
    News Doctors in shaken baby syndrome case are accused of “scientific prejudice”) and the timing of the GMC decision.

    Are you implying that there was another reason other than her purported "scientific prejudice" for restriction of her license. I may never know since the GMC does post the reasons for the suspension on their website, only the conditions of the suspension (

    I fail to see any particular smear. I am not even suggesting that Dr. Cohen is unquestionably correct. I know what has been reported - she holds an opinion that is different from what is currently held by many other experts and that her opinions are guided by her scientific work which has been published in peer reviewed journal. I also know she was publicly chastised for this followed by restrictions placed on her license. Is there something which is incorrect about any of those items?

    If I am smearing anyone it is those who have no doubts that they are correct. I am certain only that there are few things that one can be certain of.