More safety points&statistical cross examination
Here a few more points to add to my last post on the general subject of safety. These came from a case where the surgeon encountered a problem using a product during surgery. The manufacturer had specific safety instructions about this problem. The surgeon decided not to follow the manufacturer’s safety instructions and decided to go ahead with the surgery anyway resulting in serious harm to the patient: So these are a few of the safety points I had in my cross examination materials:
- what could be more fundamental then the need to follow the manufacturer’s safety instructions for use of the product?
- If you don’t follow the recipe you end up with a bad result
- If you choose not to follow the manufacturer’s safety instructions about the product, you are putting the patient at risk of serious harm – which is exactly what happened here
- If the written safety rules are ignored, it is always an invitation to disaster
- When the doctor is faced with a choice between safety and risk of harm to a patient, we trust our doctor to always chose safety
This particular case also involved statistical issues. Many cases have such an issue in them. It might be the predicted life expectancy of the plaintiff or the odds of a particular harm occurring in the future from the injury or even an issue of causation. The subject is complex, but here are just a few generalities to consider when cross examining an epidemiologist or statistician. There is a lot of professional literature as well as legal literature about this subject for you to review.
- statistical relationships may hold true as a matter of averages, but do not control a particular case. For example, tall people tend to weigh more then short people, but that is not always true.
- Life insurance mortality tables are used to predict life expectancy, yet they are at best only gross averages where someone might live considerably longer then the prediction or less then the prediction.
- It is possible to draw false conclusions from research studies. An example, is a study where people were given drinks of whisky & water, others were given drinks of rum and water and still others drinks of vodka and water. All of the people ended up intoxicated. The conclusion was that the common factor was water and therefore water causes people to become intoxicated!
- There are numerous things can influence the conclusion drawn from studies. For example,
- How variable the data relied upon? (Exact comparisons are impossible – age, pre existing health and many other factors inhibit the ability to be certain)
- how accurate an answer do you want?
- how large the sample size used "sample size"
- date extraction bias involving clinical interpretation of the result or interpretation of the data i.e. natural human bias.
We have all seen many examples of studies over the years which end up being in conflict with each other over issues of medical health. One study says some food or drink is harmful to health only to be followed by a later study reaching exactly the opposite conclusion
In 1985 the New England Journal of Medicine published two studies, in the same year, on issue of cardiovascular disease in women that arrived at two totally conflicting conclusions
We’ve seem influence of the drug industry on research studies which they pay for and involve their product, but which later turn out to have been biased or incorrect
Yet these questionable research studies have been published in reputable journals by qualified medical people
Statisticians use studies like these to arrive at conclusions which are only generalities and do not involve a research study done on this particular patient