According to the book “The Body” a French lawyer struck a deal in 1965 with ninety year old Jeanne Calment in which he agreed to pay her twenty hundred Francs until her death. When she died, he would own her apartment Arles. Given her age, it seemed like a good deal. However, she went on to become the oldest person on record by living another thirty two years. In fact, she outlived the lawyer who died at age seventy seven. By then he had paid twice the market value for an apartment he never occupied. The lawyer learned the hard way about relying upon statistical significance, mathematical probability and statistics in general. Yet, we are faced, more and more frequently, in our cases with expert testimony relying on statistical credibility. We need to understand the subject in order to deal with this kind of proof offered in support of “scientific evidence.”
The well respected medical journal, Lancet, has published articles about the waste of money involved in medical research. The reason much research is poor is reported to be because “most scientific studies are wrong, and they are wrong because scientists are interested in funding and careers rather than truth.” Researchers are publishing studies that are too small, conducted over too short a time, and too full of bias in order to get promoted and secure future funding. A 2023 article in the medical journal Nature, https://www.nature.com/articles/d41586-023-02299-w, was titled “Medicine is plagued by untrustworthy clinical trials. How many studies are faked or flawed?” The article reports that the editor of a medical journal, employed by England’s National Health Service, studied more than 500 studies over a period of several years. He judged that at least 44% of these contained flawed data: impossible statistics, incorrect calculations or duplicated figures. At least 26% had problems so widespread the trial was impossible to trust, either because the authors were incompetent or because they had faked the data. He labeled these studies “zombie” trials because they looked like good medical research, but closer scrutiny showed they were actually masquerading as reliable information.
In the United States, many high profile studies are now widely believed to have relied upon unreliable data and arrived at unreliable conclusions. As an example, a gigantic study of Aspirin involved twenty two thousand people over five years concluded that taking daily small doses would reduce the risk of heart attack. More recently, The American Heart Association and cardiology specialist advise the risks, including bleeding, outweigh the benefit and people over seventy are likely to do more harm than good by taking the aspirin. It’s our job to not accept the documentation offered by the defense in support of their claim when it is based upon medical statistics and medical studies
Factors Impacting Epidemiologic & Statistical Studies
There are many factors that impact the accuracy and validity of epidemiologic and statistical studies. Some of the more important ones involve these:
1.In dealing with statistical medical issues, one problem is that unlike other sciences, it is usually impossible to make exact calculations. While the laws of physics allow exact calculations, in medicine, such exactness is rare. Instead of exactness, medicine usually has to look at “statistical relationships “that apply to average situations. Statistical relationships may hold true on average but not in every case. For example, tall people tend to weigh more than short people or older children tend to be taller than younger children. However, one cannot therefore conclude that in every case tall people are heavier than short people or that all older children are taller than younger ones. In addition, statistical significance is not the same as medical significance which deals with the exercise of medical judgment. Something important to a physician’s medical judgment based on experience may not be statistically established in medicine. Many problems in medicine are medically treated even though there is no statistically accurate research to dictate the treatment.
2.“Statistical confidence” refers to the degree of certainty that a particular outcome can be predicted for a conclusion. In medicine there are limits for drawing accurate associations between a variety of medical findings. The accuracy of making associations depends upon a proper foundation for comparison Statistical confidence requires a foundation for the study that is appropriate and the correct size. If the foundation is faulty, false conclusions can be drawn from studies that appear valid.
3. Experience has shown that statistical studies are subject to “investigator bias” that influences the outcome of a study. The sponsor of the study or the person conducting a study may have a strong self-interest motive regarding the outcome of the study. This bias regarding the hoped for outcome may consciously or unconsciously effect interpretation of data and study results
4. Some studies have used a small sample portion taken from another larger study. Taking a smaller sample from a larger group allows one to wrongly argue the whole group has the same characteristics as the sample group when that is likely untrue. Doing so makes the study unreliable because it invites misinterpretation.
5.The size and makeup of the study group is directly connected to accuracy of the finding. An epidemiological or statistical analysis study has to have an appropriate make up for the study involved. It is possible to have two investigations conducted with the same methodology but resulting in conflicting results due to the size and the subjects selected for the group. Examples of this problem have been illustrated by studies published in reputable medical journals regarding the same medical issue but resulting in conflicting conclusions due to the group used in the studies.
6. In evaluating statistical evidence, there are a number of factors to consider. In preparing for epidemiology and statistical evidence here are a few issues to consider. These subjects involve factors that can influence conclusions:
In addition, to those questions, consider the following subjects as well. When evaluating the accuracy of statistical conclusions, it’s important to ask a range of questions to ensure that the conclusions are reliable. Here are some questions to consider:
1. Sample size and representation
2. Data collection
3. Measurement and variables.
4. Statistical methods
5. Replicability and peer review.
6. Publication bias in reporting
7. Alternative explanations.
8. Longitudinal and cross-sectional studies.
9. Control of Variables
10. Ethical considerations
The use of expert testimony and evidence of statistical findings is commonly involved in our trials. Evidence regarding issues of proximate cause, causation and negligence often involves statistical studies. It’s important to understand the subject and prepare in advance if we are going to properly represent our clients.