Medical Research: The Basis of Evidence

There is an old joke – When a doctor has seen a medical condition once, he says “In my experience.”  When the doctor sees a condition twice he boasts, “In my series.”  However, when he has seen something three times, he exclaims, while pounding his fist on the table, “In case, after case, after case!”

This joke is twice old.   First, it was told more then 50 years ago. Second, it is the old way of practicing medicine.  Once upon a time, doctors made decisions based primarily on their personal experience.  What a physician had seen, what treatments had succeeded or failed by his own hand, was the basis of his recommendations. What changed in the last 50 years has been formalization the scientific model in medical research. Today at the physician’s fingertips is a plethora of carefully constructed data.

Last week Princeton graduate Dr. Paul Meier died.  In an excellent obituary (NYT 8/14/2011) there is discussion of the profound contribution by Dr. Meier to medical research and patient care.  Dr. Meier was an early advocate of doing research by comparing two different treatments in two groups of patients chosen by randomization.  He taught researchers to assign patients to one group, usually  to receive standard therapy, and a second, matched group, who get the experimental treatment.  Dr Meier showed that by randomized testing it is possible to prove one therapy better then another.

This research model has four parts.  First, which patients are going to be tested? What is the medical problem?  It is critical to be very exact. It is not enough to say “breast cancer.”  One needs to be clear that you want to test “Stage I, estrogen receptor negative, invasive ductal breast cancer in premenopausal women who have never had cancer before.”

The second step in designing a research trial is defining the treatment you wish to test. Again, one has to be very complete.  If the treatment is surgery, how exactly will the operations be performed? If is medicine, the dose, how the drug will be supplied and how it will be administered.  If the research will be watching a group of patients (i.e. do yearly chest x-rays on smokers), how often will observations be performed?

Dr. Meier’s major contribution was step three — to what are you going to compare your results?  What are your “controls”?  If you are going to give a new medicine to group A, you must have a randomized, matched group B that does not get the new medicine.   He proved that without a matched group B, it is very hard, usually impossible, to tell if the new medicine in Group A really worked.  Hence, the problem when a doctor uses personal experience to make complex decisions…he does not have a clear group B.

The final step is measurement.  When and how are you going to decide if the experimental treatment worked?  Do patients live longer? Live better? Lose less hair? Run faster? Are less depressed?  It is critical for the researcher to decide before the study starts, what, when and how they are going to measure.  Changing what a study measures in the middle frequently produces confused and useless results.

Patients involved in research studies find these four steps can be confusing and frustrating. Patients that do not match the study group are excluded from any chance of getting the new therapy. As the study controls the treatment, there is no flexibility. Since randomization (i.e. flipping a coin, rolling dice or letting the computer decide) means someone else picks the treatment, it can feel like a loss of control.  In “blinded” trials, it can be frightening not to know what treatment a patient is receiving, even if they are at least getting standard therapy.  Finally, the need to gather lots of data and measurements (more tests) can make a patient feel “like a Guiney pig.”

In the end, however, this carefully controlled experimental model is the key to future therapies.  It can produce clear answers that allow doctors and patients to judge the likelihood that a medical intervention will help.  This type of research is the core of  “evidence-based medicine.”  Physicians pick treatment by data and then use experience to apply therapy and adjust as needed for individual patients.

The doctor’s major role is diagnosis, communication, education, certain skills (i.e. surgery) and judging exactly how to administer proven treatments in specific patients.  Evidence-based medicine allows patients a greater ability to know what can help, hurt or should be avoided.  The research process improves the quality of health and healthcare.


There is no comment on this post. Be the first one.

Leave a Reply