A look at how medical decisions are made

Every time something is done to a patient, whether it is giving a drug or performing a test or procedure, there must be compelling reason, based on scientific evidence.

This evidence is obtained by clinical studies, and published in journals edited by independent, highly credentialed experts in the field.

Clinical studies

There are several types of clinical studies.

One is where patients who have a certain specific condition are treated and the results compared to other treatments or no treatment. In good studies, the patients in both groups have similar characteristics, such as age, race, other medical conditions, etc.

An example of this would be the recent studies comparing angioplasty to medical therapy with statins (such as Zocor and Lipitor). Incidence of heart attacks, death and other outcomes for chronic angina patients were compared, as well as complications from treatment. Because the results indicated that medical therapy was superior for the selected population studied, many cardiologists now recommend drugs rather than angioplasty for chronic angina.

Drug studies are often done as part of new drug investigations, registered with the FDA in order to provide this agency data needed to allow its sale.

In some of these, patients are given either the drug or a placebo (sugar pill). To prevent bias, neither the patients nor the investigator knows which the patient gets. The results of these studies are usually presented to show what percent of patients getting the real drug benefit compared to patients getting a placebo. The placebo effect is very strong for conditions that are subjective, such as depression, irritable bowel, muscle pain, headaches, etc., making studies difficult to interpret.

In the case of surgical procedures, there should be good evidence that the procedure achieves a result sufficient to justify the risks or costs of surgery. To provide the necessary evidence, the results of many separate studies are often pooled together.

Getting to the truth

This all sounds so simple. But it isn’t. The interpretation of the data is often complex. Bias by the investigator is a big problem.

A study of the effectiveness of Botox for frequent migraine headaches (more than 15 per month) was done several years ago. Patients were given Botox at various strengths as well as pure saline. The results were interpreted to show Botox was effective treatment. However, if one looked at the actual data, the saline injection was more effective than most of the Botox strengths given, and there was no pattern to the responses from different strengths. There was no way the data could have been interpreted the way it was. Since then, there have been several more studies. But it seems few have used a saline control injection, yet they draw the same conclusion that Botox is effective.

Why? In most cases, the authors of the studies are linked to Allergan pharmaceuticals, the company the makes Botox.

In other bias issues, the investigator is someone who earns his living from treating with the same procedure he is studying, and may design the study or interpret the findings to show desired results.

Trusting the writers

Another problem is that many studies are published in journals that are not “peer reviewed.” The standards used to evaluate the quality of the studies are often lacking.

Why do doctors even read them?

These are all for-profit journals funded entirely by ads from the drug companies. They are sent free to physicians. The authors in the journals frequently are involved in some way with the company that makes the drug. Possible conflicts of interest are not clearly stated, as they are in the independent, scientific journals.

Patient selection represents a big problem. Studies are mostly done on very specific patients.

Yet doctors are often presented with the findings in such a way that it appears that they can be applied to different patients. In the case of antiobiotics, studies are often for a very specific group of patients with a clearly defined bacteria causing the problem. Once the drug is marketed, the drugs are promoted (unofficially) for many non-studied uses. It is one of the major reasons that very powerful antibiotics are unnecessarily used for simple infections, resulting in the rise in resistance to the drugs.

In summary, most clinical data is obtained in an impartial, scientifically valid manner. But it is incumbent on physicians who use evidence to make sure that the sources are unquestionable. Any new proposed treatment must have multiple, quality studies. Acting on only a few studies can be dangerous.

Charles Barta retired to Green Valley after 10 years as a medical director for several health care insurers. Before that, he was physician-in-charge of Kaiser Permanente of Colorado and a private internist in Las Cruces, N.M. He had previously held a management position in the Medical Systems Division of Pfizer. His column is published Sundays. He can be reached at Cbar52@aol.com.