Dear Dr. Roach: I read your column on a regular basis, and on occasion your response to those seeking advice on the benefits of some particular medication or treatment includes the comment that the trials were not large enough or powerful enough to show any significant benefit.
Are there any medical guidelines for setting up expensive and time-consuming medical trials so that there can be a high level of confidence in the benefit of a particular medication/treatment one way or the other? Is there a minimum success rate of a medical trial before the benefits of the medication/treatment can be considered significant enough for it to be recommended?
Dear A.D.: In the first place, the editors and reviewers of the medical journals publishing papers evaluate the papers for statistical and clinical significance of the research findings. For large randomized, controlled trials (which usually reveal the best evidence for effectiveness of a treatment), there is a statistic called, appropriately enough, “the power.” The power gives the likelihood that a study will be able to find an expected difference between the treatment group and the control group. A second statistic, the p-value, gives the likelihood that the observed difference between groups could have occurred by chance if the two treatments were equally effective. Larger and longer (and thus more expensive) trials tend to have more power to detect a true difference and to reject the hypothesis that the two treatments are identical.
The power level varies considerably among trials, and it depends also on how large a difference between the groups is expected.
The minimum p-value that is usually accepted is 0.05. This means that there is a 5 percent probability that the apparent difference is actually due to chance. A lack of statistical certainty is one reason that “truths” in medicine seem to change over time. There are other reasons, though, including mistaken assumptions, poor research design and, very rarely, outright fraud.
Once a write-up of a trial has convinced a journal, it still needs to be accepted (or not) by the scientific community, who are not a unified body. Some physicians may begin using the new treatment immediately, while others wait for confirmation. I tend to be conservative, especially for the column, and usually wait for very strong evidence before making recommendations.
Dear Dr. Roach: I had toothache and went to a dentist, who sent me to an oral surgeon, who removed a tooth. Now (on and off), I have a toothache where the tooth was removed ( I realize it can’t be a toothache, because there is no tooth there). I’ve been back to both the dentist and the surgeon many times.
They have checked and rechecked, and have come up with nothing. The oral surgeon told me to see a neurologist, because it could be a nerve that is causing me this discomfort. Does that sound right?
Dear B.C.: Feeling pain at the site of a tooth extraction is a nerve pain syndrome called “phantom tooth pain,” or atypical odontalgia. This condition is very similar to other types of face and head pain, such as trigeminal neuralgia. A neurologist is absolutely an appropriate person to confirm the diagnosis and treat it, which often is done with medications usually used for seizures or depression.
I would be concerned that the initial tooth pain might not have been from a tooth at all, but from the beginning of the nerve pain syndrome. I have heard of people who have undergone multiple extractions for what ultimately was found to be a nerve problem, not a dental one.
More information is at fpa-support.org.
Email questions to ToYourGoodHealth@med.cornell.edu.