2005 - CRITIQUE OF "ANOTHER PERSPECTIVE ON THE MEMANTINE GRAPHS: CLINICAL VS. STATISTICAL SIGNIFICANCE"
Steve Davis has made us think. He has pointed out aspects of two major papers that he finds troubling, re-graphing for us data that he thinks misleading. His title, obviously provocative, is itself, I believe, a bit misleading. Hence I think his comments require a counter balance. (Unlike the authors of the articles Davis cites, I have not received any money from involved companies.)
Davis is troubled by the magnification of the graphs, showing one point differences as bigger than they'd appear in graphs that detailed the entire scale. However these numbers were not meant to deceive or mislead. The actual numbers and their standard errors were provided. If we look at fever charts a one-point difference can mean a lot, even though a temperature chart could include a y-axis that goes from room temperature to 1100F. Graphs of blood pressure changes that look only at values in the desired range, not starting at O would be thought perfectly appropriate by most readers. Graphs of IQ changes, or weight or height changes, routinely focus on the part of the graph that changes, or simply graphs the change itself. There is nothing particularly valuable in including the lower limits of a scale on a graph intended to highlight the results (since the actual numbers are provided). And highlighting the results isn't necessarily sinister. More importantly, Davis overlooks two very important aspects of the New England Journal of Medicine article he cited. The first is the measure of a 48.5 hour reduction in time spent by the caretaker over the course of a month for memantine-treated patients. This is a huge amount of time, 1 1/2 hours each day for caregivers who may have an hour of free time each day or less. In marginal terms, this might represent a doubling of "free" time. Secondly, the article itself worries about the meaning of its findings. "The clinical relevance of treatment effects has been an issue in all trials of medication for Alzheimers disease. Point differences between drug- and placebo-treated patients in quantitative scales do not necessarily indicate that these effects are clinically meaningful...The treatment effects...seemed to translate into improvements in the patients' behavior (less agitationÖ) and mitigation of the burden on caregivers (fewer hours spent assisting the patient.)"
The clinical relevance of the memantine findings were determined by an expert panel of the Food and Drug Administration. And while it has become clear that the FDA has become tainted by political interests in very recent years, there is no reason to think that the approval of memantine represents a political intervention. I think that these Alzheimer drugs, do, in fact, help people. We certainly can argue about whether their effects are either worth their money, or whether their effects were large enough to justify approval by the FDA (think here about the effects of drugs approved for MS, ALS, stroke and Parkinsons disease), but I don't think it is correct to insinuate that the effect of memantine has been exaggerated by misleading graphs.
Copyright Rhode Island Medical Society Jan 2005
Provided by ProQuest Information and Learning Company. All rights Reserved.
findarticles.com |
