medpundit |
||
|
Thursday, November 14, 2002The latest research is likely to encourage many doctors to make blood tests for inflammation part of standard physical exams for middle-aged people, especially those with other conditions that increase their risk of heart trouble. The study, based on nearly 28,000 women, is by far the largest to look at inflammation's role, and it shows that those with high levels are twice as likely as those with high cholesterol to die from heart attacks and strokes. But what does “twice as likely” mean, exactly? Do ten percent of people with high cholesterol die of heart attacks and stroke? Do twenty percent? Five percent? Two percent? What percentage with high C-reactive protein die of heart attack and strokes? Five? Ten? One? Only the authors know. They haven’t shared that information with CNN, and they haven’t revealed it in their paper, either. They mention in the abstract, and in the paper that “77 percent of all events occurred among women with LDL [bad cholesterol - ed.] cholesterol levels below 160 mg per deciliter (4.14 mmol per liter), and 46 percent occurred among those with LDL cholesterol levels below 130 mg per deciliter,” but that isn’t at all the same as telling us what percentage of people with those cholesterol levels had “events.” The paper itself only portrays the results in terms of “relative risks” and “predicted events,” a statistical technique that inflates the signficance of findings. It never once gives absolute numbers. This is particularly maddening since it was sponsored by a grant from the NIH and since it’s being used to promote the widespread use of the test for screening purposes. As the CNN article points out, some doctors are already treating patients with expensive cholesterol-lowering drugs (the statins) on the basis of their C-reactive proteins, a move which would only further inflate overall medical costs in return for uncertain benefits. The main author of the paper, Dr. Paul Ridker, has a patent on the process of testing for C-reactive protein - a test that can cost anywhere from $25 to $50. Clearly, it’s to his advantage to give the data the most positive spin possible. The editors of the New England Journal of Medicine, however, owe their readers more. They should require that papers like this include the raw data so the results can be given the scrutiny they deserve, and they should be holding their authors to a higher standard of integrity. posted by Sydney on 11/14/2002 07:49:00 AM 0 comments 0 Comments: |
|