I do hope Fred Hassan is sitting down!
258. To reduce the possibility of the results being biased, clinical trials are often “double-blinded,” meaning that neither the patient nor the doctor (nor, with most trials, the pharmaceutical company sponsor) knows whether the patient is in the experimental group (i.e., the group taking the drug under study) or the control group (i.e., the group taking placebo or another drug with known effects in the study population). Despite the obvious (and intended) limitations of blinding a clinical trial, if certain data about the population in the trial are made available, it is possible to discern useful information about the trial’s results, even while the treatment assignments remain blinded. Recognizing this, on February 11, 2008, in the Second Senate Letter, Senator Grassley wrote to Hassan that ENHANCE statisticians would not have needed to unblind the ENHANCE data to know that the study was not likely to show a statistically-significant difference between treatment arms:
It has come to my attention that Schering Plough and Merck would not need to unblind the data to understand that Vytorin performed no better than generic simvastatin. . . . These studies try to detect a statistically significant difference between treatment groups on the primary endpoint. Once the results are recorded, the study is then unblinded to determine which drug is the better performer. However, if the drugs performed the same, meaning there is no statistically significant difference in the treatments, then this information is apparent before the study has been unblinded.
259. Dr. Allen Taylor of Walter Reed reached the same conclusion. Dr. Taylor told Heartwire that, “Somebody had looked at the end-point examination, the IMT results, and, irrespective of group assignment, could know that a groupwise comparison of CIMT changes showed no statistically significant difference. . . . In my view, once that is known, the trial is functionally unblinded.”
No comments:
Post a Comment