WHY DOES THE FDA REFUSE TO USE MULTIPLE REGRESSION?

WHY DOES THE FDA REFUSE TO USE MULTIPLE REGRESSION? Patti Waldmeir has an article in the Financial Times (March 7 on internet. March 8 in the American printing) with the snappy headline, “Why the FDA must do a deal with death.” She writes about a court case challenging the FDA policy of limiting access to a terminally ill patient to experimental drugs that might save her. The case is being argued on Constitutional grounds, but the author does not even blink at the appalling FDA rationale for the policy, that only a controlled experiment should be used to test a drug. She quotes the questions from the appeals court judges, “if terminally ill patients can get all the drugs they need without participating in clinical trials, how will drug companies recruit subjects to test drugs for the public good?” Let me restate that argument so that the full horror of it will register: We want to deprive you of what you think may save your life because we want to force you into a controlled experiment. What is weighed in the balance against the life of the patient is a preference for a particular statistical technique. Economic and climate studies necessarily use multiple regression and people are willing to make major decisions on the basis of those studies. If a patient lives longer than expected after taking a drug, that information will be known –and the patient will have lived longer.