To the Editor.
—At a time when escalation of medical costs is causing increasing public and private concern, I must applaud the publication of studies aimed to show that a less costly or more efficient method of providing treatment is fully satisfactory. However, before such studies are accepted as a reason for altering standards of treatment, they must be carefully scrutinized. The likelihood that a given study is erroneous and misleading, and the seriousness of the consequences of adopting a false conclusion must be considered. Therefore, I believe Baughman et al in their article entitled "Early Discharge Following Acute Myocardial Infarction: Long-term Follow-up of Randomized Patients" published in the May 1982 Archives (142:875-878) have not fully shouldered that burden.Many physicians, and virtually all clinical researchers, are familiar with the P value that emerges from statistical analysis of medical data. The numerical value of P measures the probability that, in