A bioassay computer program for analgesic clinical trials

Abstract

A Fortran program to calculate estimates of relative potency of a test to a standard analgesic is described. The program considers, at the option of the users, four distinct populations: completers, all of those patients who have completed a full crossover round of medication; two‐rounders, patients who have completed two or more rounds of medication; incompleters, patients who have not completed one or more rounds of medication; and first dose only patients, the initial administration received by all patients in the study whether or not they have dropped out. Up to six observations plus the initial reading for pain intensity scores and pain relief scores may be processed separately. The program also creates several summary variables including: SPID, an estimate of the area under the reciprocal of the pain intensity curve; TOTAL, an estimate of the area under the pain relief curve; and estimates of onset and duration as measured by relief scores and by pain intenSity scores. The program automatically performs calculations‐ for the analysis of variance and produces F values when appropriate for testing linearity of the log dose relationship, parallelism of the two curves for test and standard, and so forth. The paper discusses the rationale behind the approach adopted in the program.