A SOFTWARE PACKAGE FOR EXACT LOGISTIC REGRESSION

IN MANY CLINICAL OR EPIDEMIOLOGICAL STUDIES, THE OBJECTIVE IS TO RELATE A BINARY RESPONSE VARIABLE TO TREATMENT, EXPOSURE LEVELS, AND OTHER COVARIATES. USUALLY THIS IS DONEBY USING A LOGISTIC MODEL WITH THE DATA ANALYSIS BASED ON LARGE-SAMPLE ASSUMPTIONS. FREQUENTLY, HOWEVER, THE SAMPLE SIZE IS SO SMALL OR THE DATA STRUCTURE SO SPARSE AS TO INVALIDATE THESE ASSUMPTIONS. FOR EXAMPLE, STUDIES OF CANCER ETIOLOGY, PREVENTION, OR TREATMENT OFTEN HAVE A SMALLNUMBER OF SUBJECTS. IN SUCH CASES, SMALL-SAMPLE OR EXACT METHODS OF DATA ANALYSIS NEED TO BE USED. THESE METHODS, ALTHOUGH WELL RECOGNIZED IN STATISTICAL THEORY, HAVE TENDED TO BE COMPUTATIONALLY INTENSIVE AND THEREFORE CONSIDERED TO BE IMPRACTICAL BY DATA ANALYSTS. THE RECENT DEVELOPMENT BY CYTEL SOFTWARE CORPORATION OF A COMPUTATIONALLY EFFICIENT, MULTIVARIATE SHIFT ALGORITHM FOR THESE PROBLEMS HAS, HOWEVER, CHANGED THE PICTURE AND PLACED EXACT LOGISTIC REGRESSION WITHIN THE REACH OF THE WORKING STATISTICIAN. THE MAIN GOAL OF THIS PROJECT IS TO CONDUCT THE PRELIMINARY DEVELOPMENT OF A USER-FRIENDLY AND WIDELY MARKETABLE SOFTWARE PACKAGE FOR THIS ALGORITHM. SPECIFICALLY, THIS PROJECT WILL (1) UNDERTAKE A STUDY OF THE FEASIBILITY OF IMPLEMENTING THIS ALGORITHM ON MINI- AND MICROCOMPUTERS, (2) EXPLORE THE POSSIBILITY OF ATTACHING THE ALGORITHM TO EXISTING STATISTICAL PACKAGES LIKE SAS AND EPIC, AND (3) LAYTHE GROUNDWORK FOR DEVELOPING A CONVENIENT DATA MANIPULATIONINTERFACE FOR THE ALGORITHM.