Abstract
Howell (1968) and Howell and Loy (1968) have described the 'fruit machine' method of diagnostic coding which, using a large computer, will code at the rate of 2,000 diagnoses a minute. More recently, the Medical Research Council's Computer Unit has undertaken reprogramming and extension of this 'fruit machine' method to enable coding to be carried out on a medium-sized machine, currently at the rate of about 60 diagnoses a minute. The data from the present series were processed by courtesy of the Medical Research Council. The Research and Intelligence Unit of the Scottish Home and Health Department suggested that a comparison be made, using Scottish morbidity data, of the merits of machine and manual coding. The objects of the study were to examine approxi mately 25,000 hospital discharge summaries: (1) to assess the present accuracy of manual coding over various diagnostic ranges for the different regional boards; (2) to compare the accuracy of machine and manual coding; (3) to assess the feasibility of future machine cod ing of Scottish morbidity data, i.e., to establish that automatic coding is operational on 'real data'; and (4) to provide further information which would be used to update the fruit machine dictionaries, i.e., to make them more comprehensive.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have