Background: The registration book for admission and discharge of patients was the basis of a survey of hospital morbidity and the main source of information concerning the diseases attended in the hospitals in Spain. The aim of this study was to evaluate the quality of this information from the data of a hospital with a computerized patient registration system.
Methods: The sample collected from the registration book in 1985 by the National Institute of Statistics studied three types of errors: the main error being selection of diagnosis, coding and transcription of the principal diagnosis, and comparison of data contained in the patient discharge form.
Results: In the 896 releases studied an error oscillating between 1 and 2% was found in transcription. The principal diagnosis had been erroneously selected in 26% of the reports with more than one diagnosis. Important coding errors were found in 11%. Transfer between different hospital wards or death were variables found to increase the probability of error.
Conclusions: Hospital registers should be submitted to quality control processes in which the physicians facilitating the data should participate.