Some Sources and Potential Consequences of Errors in Medical Data Recording
L. Hendrickson, J H Myers · Methods of Information in Medicine · 1973
One year’s patient data from an urban teaching hospital was retrieved from the computer tapes of a large abstracting service. A variable but significant error rate was found in the recording of each entry studied. Such errors caused the data to be unsuitable for a quality of care study since the error rate was either very high or there was evidence of a recording bias. In either case, chart review would have been required and thus the data would have been, at best, only a guide to selecting parameters to study in a chart review. The »information« obtained was so obviously inaccurate that it is unlikely that anyone would consider using it. However, in-hospital monitoring of data summaries received from the abstracting service might well lead to the elimination of obvious errors while retaining plausible but still inaccurate data. The acceptance of such could lead to serious false conclusions. For example, our »data« suggested that ward patients and non-ward patients received a different quality of care, a result not confirmed by chart review. Given the system used, it was concluded that possible errors in data recording could be recognized most efficiently through »check« programs at the abstracting service and that by using such programs, the abstracting service could and should initiate improvements in hospital practices and coder training to minimize errors.