The patient's history: An appraisal of computer help

The traditional approach in establishing the doctor‐patient relationship is that of “taking the history.” This is fundamental to medical records, but it is a performance seldom re‐explored after the formative period in one's medical education. The patient's history remains fundamental and with systems analysis of patient care, the information furnished by this is designated as a part of the data base, in accord with principles elucidated by Weed and previously reported by our SARCI group specifically for otolaryngology. Our prior report described the struggle with language related to the things we examine and try to describe in terms understandable to a computer. The next consideration has been the evaluation of what kind and how much information would be accumulated for each patient to constitute the standard data base for the otolaryngologist. For greatest computer help, it is clearly required that the same information be gathered on every patient before proceeding to define any additional data which will be used only for further definition of an individual patient's problem. Almost endless sets of patient history data are available to us and our problem has been to select and decide what set would constitute our standard data base for most patients in otolaryngology. Certain principles concerning the concept and content of the data base have become evident. These are as follows Evaluation of the information in the history of a patient taken and recorded by different physicians was conducted and analyzed. The overall result can be summarized to the effect that there was more variation within the results of any one physician than between physicians. Knowing of such variability in contrast to the standardization required to mesh with the computer led us to evaluate questionnaires. Patient questionnaires need not imply leaving a patient alone with a form or a video display tube staring at him. A paper and pencil format needs no time limits and even computer controlled questionnaires can work at the patient's place. Both of these systems gain information that is as complete as we care to make it; but having the same approach to each patient makes the information comparable. We wanted a combination of questions so as to elicit a “signal” whenever there was a need for further physician evaluation. A format containing 65 “Inquiry by Systems” questions was found. It has been tried on seven population samples and has been modified four times during the course of these trials. Any such tool needs evaluation for reproducibility of results and for validity. As an example, we compared positive answers in the ear, nose and throat area with the findings recorded in the medical record. This gave a mean agreement of 75 percent with a mean error of 16 percent. Classical narrative form for recording the patient's history requires a change in language for the computer. At the present time a “yes” and “no” or numbered answer to a question is required to make the potential of the computer useful to the physician. We have established to our own satisfaction that questionnaires are likely to generate useful information consistently and easily, but the physician's judgment is required to deal with false positives as well as inconsistencies.