ABSTRACT
The AEDMI project, carried out in the Bellvitge University Hospital, near Barcelona, is proposed as an example of clinical decision system prototype development. Several difficulties have been encountered to develop this project. One of them was to find the adequate understanding for the project and proper funding for personnel resources. The support by FIS grants and from the hospital structures was essential to the survival of the project; however the resources always have been below the real needs for development. The difficulties with clinical coding systems, the excessive expectancies upon prototypes, and the lack of culture in information methods were other relevant hinders. The complexity of clinical realm, the uncertainty of medical decisions, and the current trends to substitute the clinical information by data obtained from electronic devices, or analytical monitoring of patients, are additional problems for clinical information managing.
Subject(s)
Decision Support Systems, Clinical , Attitude of Health Personnel , Bayes Theorem , Hospitals, University , Humans , SpainABSTRACT
BACKGROUND: The methods and characteristics of clinical data gathered at the initial steps of development of a computerized system to aid medical diagnosis are reported. The objectives of the study were as follows: to describe the overall method and to set a framework for developing an intellectual model of the medical diagnosis procedure. MATERIAL AND METHODS: A structured medical interview and physical examination using an informatic program on PC compatible portable computers were completed in a sample 1,238 patients attending the outpatient clinics of our institution. Data obtained were compared with information in the patient's medical record taking as reference pattern the record of physicians in charge of the patients. Diagnosis were codified according to WHO International Classification of Diseases (ICD-9-CM). RESULTS: The distribution of symptoms and signs corresponding to the different organs and systems was analyzed. Each subdivision afforded a range of 1.3 to 3.9 abnormal findings per patient. A total of 3,571 diagnoses were codified for the whole group 1,238 patients with a mean (standard deviation) of 3 (2) diagnoses per patient (range 0-12). The distribution of diagnostic groups varied depending on the consideration of the main diagnosis or the concomitant diagnoses that defined the patient's clinical context. The most frequent main diagnoses included tumors, cardiovascular diseases, gastrointestinal disorders, and genitourinary tract diseases. CONCLUSIONS: As shown by results obtained in a sample of 1,238 patients, there is a very complex situation in clinical practice due to the simultaneous occurrence of several clinical patterns. This finding should be taken into account when developing clinical decision making support systems. The use of a structured medical interview or a structured and standard medical visit may be an adequate tool to clarify this matter and to contribute to standardization of clinical concepts and situations.