Seven years experience with a computerized diabetes clinic database

With the emergence of information technology applications in medicine, a computerized medical record system that could be used to : (1) maintain patients' clinical records over time, (2) communicate with referring practitioners, and (3) form the basis of a potential research database of information, was sought. In 1987, we developed such a clinical database to register patients attending our busy Diabetes Clinic, now seeing in excess of 300 new referrals and, on average, 3,000 clinic visits per year. Baseline demographic data, clinical history, and examination and investigation results are recorded. We also record diabetes therapy and other medication dosage and changes, monitor follow-up, assess health outcome information (such as stroke or amputation), and generate results, summaries, and reports to referring practitioners and other health professionals. We now have almost seven years of experience using the system. Initially established on a single PC with paper-based data collection and subsequent data entry (running as a DOS application), it is now established on a PC Local Area Network [LAN] with terminals in the clinic consultation rooms enabling direct data entry and allowing patients to view their results in graphic form on screen. From its inception, the Diabetes Clinic Database System has maintained patient demographic and clinical data (which facilitates efficient clinic management) with patient clinic lists and adhesive address labels generated from appropriate menus. Batch mode processing produces daily work sheets which facilitate the running of clinics as well as ad hoc, daily, and weekly reports for all patients (as required). This expedites correspondence with referring doctors. A quality assurance report to the clinic doctor highlights missing clinical information which must be obtained in order to ensure data completeness. The initial system was relatively inefficient in that it required data entry following patient consultation and provided no immediate feedback to patients themselves. In January 1994, to address these deficiencies, the system was established on a PC LAN (running under Novell); it provided on-line data entry within the clinic setting and enabled patients to participate in the recording of their information, observe their own progress by way of on-screen graphs (e.g., blood sugar control, weight, cholesterol), and receive hand-held summaries generated immediately following the clinic visit. Batch programs generate hard copies of this data to be filed in medical records. Two major assessments of the system have been undertaken. In February 1990, we undertook a survey of Local Doctors with 5 or more patient referrals on the system; this resulted in a pleasing 66% response rate. There was an almost universal acceptance and indeed a significant preference for this system over 'traditional' letters. In January 1994, following the introduction of the system onto the PC LAN for direct data entry in the clinic setting, we assessed (by anonymous questionnaire at the end of the consultation) patient attitudes towards these changes. The development of the CRS Diabetes Clinic Database System has improved our approach to diabetes outpatient care and our communication with other health professionals. It has the added benefit of providing a database of information that is suitable to address critical clinical research issues in diabetes management. This system provides an acceptable blend of information technology and clinical medicine, redesigning and enhancing the way we deliver medical care to patients with diabetes. Involving the patient in the collection and interpretation of their clinical data via a computer system (as utilized within our clinical unit), is both acceptable to the patient and her referring doctor. Ongoing system refinement and assessment remains integral to our use of information technology.