From clinical data records to research: a database system for the study of clinical and functional indicators of...

The identification of reliable clinical and functional indicators of Chronic Heart Failure (CHF) is currently a major research challenge for physicians dealing with this pathology. With the aim of providing an adequate infrastructure for this research, we have developed a Database System where all relevant information concerning CHF patients during follow-up can be efficiently recorded, monitored, extracted and easily transferred to commercial packages for data processing and statistical analysis. Patient clinical status is recorded every 6 months, whereas data from several laboratory investigations are recorded every 12 months. All complications or events between two successive controls are recorded too. Patients needing cardiac transplantation are entered in a transplantation waiting list. The Database is fully integrated into the Hospital Information System and meets the standards of the national database of CHF patients. It grows with the concurrent activity of several independent teams, all of which have access to complete and structured information for research purposes. A user-friendly procedure allows the export in a standard format (Microsoft Excel) of all patients' data pertaining to a selected set of controls and variables of interest. These data are used directly for analysis or sub-selections are performed by the investigator through a simple query language. Depending on the objectives and complexity of the research, different commercial statistical analysis packages are available, which are used by clinical investigators autonomously or with the aid of a statistician. This overall approach allows great autonomy of each user in extracting, manipulating and statistically analyzing data. The Database has been in use at the Heart Failure Unit of our Rehabilitation Center since October 1992 and 595 patients have been enrolled since then. Several studies based on these data have been performed and more than 120 scientific communications and 30 articles in national and international journals have been produced. Hence, this experience represents a successful example of how clinical data records can be efficiently and effectively linked to valuable clinical research.