Development of a database of patients supported by ventricular assist devices

An increase in the number of patients with end stage heart failure is leading to increased use of ventricular assist devices (VAD). However, sometimes the optimal time point for implantation of left ventricular or biventricular support remains unclear. Data analysis using an electronic database may help to make the decision making process more precise and thus improve outcome. However, it is not easy to find a balance between sufficient comprehensiveness of the data, which are selected from a huge amount of available information, and practicability of database maintenance and data analysis. We developed the Assist Database based on Access for Windows. The Assist Database consists of five main parts: (1) demographic and admission data, diagnosis, goal, and type of VAD; (2) preoperative period; (3) postoperative period up to 30 days; (4) follow-up period; and (5) statistical evaluation. The preoperative and postoperative parts include hemodynamic data; ventilatory support; laboratory results; results from echocardiographic, neurologic, pathologic, and other examinations; medication; and complications. The follow-up part documents readmissions, complications, and outcome. From April 1987 to October 2002, eight different types of VAD were implanted in 654 patients in our institution. Their data were retrospectively added to the Assist Database using medical records and different previously used electronic databases. Since the Assist Database came into routine use, it has been supplied daily with selected data of current patients. On the data entry level, the data arising from medical records are entered either manually via standard forms or automatically from other electronic documentation systems used in our hospital in routine patient care to collect laboratory results, demographic data, blood transfusion data, and operative data and from electronic patient charts via interfaces. The structure of the database is designed to facilitate the data analysis level. The database presented is one of three databases united to form a network. The structure of the Assist Database facilitates comprehensive, time saving data collection, which allows different online data analyses. These analyses may affect the decision making process and thus improve outcome. However, achieving a balance between the volume of available information, the time consumed, and the relevance of the data for further analysis remains difficult. The Assist Database should include information relevant for the decision making process and for the prediction of outcome. In particular, data collection should be focused on patients' preoperative condition and on postoperative organ function and quality of life. Further, different databases (for patients with congestive heart failure, assist device patients, and transplanted patients) should be unified to form a network to avoid the repeated collection of identical data, to save time, and to increase the quality of analysis. In the long term, multicenter use of the Assist Database could be considered.