Database: relevant or not.

Progress in the field of ventricular assist devices requires a more rigorous and systematic method of collecting outcomes data. A worldwide registry of device implants and results is proposed. With widespread participation, data from this registry would improve the identification of risk factors and complications, and allow for the creation of predictive models that would enhance patient selection. Professional societies should lead the development of a registry in close partnership with government and industry. Data collection using the Web, with rigorous security measures to protect patient privacy, would offer numerous advantages in efficiency and immediacy of communication for all participants.

The conduct of clinical trials: a Food and Drug Agency perspective.

The conduct of a clinical trial required by FDA regulations should be executed in conformity with a protocol developed as a cooperative endeavor involving the Agency, the sponsor requiring device approval for marketing, and the investigators. This is the process most likely to assure a scientifically satisfactory study design that will address the primary concerns of all the parties. More importantly, timely access to safe and effective technologic advances will be made available to the health provider community based on the best scientific evidence.

FDA's requirements for in-vivo performance data for prosthetic heart valves.

The Food and Drug Administration (FDA) has recently revised its "Replacement Heart Valve Guidance". That document lists the data FDA deems necessary to support the approval of new prosthetic heart valves of all designs, and which should be contained in Premarket Approval Applications for these devices. The guidance covers detailed data requirements for in vitro, animal, and clinical data. This paper is intended to briefly summarize FDA's requirements for in vivo and clinical data. The clinical study must establish that the device is both safe and effective, as compared to currently marketed replacement heart valves. It is possible to achieve this goal using hypothesis testing to compare the results of an observational study against a set of Objective Performance Criteria (OPC) which have been established by the FDA. The establishment of the OPCs was facilitated by a standardized set of definitions of complications published by the American Association of Thoracic Surgery and Society of Thoracic Surgeons (AATS/STS) in 1987/1988. Papers published in peer reviewed journals have utilized this set of definitions for data analysis, providing an ample pool of data from which to establish OPCs. The number of patients required to establish the safety and efficacy of a replacement heart valve, using this approach, is 800 valve years, 400 in the aortic and 400 in the mitral position. Advantages of this approach are reduction in the number of patients and duration of the study.