FDA Medical Device Requirements

Copyright Charles S. Tritt, Ph.D.
April 6, 1997

  1. FDA Device Classes

    1. Class I: External devices (surgical scissors, manual stethoscopes, etc.) are subject to general controls only such as the current good manufacturing practices (cGMP), quality assurance and labeling.
    2. Class II: More invasive devices (x-ray machines, ultrasound devices, etc.) are subject to general controls and special controls, such as compliance with performance standards, postmarket surveillance (tracking), and additional preclinical and clinical performance data as the FDA deems necessary.
    3. Class III: Life support and life sustaining devices (heart valves, VAD's, etc.) are subject to general and special controls and require formal premarket approval.

  2. FDA Applications

    1. If a device is substantially equivalent to one in use before 1976 only a 510(k) application is required.This is the "easiest" way to get a device approved.
    2. An Investigational Device Exemption (IDE) allows an unapproved device to be used on human subjects.The IDE must be filed and approved before you can start clinical trials.
    3. The results of the clinical trails are used to file a Premarket Approval (PMA) application that must be submitted and approved before a new device can marketed.
    4. Devices must be shown to be reasonably safe and effective for specific pathological conditions in order to be approved.

  3. Requirements to Start Clinical Trials

    1. Risk/Benefit analysis based on objective criteria.
    2. Informed consent must be obtained from all subjects.
    3. A local Institutional Review Board (IRB) and the FDA must approve the protocol.

  4. Issues Involved in Clinical Trials

    1. A clear hypothesis must be stated and the study must be designed to support or refute this hypothesis.
    2. Clear entry criteria must be established and enforced. A major problem with clinical steadies in the past has been physicians admitting patients to the study group that do not meet the entry criteria. This in effects creates a new study with a different hypothesis and may render their data useless.
    3. Some type of control group or baseline data is required. There are at least four types of controls possible. Each has its own advantages and disadvantages. These are:

      1. Prospective randomized groups. This is the best scientific/statistical approach, but runs into ethical problems when applied to potential fatal conditions without any viable alternative treatments.
      2. Non-randomized groups (possibly patients at a different center). Not as statistically sound as the randomized groups, but possibly more practical and ethical.
      3. Patients used as their own controls. Allows only limited questions regarding the immediate response to treatment to be answered, but avoids ethical dilemmas.
      4. Historical control groups. Same problems and advantages as item b above.

  5. Device Tracking

    1. Required by Safe Medical Devices Act of 1990.

  6. Manufacturing Control

    1. Requirements listed in the Quality System Regulation (also known as the current Good Manufacturing Practices or cGMP).

These notes are based in part on Chapters 19, 20 and 21 in "Cardiac Mechanical Assist Beyond Balloon Pumping," Susan J. Quall, Ed., Mosby, 1993.