DEVICE STUDY SPONSORS NEED TO CONSULT "EXPERTS" ON INCLUSION CRITERIA

DEVICE STUDY SPONSORS NEED TO CONSULT "EXPERTS" ON INCLUSION CRITERIA for clinical trials, FDA says in its draft guidance on medical device clinical studies. According to the agency, a study's "population should be defined before the study by the development of rigorous, unambiguous inclusion/exclusion criteria. Experts in the field of clinical study should develop the criteria." The agency adds that the inclusion/exclusion criteria should "uniquely characterize the study population and relate to the intended use for the device." The draft, which was released at a Sept. 27 FDA-sponsored workshop, says that in determining criteria, baseline factors such as age, sex and concomitant disease should be considered (see preceding story). "Extreme care should be taken to identify those influencing variables that are likely to impact study outcome" in order to minimize "the chance that conclusions drawn at the end of the study will be spurious." The guidance cautions against "investigators or sponsors [attempting] to reach their desired sample size by including some borderline patients that "just miss" inclusion by the specified criteria," as happens "frequently." Borderline patients are "inevitably" discovered and disqualified, FDA says. Inclusion of such patients also would likely lead to the agency questioning "the validity of other study components." FDA says it developed the draft because of "a concern about the quality of clinical studies submitted to support medical device applications." Several of the recommendations in the guidance address questions raised by the Committee for Clinical Review (the Temple Group) in its March report on a sample of device premarket submissions. For example, the Temple Group's analysis found that there was "poor specification and characterization of patients entering studies." The guidance also suggests ways to steer clear of what the Temple Group considered the fundamental problem with device applications: "lack of attention to basic study design." The draft outlines specific issues that should be considered by investigators in order to clearly define a study hypothesis and establish endpoints that will provide a precise measure of a device's safety and effectiveness. For example, sponsors should think about how it is "to be determined that the device is effective" and whether the device is "as effective or more effective than another intervention." The relative safety of the device compared to other treatments also should be considered, as should the "best clinical measure of safety and effectiveness." The draft guidance further recommends that feasibility, or pilot, studies be conducted "if a sponsor cannot answer the key questions needed to focus the study because of insufficient experience with the device in human populations." These studies enable a sponsor "to identify possible medical claims for the device," monitor and determine the precision of potential endpoints, "test study procedures" and refine the prototype device." FDA currently does not require feasibility studies but would be likely to encourage them in certain situations, such as the development of a groundbreaking new device. At the Sept. 27 workshop, Richard Chiacchierini, director of the division of biometric sciences in the device center's office of surveillance and biometrics, noted that "pilot studies usually are not suitable to be evolved into clinical studies." Rather, the study "is a good place to do...exploratory data analysis. If you can, crudely test as many hypotheses or follow as many leads in the data as you can to try to get focused study questions for future studies." The draft guidance also says that sponsors need to "monitor intensely the conduct of the study" to ensure compliance with the study protocol. The Temple Group noted that protocols, when developed, were often not well-followed. FDA acknowledges that studies rarely can be conducted exactly as envisioned. However, the effect of unforeseen circumstances can be mitigated if "contingency plans" are "carefully crafted with the goal of preserving the integrity of the established design."