Medical Device News and Insights

Thoughts on InHealth’s Evidence and Reform Summit

Wednesday, November 18th, 2009 | Industry News and Innovations | No Comments

I recently attended a special health technology summit meeting on “Evidence and Reform”. The summit was presented by InHealth, an independent, non-partisan 501(c)(3) charitable foundation, founded in 2003 at the initiative of the board of directors of AdvaMed.

The summit meeting addressed topics in healthcare policy and health technology research in the light of the current US administration’s focus on health care reform.  Attendees included medtech industry leaders, academic researchers, and clinical professionals. The objective was to increase understanding of how rapidly advancing healthcare technologies contribute to the enhancement of healthcare systems in the United States and around the world by improving patient outcomes, productivity, and quality of life. › Continue reading

Tags: health technology research, healthcare reform, InHealth

Adaptive Design for Medical Device Clinical Trials

Monday, July 13th, 2009 | Medical Device Clinical Trials | No Comments

While adaptive methods for prospectively designed clinical trials have been around for over thirty years, it is only fairly recently that we are seeing increasing interest in applying such methods to medical device clinical trials.

Adaptive design is a term that encompasses several concepts, the most prevalent of them being the notion of early stopping of a trial. In a trial designed for possible early stopping, one or more interim analyses are conducted prior to the final analysis, the idea being that the accumulated evidence at the interim look may be sufficient to draw an appropriate inference and make a sound decision.

Not all medical device clinical trials lend themselves to early stopping. For example, if a trial is such that the primary endpoint for each subject cannot be ascertained until two years of follow-up, an interim look may not be feasible-accrual into the study may be complete before a practical early look can be conducted.

Early stopping of a trial is sometimes used to prevent subjects from receiving an inferior treatment, to allow more subjects to receive a superior treatment, or, to end a trial that is futile, that is, one for which there is no reasonable expectation of showing any difference between the treatments being compared.

There are drawbacks to studies designed for possible early stopping, not the least of which is that if a study is stopped early for superiority, there may not be sufficient experience with the superior treatment either for deliberation by the medical scientific community, or for the sponsor of the trial to know how best to market the product. One of the earliest papers on the subject, and one of the best in terms of understanding the concepts and the pitfalls appeared in 1987 (Interim Analyses in Randomized Clinical Trials: Ramifications and Guidelines for Practitioners; Geller, NL and Pocock SJ; Biometrics 43, 213-223, March 1987).

One of the more useful applications of adaptive designs is the ability to modify sample size during the course of the clinical trial. The need for sample size re-estimation comes about because all sample size calculations make key assumptions about the primary study outcome. For example, in comparing two treatments, one might specify a priori both the expected difference in average response and the expected variation inherent in measuring response at the subject level. It would be ideal to have an opportunity to examine these specifications before the trial is terminated in case the specifications are wrong. For example, if in doing the power analysis we assumed a standard deviation of 2.00 units, but we are finding that the standard deviation is closer to 4.00 units (based on say, 20% of the data), we might wish to adjust our sample size to reflect this fact. Greater uncertainty requires larger sample sizes to detect a fixed response difference between the two treatments.

Specialized, commercial statistical software is available for designing clinical studies with adaptive features. Appropriate use of such software depends upon the input of a highly knowledgeable and technically trained statistician who is familiar with the concepts and who can ask the right questions of the sponsor in order to identify the relevant input parameters and interpret the output.

You may be interested in two upcoming events addressing the topic - the 4th Annual Trial Design Innovation Conference and CHI ADAPT 2009 Conference.

If you are presently designing a medical device clinical trial or are contemplating doing so, you would be well advised to consider whether one or more adaptive design features should be incorporated into your study protocol. Feel free to contact me, (feldstein[at]mdci[dot]com] to explore this topic further.

Tags: adaptive trial design, medical device clinical trial design