| Comment Record|
Dr. Richard Sylvester ||
2002-02-14 03:59:13 |
EORTC Data Center |
| Comments for FDA General |
1. General Comments
Comments of the European Organisation for Research and Treatment of Cancer Data Center on Docket No. 01D-0489: FDA Draft Guidance for Clinical Trial Sponsors: On the Establishment and Operation of Clinical Trial Data Monitoring Committees
Most of the draft (all but section 6) deals with “Data Monitoring Committees (DMC)” as opposed to “Independent Data Monitoring Committees (IDMC)”. This distinction should be made earlier in the text. Then one should stress from the beginning the need for the independence of committee members from trial participants and industry or sponsor representatives and define what is acceptable and what isn’t. We feel that the need for complete independence of committee members does not come across strongly enough in the draft since it is discussed starting only in section 6 on page 18.
At the beginning of the document the distinction between phase III trials on one hand and phase I or phase II trials on the other hand is not made clear. Except for section 4.4.2, is this document meant only for phase III trials ? This should be clarified.
The document emphasizes the situation where investigational products are being evaluated for marketing approval. But what about academic studies where this is not the case: supportive or strategy studies, trials of surgery or radiotherapy ? Do the same procedures apply, both for reporting to FDA and otherwise ?
Section 4.1: Preparation and analysis of interim data (page 7).
Section 220.127.116.11 Format of Interim Reports to the DMC and Use of Treatment Codes (page 9)
Section 6.4 Conduct of the Interim Analysis (page 20)
In the multicenter, cancer cooperative group setting of the EORTC, similar to that of the ECOG and SWOG in the United States, the trial data manager prepares the data for the interim analysis which is carried out by the trial statistician. Within this setting we feel that it is difficult for a completely independent third party to carry out the analysis for the following reasons:
1. They will not be as familiar with the trial or the data as the trial data manager and trial statistician and are thus more likely to make mistakes in the analysis.
2. With a large number of trials it becomes much more difficult both practically and financially to have the trial analyzed by a third party. Many of these trials may not have any financial support from the industry.
In cases where the trial statistician’s knowledge of interim efficacy data might pose a problem for future changes or amendments to the protocol, the advice of a completely independent statistician should be sought for the changes. In any case all such amendments must still pass through the appropriate protocol review channels within the organization, to include external expert review.
Section 4.2.2 Interim reports to the DMC (page 7)
The first paragraph states that safety data should generally be available only to DMC members during the course of the trial. Does this restriction apply to toxicity data from phase III trials where toxicity is not a primary endpoint ? And what about phase I and phase II studies ? In many cases a DMC may meet only after one or two years to review efficacy data, however detailed, non aggregated toxicity data need to be reviewed on a much more regular basis, by the sponsor and/or the principal investigator who may also be entering patients. The need for this review is even stated in the last paragraph before section 18.104.22.168 on page 13. In addition ICH Topic E6, Guideline for Good Clinical Practice, states in Section 5.17.1: “The sponsor should expedite the reporting to all concerned investigators(s)/institutions(s), to the IRB(s)/IEC(s), where required, and to the regulatory authority(ies) of all adverse drug reactions (ADRs) that are both serious and unexpected.” We feel that the recommendations proposed in the FDA document may result in the toxicity data not receiving the close review that it needs.
Section 22.214.171.124 Initial meeting (page 9)
Section 4.3.2 Statistical Methods (page 10)
In many organizations such as the EORTC there is a separate, independent Protocol and/or Ethics Review Committee (PRC) that must review and approve all new protocols including the statistical aspects of the interim analyses. This is not the role of our IDMC. When both a PRC and a DMC exist, the roles and potential interaction of the two committees must be made clear so that there is no conflict between the two.
Section 126.96.36.199 Maintaining Meeting Records (page 17)
The last paragraph states that the sponsor should submit to FDA all meeting records including the confidential reports that were prepared for the DMC. If the sponsor is a pharmaceutical company, then this implies that the company has access to the confidential reports. This is contrary to the first paragraph of this section and section 6.3 that states that the sponsor should not receive the comparative unblinded outcome data.
Section 6.5 Sponsor Access to Interim Data for Planning Purposes (page 21)
The statistical dangers of designing further trials based on immature, unblinded, interim data should also be mentioned.
On behalf of the EORTC Data Center,
Richard Sylvester, ScD
Assistant Director, Biostatistics
EORTC Data Center
83 avenue E Mounier, Bte 11
1200 Brussels, Belgium