2004D-0440 - Computerized Systems Used in Clinical Trials Guidance
FDA Comment Number : EC3
Submitter : Mr. Jesper Rosendal Date & Time: 12/21/2004 06:12:34
Organization : H. Lundbeck A/S
Drug Industry
Category :
Issue Areas/Comments
GENERAL
GENERAL
At Lundbeck we are pleased to see that the new guideline is more focused on the risk based approach. We believe that this opens up the possibility to add a lot of new value that the new technology can provide.

To highlight some of the individual areas where we see the guideline as an improvement, we list the most important ones here:
1) Data Retention. Same demanding as for paper trials ? Enables us to migrate instead of maintaining old hardware and software
2) Printed name not needed to be displayed on the screen.
3) Change Control. The term revalidation is substituted by the level of validation activities

Still however we do have some questions to interpretation of the guideline in some areas:

General: The term ?durable media? is no longer refereed to in the guidance. Therefore, we urge the agency to alter the guidance to reflect that is considered permissible to define the source data for eDiary data as the first storage at a central computer, instead of at the eDiary itself? (based on a risk assessment this would be the best alternative).

Line 18: This sentence may be interpreted to mean that the patient medical records could be covered under this guideline, whereas line 65-66 points more directly to the medical records as being optional (the principles of this guideline may be applied). As the industry has low influence on systems managing the medical records, please be more precise in the definitions regarding this.

Lines 76-77: Identiying, in the study protocol, all steps at which electronic records are created, modified, maintained, archived, retrieved or transmitted will cause redundant information and is error prone. We suggest that the agency allows this information to be described in procedures that may be referred to in the study protocol.

Lines 84-85: We fully appreciate the requirement for designing and validating fields to ensure that only input of a given format may be entered, thereby preventing illegal input to a system. However, a significant meaning of the term ?preclude? is to ?rule out in advance? (Merriam-Webster Online Dictionary). Since it is impossible to rule out errors totally, we suggest that the agency replaces ?preclude? with a more lenient term (e.g. ?strongly reduce the possibility for?).

Line 101: FDA will exercise discretion regarding specific requirements regarding computer generated audit trails and opens for the opportunity to have other sorts of audit trails, e.g. paper based audit trails. However according to the definitions on line 537 an audit trail is per definition computer generated. There is an apparent discrepancy between the body text and the definitions section. Therefore, we urge the agency to ensure consistency between the two.

Line 107: We suggest that the sentence be changed to ?and any KNOWN system vulnerabilities?.

Lines 243-245: For systems that span different time zones, a good way of tracking time is by using one central server to apply all time stamps (i.e. not local time). We suggest that the agency modify this sentence to clearly acknowledge that central time stamps are acceptable.

Lines 404-409: The agency expects that a design specification be available and also that testing based on the design specification be performed. In the case of off-the-shelf software, no testing based on the design specification will usually be performed by the sponsor organization. Testing by the sponsor is usually only performed to test functionality (PQ). We recommend that the agency consider revising the paragraph to reflect this fact. We recommend that the agency should expect to see, a written specification that describes what the system is intended to do (Requirements specification), A written test plan based on the sponsors requirements and test results/evaluation. In addition to this, we urge the agency to extend the definitions section to include the specific terms used with reference to computer system validation.