«

»

Apr 28 2011

Statistical Reviews of Case Report Forms

Case Report FormA statistical review of a trial case report form (CRF) is highly important for various reasons. The project statistician is aware of the planned primary and secondary endpoint analyses and knows what data will be needed to implement these analyses. The project statistician also has a clearer view of how the data will affect the production of tables, listings, and figures, since he or she understands the data programming that will be needed for display production. Inadequate data can lead to a lot of frustration, lost hours, and extra costs.

When I am performing a statistical review of a CRF, there are several key items that always receive my attention.

1. The Primary Endpoint(s)

Reason: Without adequate primary endpoint data, the respective analyses cannot be properly performed.

Q: Is the data being collected at all necessary timepoints?

Q: Will the resulting format of the data be acceptable for performing the primary analyses as planned in the protocol?

2. The Secondary Endpoint(s)

Reason: As for the primary endpoint(s), without adequate secondary endpoint data, the respective analyses cannot be properly performed.

Q: Is the data being collected at all necessary timepoints?

Q: Will the resulting format of the data be acceptable for performing the secondary analyses as planned in the protocol?

3. Dose Administration

Reason: Dose administration data is needed for determining both exposure and compliance.

Q: Is the date, time, amount, and duration (if applicable) of each dose being recorded?

4. Date of First Dose

Reason: The date of first dose is highly important for establishing whether medications are concomitant and whether adverse events (AEs) are treatment-emergent. It is also important for many other reasons, such as establishing duration of treatment, determining study days, etc. Sometimes the date of first dose will be collected in a subject diary that is completed at home. This diary data is usually not queried. Therefore, it is important that this data value also be collected in the CRF to be checked and queried if necessary. It is not uncommon for a subject to insert the wrong date or not even provide a date at all in a subject diary.

Q: Is the date (and time, if applicable) of first dose being adequately collected in the CRF?

5. Date of Last Dose (if applicable)

Reason: The date of last dose is important for determining duration of treatment, how many days past last dose an AE occurs, and in certain cases, whether an AE is considered concomitant.

Q: Is the date (and time, if applicable) of last dose being adequately collected in the CRF?

6. Date of Informed Consent

Reason: Date of informed consent is typically used in determining subject age and is simply in and of itself an important date to have recorded somewhere in the CRF. It is frequently expected that this date be provided in study listings.

Q: Is the date of informed consent appropriately collected?

7. Did Subject Complete Study AND Date of Completion or Discontinuation

Reason: I have seen CRFs that recorded if a subject discontinued early, but did not record if the subject actually completed the study. This meant that for subjects who discontinued early (and for which we had the date of discontinuation), we knew when we had received all subject data. However, for subjects who did not discontinue early, we had no way of knowing when (or if) the subject was finished with the study. We also had no date of completion to provide in the study listings, which appeared to be erroneous (but was not).

Q: Is whether the subject completed the study recorded?

Q: Is whether the subject discontinued early recorded?

Q: Is date of completion or discontinuation recorded?

These are just a few of the items that I check when reviewing a CRF. Clearly, if any of the above items is missing or is collected incorrectly, issues will arise, extra work will be necessary, frustration will mount, and costs will go up. For these and many other reasons, it is highly beneficial to a project team to include a statistician in the review cycle for all CRFs.

This is a post by Kara Snoke, M.S. Kara is a Biostatistician at Cato Research.