clinical data management

It says that the Investigator considers the data for that particular (part of the) CRF accurate, reliable and complete.

In a completely paper-based workflow the Investigator signs off, for example, each patient’s CRF visit. This complete visit is taken to the data entry staff of the clinical data management department. Additions, modifications to the CRF visit practically can’t be made because the completed CRF visit is at the clinical data management location. Any queries popping up are sent to the Investigator and he/she makes corrections or reacts upon the queries on the query form. This query form is signed for by the Investigator too. This way the circle is closed. All CRF data is owned by the Investigator(s). (Notice that Data Managers only verify the data. They aren’t allowed to add, modify, delete received data without the Investigator’s signature.)

In an electronic data capture (EDC) study the Investigator enters the CRF data and signs electronically for accurateness, reliability and completion after, for example, each CRF visit. However, the Investigator can still add, modify or delete data in that CRF visit at a future time-point, until the CRF visit is locked for editing. E.g. a data correction as a reaction upon a query. Then again the Investigator should sign for the changed CRF visit or for the changed data point only. In the end all CRF data entries, modifications or deletions should be somewhere along the way signed for by the Investigator. He/she is the ‘CRF data’ owner.

So this is why signing by the Investigator is so important. You want the conclusions and interpretations in the Clinical Study Report to reflect the truth. (1) Showing what actually happened with the subjects under study. The Investigator signs for that. The Data Manager verifies.

And secondly, the signature documents (2) the Investigator’s ‘timely’ entry & review of CRF data.

OK, now more practical.

1. The Sponsor decides what he finds suitable; signing after each visit, per patient, or once for the complete site. You can add room for an electronic signature in your EDC system or you can have the Investigator sign for the data entries on paper, next to the EDC system. It’s your choice. In what way do you think the Investigator can truly set his/her signature for a set of CRF data? In my opinion after each CRF visit and for each data change, addition or deletion in an already signed CRF visit.

2. A digital signature should at least consist of a login name and a password. For a series of consecutive data entries (e.g. a CRF visit=expected to be entered in one session), the combination of logging in with name and password, and after entering, signing off with password only is enough (21 CFR part 11.200 a.1.i.).

3. An audit trail only is for EDC not enough (GCP 4.9.3.). An audit trail takes care that initial entries and modifications are not obscured. But it isn’t the same as conscious signature for the data added, deleted or modified in the EDC’s CRF.

4. A digital signature is linked electronically to (part of) a CRF. In my opinion, follow the sites logistics in patient care. What part of the CRF is expected to be entered in one go? Are there specific measurements like X-ray results that are expected to be entered at a later time-point? Then also ask for a digital signature at the end of the X-ray visit entries.

5. Any coding (e.g. for adverse events) should not be signed for by the Investigator. He/she doesn’t own the codes added. Coding is added by a Data Manager and approved by a Medical Reviewer (Physician) for all patients.

6. You could think of setting Investigator signature rights in your EDC system. Who is allowed to sign for CRF data? All people on the site delegation list? Or only the (sub-)Investigator(s)?

7. Before database lock a Data Manager should check amongst others that sign off is present for all CRF data in the EDC system.

8. For obvious CRF inconsistencies, like spelling errors, you can make a data convention list upfront. Have the Investigator sign this data conventions list before first real entry, and then you can make corrections as on the list without having the Investigator’s signature for them. E.g. for obvious spelling inconsistenties.

That’s it. You could give it a thought. Just remembering why Investigator signatures are essential in clinical data management.