Outcome adjudication

The endpoint committee classifies the study's events blinded to the treatment: two reviewers, an adjudicator when they differ, the settings and the results.

In an outcome trial the primary endpoint is often a clinical event: a myocardial infarction, a stroke, a cardiovascular death. Sites report events in their own words, so an independent endpoint committee classifies each one by the definitions of its charter, blinded to the treatment. Outcome adjudication runs that work in YouEDC Capture: two committee members review each event independently, and a third decides when they differ.

Who does what

• Endpoint committee member (a role on the Team page): sees their queue and the events as dossiers, classifies them. They never see the site, the patients' other forms or anything that reveals the arm.
• The project administrator and the data managers choose the event forms and the categories, follow the reading of every event and download it to Excel; the medical monitor follows it too.

Which events

By default every serious event (AESER = yes) of every form with the AETERM field. In Settings choose other event forms, include the non-serious events, and write the categories of the charter, one per line as CODE|name. The default list is that of a cardiovascular outcome trial: cardiovascular death, non-cardiovascular death, myocardial infarction, stroke, hospitalisation for heart failure, unstable angina, not an endpoint event.

Review an event

1. Open Outcome adjudication: My queue lists the events waiting for you, the ones for the adjudicator first.
2. Open an event: its form as the site reported it, the patient's sex and age at the event.
3. Choose the category, write a comment if you wish and press Save the review.

The first and the second review are independent: neither reviewer sees the other's review until the event is decided. When both chose the same category, that is the result. When they differ, the event goes to a third member, who sees both reviews and decides; the reason is required. Nobody reviews one event twice and the adjudicator is never one of the two reviewers.

When the site changes the event

Each review keeps the event as it was read. If the site changes the event form afterwards (a new diagnosis, a date, the outcome), the event is read again from the start; the earlier reviews stay in its history and in the Excel file, marked as not current.

Audit trail and Excel

Every review, decision and settings change is a record in the audit trail of the study. Excel gives the events with each review, the decision and its reason, and every read with its time, for the committee minutes and the trial master file.