Edit checks and validation rules: understanding what the system is telling you

Free Lesson Preview

This is just the beginning

The full CRC track covers 8 courses from study start-up to close-out — the skills sponsors actually look for.

Already enrolled? Sign in

Three alerts, three different messages

You enter a blood pressure value of 42 into the systolic field. The system blocks the save entirely -- a red warning, no option to proceed. You correct it to 142, and the field accepts the entry. On the next form, you enter a diastolic blood pressure of 96 and a systolic of 88. The system does not block the save, but an amber dialogue appears: "Diastolic value exceeds systolic value. Please verify and provide an explanation if correct." You check the source document, realize you transposed the values, and correct them. The next morning, you log in and find a notification: an overnight validation check flagged that a participant's hemoglobin dropped from 14.2 g/dL to 7.1 g/dL between Visit 3 and Visit 4 -- a cross-form query asking you to investigate.

Three alerts. Three entirely different messages from the system. And three different responses required from you.

This is what edit checks and validation rules look like in practice. They are the EDC system's quality safeguards -- programmed logic that evaluates your entries against defined criteria and tells you when something appears wrong. But "appears wrong" covers a spectrum, from "definitely wrong and you cannot proceed" to "possibly wrong and you should investigate." The coordinator who treats all edit checks identically -- either ignoring them reflexively or panicking at every alert -- is not using the system effectively.

Understanding what the system is telling you, and responding appropriately to each type of message, is one of the most consequential skills you will develop as a coordinator. It is also, in my experience, one of the most underappreciated. Edit checks sit in an odd pedagogical space: they are too routine to feel important and too technical to feel interesting. But they are the front line of automated data quality assurance, and your response to them directly determines whether that quality assurance succeeds or fails.