A data validation is a rule attached to a question. When an answer breaks the rule, Datacapt shows the site a message immediately and raises a query on that data point. Errors get caught at the moment of entry rather than months later, by which point the site staff no longer remember the visit.
š Prerequisite: the eCRF Builder permission on the study, and a study in Draft or Suspended status, since the Builder is closed while a study is Live.
Create a validation rule
Each rule is attached to one question, and can test the answers of several questions together.
š ļø Step by step guide
Open the study and go to the eCRF Builder.
Find the question you want to control, and select Edit.
Open the Data validation tab.
Select + Add new validation.
Configure the check on this question's answer.
To bring another question into the same rule, select Add new condition, choose the question, and configure its check. Repeat for as many questions as the rule needs.
Choose a message severity: Information, Warning or Exclusion.
Write the message the site will read.
Select Save.
š” Tip: to add a second, independent rule to the same question, use + Add new validation again rather than adding a condition. Conditions combine into one rule; validations stay separate.
Choose the severity
The severity decides the colour of the message and nothing else. In the eCRF, none of the three blocks the site from continuing.
Severity | The site sees | Blocks entry? |
Information | A blue message | No |
Warning | An orange message | No |
Exclusion | A red message | No |
ā ļø Warning
Exclusion does not exclude anyone. It is the strongest of three colours, nothing more. The subject stays included, and someone still has to exclude them by hand if that is the intention. See Exclude or re-include a subject in the eCRF.
The same three severities behave differently in the ePRO, where Exclusion does stop the participant. If your study uses both modules, do not carry the ePRO behaviour across in your head.
Write the message for the person who will read it
The message is the whole value of the rule. The site staff member reading it has your question in front of them and no idea what rule fired.
Say what looks wrong, not that something is wrong. Age is outside the inclusion range of 18 to 65 beats Invalid value.
Say what to do next: check the source, correct the entry, or confirm the value is real.
Keep it short enough to read in the flow of data entry.
Test the rule before the study goes Live
A rule that never fires and a rule that always fires look identical from the Builder. Testing is the only way to tell them apart, and once the study is Live the Builder closes.
š ļø Step by step guide
Keep the eCRF Builder open.
Open a second browser tab on the same study.
In the second tab, go to the Inclusions list and open a test subject's eCRF.
Go to the form holding the questions in your rule.
Enter a value that should break the rule, and check the message appears.
Check a query is raised on the data point.
Correct the value, and check the query resolves.
š” Tips
Work with both tabs open. Change the rule in the Builder, reload the test tab, see the result immediately.
A validation only re-runs when one of the data points in it changes. To re-test, re-select or retype at least one of the answers in the rule.
Test the case that should pass as well as the case that should fail. A rule that fires on everything is worse than no rule, because sites learn to ignore it.