Question validation

A question's validation conditions allow you to detect entered values that are not within an expected range or do not comply with certain rules. For example, you may want to detect a value which is not between 100 and 300, or a date value which is after a date on another eForm. When you specify a validation condition, you also specify a message that will be displayed to the user if they enter a value which satisfies that condition.

If a question has several different validation conditions they will be tested in the order that they appear in the Question Definition window. All conditions that are true will be displayed to the user at the time of data entry, but only the first one can be overruled, and only the first one is saved with the response in the database. If more than one validation fires, the second and subsequent validation messages are not saved.

Question validations are only run after the user has entered a response value. If a question field is left blank, any validations defined for that question will not run.

Questions are automatically validated every time that an eForm is opened by a user with the Change Data permission and every time that an eForm is saved. It is also possible to revalidate all eForms for single or multiple subjects within the Data Entry module.

You cannot define validations for lab test questions, because they are validated against laboratory normal range information.

By default, if a warning validation fires for a derived question, the warning icon will display but the associated warning message will not. The reason for this default behaviour is that the data entry user is not able to edit the value in a derived question and so cannot make any changes in response to a warning. However each data entry user can switch on warning messages for derived responses via Tools > Options > eForm.

You should not add or change 'Reject if' validation conditions once the study has been opened for subject data entry.

The Arezzo Terms report in the Study Data Diagnostic Utility module can be used to list all existing validation conditions in a study definition.

If you are concerned about the risk of data entry users repeatedly entering different data into a question until it is accepted , we strongly recommend that you use Source Data Verification to check specific subject data against the original information source. This will confirm the accuracy and integrity of the data that is required for GCP compliance.

There are 3 levels of validation:

Reject if...

If data is entered that satisfies a 'Reject if' condition, the validation message is displayed to the data entry user, and the value is not be accepted. The data entry field automatically reverts to the previously stored value unless the value is derived, in which case the field is cleared.

Warn if...

If data is entered that satisfies a 'Warn if' condition, the validation message is displayed. The data entry user can choose to:

accept the warning and keep the value in which case the 'Warning' question status icon will be displayed
overrule the warning and keep the value in which case the 'OK Warning' question status icon is displayed
enter a different value which does not generate a warning.

Flag if...

If data is entered that satisfies a 'Flag if' condition, users with the 'Monitor/review data' permission see the 'Inform' question status icon and the validation message. Users without this permission will be unaware of the condition.
Questions with the 'Inform' status cannot be overruled.

Open the question definition window for the question.
Select the Validation tab to display the Validation window. See image You may alter the column widths by placing your cursor on the heading intersection between two columns. When the cursor changes to a double-headed arrow, drag the intersection to the left/right to reduce/increase the column widths.
Click in the next row in the Validation Type column and select the type from the drop down list. See image
Click in the same row in the Validation Term column and enter a validation condition, either by typing directly into the field, or by clicking the fx button and using the Expression Builder. See image If the value entered by the data entry user satisfies this condition, the validation will be triggered.
Click in the same row in the Validation Message column and enter the message that you want to display if the condition is satisfied. You may type directly into the field, or once again launch the Expression Builder. See the section below on validation messages.
Repeat steps 3 to 5 for each condition that you wish to attach to the question. You may attach as many conditions as you wish to an individual question, although only the first condition to be satisfied will fire. You should not rely on conditions being evaluated in any particular order: instead you should make your conditions mutually exclusive where possible.
Click OK to save your work, or Cancel to abort.

Open the validation tab in the Question Definition window - see steps 1 and 2 above.
Left-click on the grey column to the left of the validation conditions in the row you want to delete. The row will be highlighted.
Press the Delete key on your keyboard. The row will be deleted.

A validation message can be free text (which must be enclosed in single quotes), or an Arezzo expression.

For example, you may enter:

'The value ' & me:value & ' is too high. It must be less than ' & screening:baseline:basevalue

When displayed, this message will indicate that the value entered needs to be less than the value of basevalue from the baseline eForm in the screening visit.

If you reference a date in a message, it will automatically be formatted as yyyy/mm/dd. If you want to use a different format, use the format_date function.

If you use an Arezzo expression to create the message and that expression references data that is missing at the time the validation is fired, then the actual Arezzo expression will be displayed in the message instead of the value, which can be disconcerting for the data entry user. To avoid this, only reference data in the message if it is also referenced in the actual validation condition or you are sure that it will have a value. Alternatively, you could use a function such as isknown within the Arezzo expression.

Remember that if you want to use a text message which include a single quote, you must type it as two single quotes, for example

'This subject''s pulse is too high'

Use the Arezzo term checker to check all the validations within your study.