REDCap Variable Naming
It is recommended that variable naming follow the convention below, but it is not required. A section of a variable can be omitted if it is not applicable, for example, if a form does not have multiple sections or questions then the variables do not need to be numbered.
Due to the limitations of some statistics programs not being able to use variables longer than 26 characters, all of the variable components below are typically shortened words. They are shortened either by using the first part of the word, an acronym for the word, or using the full word with vowels removed. Use your own discretion to best maintain the word’s meaning while keeping the variable names at a reasonable length.
Form_Section_Question_Number_Desc
Form | Part of the variable used to designate which instrument/form a variable is on. | |
demog_ | Demographics | |
mh_ | Medical History | |
sf_ | Self-Reported Questionnaire | |
Section | Part of the variable used to designate which section a variable is in. This part is applicable if a form has multiple sections. | |
sf_conmed_ | Concomitant medications on a self-reported questionnaire | |
cg_seprof_ | Drug side effect profile questions on a caregiver-reported questionnaire | |
Question | Part of the variable used to designate the question itself, this can be multiple parts if there are multiple parts to a question (i.e a sub-section). This part is applicable if questions are not numbered or if you want to have the variables be more representative of the question text. | |
demog_date | Date the demographics form was completed | |
sf_mh_cndtn_text_ | Medical History item number 2/10 on a self-reported questionnaire | |
Number | Part of the variable used to designate the question number or item number; it should contain enough preceding zeroes to keep the variables’ lengths the same throughout the form. This part is applicable if a questionnaire has question numbers or if creating a log where multiple sets of the same variables are needed. | |
cg_mh_cndtn_text_002 | Medical History condition 2/10 on a caregiver-reported questionnaire | |
gad_005 | Question 5 on the GAD-7 | |
Desc | Part of the variable used to designate if the variable is a Descriptive Text field. Typically, these variables will be named the same as the question variable they are associated with, with this suffix added. | |
sf_psqi_001_desc | Question text for PSQI Q1 on a self-reported questionnaire | |
gad_desc | Questionnaire instructions for the GAD-7 |
Option Coding
To ensure consistent coding of common options across Brain-CODE, some standard coding options were created. These should be followed unless there is a valid reason for an exemption. This coding should still be used if there are options that are similar but not the exact same, for example, “Do not know” and “Unknown”.
Coding | Field Option Label |
9996 | Other |
9997 | N/A |
9998 | Prefer not to answer |
9999 | Unknown |
REDCap Project Validation Checklist
Common Data Elements (CDEs)
All relevant common data elements are added to the study
CDEs are the most recent version(s)
No changes have been made to the original
No CDE variable names were used for another eCRF
The Informed Consent Form (ICF) contains SQL consent field
ICF contains a Brain-CODE subject ID naming checker
Electronic Case Report Forms (eCRFs) and Patient-Reported Outcomes (ePROs)
Consistent formatting
Fonts (type, colour, size) are consistent throughout and are legible
No spelling errors
All units are included (time, temperature, height, etc)
Option coding matches paper versions
Validation
Limited number of free text fields
Numeric fields only allow for integers or double numbers
Email address fields have email validation
When applicable ontology validation is used
Tested all fields with validation checks to ensure that they were functioning
eCRF/ePRO Variables
Consistent naming (eCRF has a prefix for all variables)
Numbers in variables have leading zeroes
No spelling errors
Naming is as intuitive as possible
Consistent abbreviations
Field labels are accurate
Fields have accurate descriptions and notes
Branching Logic
No errors due to branching logic
Branching logic is designed to mimic paper CRF
All branching logic has been tested for logic and correctness
Calculated Fields, Range Checks and Automated Queries
All calculated fields were manually calculated to ensure accuracy of the equation used
When applicable add range checks
For example - if the study is for participants older than 18, may have a range validation for age
Automated queries were tested for accuracy
Ensured no false positives
Ensured no false negatives
Range checks were tested for accuracy
Required Fields
All fields that are required are marked wherever possible
Fields are not inappropriately marked as required – causing errors in saving forms
Fields are properly designated as required – as to not prevent survey respondents from properly filling out ePROs
Survey Formatting
Autonumbering is turned off when ePROs are already labelled with custom numbering
If required, proper “STOP” actions are included
If required, hidden fields/read-only fields are included and properly formatted
Survey is appropriately formatted for its length
For example - one page for short ePROs, multiple pages for long ePROs
Font of Survey is legible and uses default OBI color scheme
Instructions for survey are available
Survey administration matches protocol
Study Design
Correct Data Access Groups for sites
Each eCRF has been assigned to at least one event in the Event Schedule Grid
Event grid matches study protocol
Arms are appropriately labeled and w/o spelling errors
Each arm has at least one event
Data entry workflow intuitive and as simple as possible