/
Best Practices in Production Mode

Click on the magnifying glass to the right. Starting typing a topic you wish to learn about

Best Practices in Production Mode

Once the project has been moved to production, you will not be able to easily edit the project fields!


Design changes after the Production phase are discouraged. You can lose data. Checking and testing the changes is harder than in the Build phase, and can result in data loss.

  • Some design changes must be re-submitted through the  REDCap administrators and will not be visible or take effect until the  REDCap administrators approve the changes.

  • Request the minimum number of design changes possible, and do so in a manner that will not destroy prior collected data, whether or not you believe you will use that prior collected data.  The data dictionary will not upload when you change the position of the field, edit the label, edit the variable name, or edit SQL code.

  • Always make a copy of the data dictionary before making any significant production changes.

  • Caution: the data import tool can override the existing data.

  • Changes to any variables may affect programmed calculations and/or branching logic. It is the responsibility of the requester to review and test all calculated fields and branching logic prior to submitting changes.


The below table provides a summary of the impact of project changes on your production data:

DATA IMPACT TABLE: TYPES OF REDCAP PRODUCTION PHASE CHANGES AND POTENTIAL DATA IMPACT



Metadata

Change Type

Data Impact

Further Explanation

Variable Name / Field Name

Add new

No Data Loss

The new field will be added to all records.

Variable / Field Name

Delete

Data Loss

This deletes the field and all the data entered for that field.

Variable / Field Name

Rename

Data Loss

This is equivalent to deleting a variable and adding a new variable, thus the data is deleted.

Form Name

Add new form

No Data Loss

New form/fields will be added to all records.

Form Name

Change via data dictionary upload

Data Loss

Form completeness data will be lost. Form names can be changed within the Online Designer without data loss.

Form Name vs. Form Label

Rename form

Possible Data Loss

Recommend renaming the Form Label rather than the Form Name. The data dictionary renames the form. All form status values (unverified, complete) for ALL records will then be reset to “Incomplete”. The Online Form Editor will NOT change the form name and only renames the form label, preserving the form status for all records.

You can re-name the survey title (if it’s a survey) within survey settings without changing form name or label

Note: Form name “back end” name (data dictionary ex:”baseline_data”) does not appear on screen.

Form label “front end” name displays on screen (ex: “Baseline Data”)

Field Units

Add, Modify, Delete

No Data Loss

This would merely change the field unit label.

Section Header

Add, Modify, Delete

No Data Loss

No data loss as it is descriptive text.

Field Type

Modify

Possible Data Loss

Depending on the change, data can be lost.

Examples of changes that can be made without data loss:

  • Radio button response sets to drop downs, or

  • Drop-downs to Radio buttons.

Examples of changes that can be made that cause data loss:

  • Radio dropdown buttons to check boxes, (and vice versa) or

  • Text box to calculated field.

Field Label

Modify

Possible Data Confusion

Changes to question caption may change the meaning of data previously entered. Simple spelling corrections or format changes are not problems.

Response Choices

Add

Possible Data Impact and Confusion

-If the choice added at the end of a list, no data is lost, but there could be confusion in the final data analyses. New choice will be added to all records.

Response Choices

Delete

Data loss

Deletes the choice and ALL data entered as that choice.

Response Choices

Recode

Possible Label Mismatch

Codes are not automatically re-mapped to new codes. Data entered remains the same in the database. Relabeling codes may change the meaning of data entered.

Calculations

Add, Modify, Delete

Data Confusion

Forms with saved calculated field values will not automatically recalculate when changes are committed. Values should be derived and confirmed in analysis. All forms with values should be resaved to update stored values.

Slider Labels

Add, Modify, Delete

Possible Data Confusion

If the changes impact how respondents answer the question, former data may not be consistent with data after the change.

Field Note

Add, Modify, Delete

No Data Loss

No direct data impact as it is descriptive text.

Text Validation Type

Add, Modify

Possible Data Loss

Data entered as free text or other type of validation text may no longer be valid.

Text Validation Type

Delete

Possible Data Confusion

Field becomes open text field.

Show Slider Number

Add, Delete

Possible Data Confusion

If the changes impact how respondents answer the questions, former data may not be consistent with data after the change.

Text Validation Minimum

Add, Modify, Delete

No Data Loss

No data impact since out of range data can still be saved.

Text Validation Maximum

Add, Modify, Delete

No Data Loss

No data impact since out of range data can still be saved.

Identifier

Add, Delete

No Data Loss

No direct data impact.

Branching Logic

Add, Modify

Data Loss

Fields that will be hidden due to updated logic but that already contain data will prompt data erasure.

Branching Logic

Delete

No Data Loss

No direct data impact, but may impact missing data analyses. Fields will remain visible.

Required Field

Add, Delete

No Data Loss

Data can still be saved without completion of required fields.

Custom Alignment

Modify

No Data Loss

Display only.

Question Number (surveys only)

Modify

No Data Loss

Display only.

Matrix Group Name

Add, Modify, Delete

No Data Loss

Display only.

Section Headers

Add, Modify, Delete

No Data Loss

Descriptive fields can be changed/deleted/added without any loss since they don’t store data

 Courtesy of Wright State website, authored by Drs. Marjorie Bowman and Rose Maxwell

This Knowledge Article was developed by the University of Utah’s REDCap Analysts using data garnered from the REDCap Consortium Library, Wright State's website and other REDCap Affiliate members’ online resources and documents.