Tip of the Week Archive

Inconsistent Outcome Measure data

• The Outcome Measure title and description need to accurately describe what the numbers reported in the data tables represent.
• Problem Example: The Outcome Measure title and description indicates that the measure to be reported is “Participant a1c levels”, but the numbers reported in the data table appear to be counts of participants.
• Problem Example: The Outcome Measure title and description indicate that the measure to be reported is “Change in BMI from Baseline at Week 6”, but the numbers reported in the data table appear to be the BMI score measured at baseline, rather than a change in scores.

 

Unclear Outcome Measures

• There needs to be enough information in the Outcome Measure description to understand what the numbers mean.
• If using a score or scale, include the full scale name, what it measures, the range of possible scores, and what higher or lower scores mean (i.e. whether higher scores = a better outcome).

 

Inconsistencies between different parts of the record without a valid explanation

• Problem Example: An unexplained mismatch in the number of participants in “Protocol enrollment” and the number of participants that “Started” the study in the Participant Flow module.
• If participants dropped out after consenting (enrollment) but before being randomized (starting), explain why in the “Pre-assignment Details” field
• Problem Example: A pre-specified Arm/Group is not reported for one of the Outcome Measures without a valid explanation.
• If you intentionally didn’t analyze an Arm/Group for an Outcome Measure, enter the “Number of Participants Analyzed” as “0” and explain participants in that arm were not included in the “Analysis Population Description” field.
• Problem Example: The number of participants analyzed for an Outcome Measure doesn’t match the number of participants assigned to that Arm/Group, and no valid explanation is provided.
• If some participants in an Arm weren’t analyzed, exclude them from the “Number of Participants Analyzed” for that Outcome Measure, and explain why they were not included in the “Analysis Population Description” field.

 

Incorrect Measure Types

• Example: if you are reporting a number of participants in your Outcome Measure data table, the Measure Type should be “participants”, not “number”

 

Incorrectly presenting Arms/Groups

• Participants can only be assigned to one Arm/Group. You can never double-count participants or assign them to multiple groups. If you have a crossover study, you would avoid double-counting participants by naming your arms something like Arm 1: “Drug A, then Drug B”, Arm 2: “Drug B, then Drug A”.
• Likewise, combining arms is problematic. Baseline Demographic data, Outcome Measure data, and Adverse Events need to be reported “Per Arm”, so combining two arms into “All Participants” and reporting Adverse Events for all arms together rather than reporting the adverse events for each arm separately will return a Major Issue during QA review.

 

For help addressing QA comments or submitting results, contact clinicalresearchsupportcenter@ucdenver.edu.