CBL-514 Missing Data: Who Is Included in the Analysis?
Follow the participant flow
The 2026 CBL-514 abstract reports 108 randomized adults but smaller denominators for specific rating outcomes.[1] That difference is a reason to consult the methods and participant-flow information, not to silently substitute one count for another.
Our recommended extraction sheet has separate fields for randomized, received intervention, completed follow-up and contributed to the named analysis. Record the count by group and visit. A participant can contribute to one endpoint while lacking a usable observation for another.
Intention-to-treat is not an imputation method
Analyzing people according to their randomized group and deciding how to handle missing outcomes are distinct choices. A label such as βfull analysis setβ also needs its study-specific definition. Read the inclusion rules rather than assuming that a familiar abbreviation guarantees an identical population across papers.
FDA’s estimand guidance emphasizes defining the treatment effect of interest and examining sensitivity to assumptions.[2] For a CBL-514 review, ask what question the analysis answers if a participant stops treatment, misses a scan or begins another intervention. These situations need not be handled in the same way.
Do not infer why observations are absent merely from a reduced denominator. The explanation may involve withdrawals, unavailable images, protocol criteria or another documented reason. If the source does not explain it, preserve that gap.
An original sensitivity exercise
Illustrative only: imagine 20 participants assigned to a group, with 12 responders, four nonresponders and four unknown outcomes. Among the 16 observed participants, 75% respond. If the four unknown outcomes are counted as nonresponse, the proportion becomes 60% of all 20. These are different assumptions, not competing arithmetic answers.
This simple example is not a substitute for a trial’s statistical analysis. It shows why a percentage should travel with its denominator and missing-data rule. A robust evidence summary describes the primary approach and any reported sensitivity analyses, including whether they meaningfully change the conclusion.
When comparing CBL-514 reports, use the same analysis definition whenever possible. If that cannot be done, explain the mismatch instead of presenting the percentages as directly comparable.
Frequently asked questions
Can a missing visit be treated as a treatment failure?
Only when that rule is part of the specified analysis or a clearly labeled sensitivity exercise. Do not add it silently.
Does a smaller analysis set automatically invalidate a study?
No. Its definition, the reasons for missing data and the sensitivity of conclusions determine how the limitation should be interpreted.
Continue your CBL-514 research
Compare the original study evidence with the documentation for the specific material you are reviewing.
