Some standardized questionnaire items were inapplicable to participants with substantial functional limitations. The resulting invalid scores reduced usable data and statistical power rather than constituting observed lack of improvement.
The trial included a broad population
Participants ranged from five to twenty-one years old and had varied functional abilities. The authors reported that standardized questionnaires contained many inapplicable items for some participants, producing invalid scores. They identified this problem as a limitation of the outcome measures.
A participant could remain enrolled and receive treatment while lacking a valid score on a particular instrument. The enrollment count therefore does not establish the analyzable denominator for every scale.
Invalid is different from unchanged
An unusable questionnaire cannot be assigned a zero change merely because its intended behavior could not be rated. Nor should missing valid scores be assumed to conceal a benefit. The scale’s applicability and data-handling rules determine what the comparison actually represents.
Compare outcomes carefully
Mixed results across clinical ratings and questionnaires may reflect different constructs, sampling and measurement limits; this paper does not isolate one explanation. Report the available denominator and instrument for each finding. A favorable result on an applicable measure cannot validate an inapplicable instrument or establish improvement across every functional level represented in the original enrollment.
Related reading: limits of proposed CBD mechanisms.
Research cited
Aran and colleagues (2021). Cannabinoid treatment for autism: a proof-of-concept randomized trial. DOI: 10.1186/s13229-021-00420-2.
Leave a Reply