Meaningful-Change Thresholds in CONNECT-FX Were Specific to Each Behavior Scale
CONNECT-FX did not define a behavioral responder by any improvement whatsoever. Its meaningful-change thresholds differed across social avoidance, irritability and social unresponsiveness/lethargy.
Caregiver anchors informed the thresholds
The researchers combined caregiver severity and change ratings, score distributions and qualitative caregiver research. They identified decreases of at least three points for social avoidance, nine for irritability and five for social unresponsiveness/lethargy as meaningful within-patient changes over twelve weeks.
A responder proportion differs from a mean change
A mean score can move without every participant crossing a threshold. Conversely, some participants can cross a meaningful threshold even when the overall between-group test is nonsignificant. In the full analysis set, responder comparisons were not statistically significant for the three cited domains; the irritability result was near, but did not cross, the conventional threshold.
Thresholds do not establish treatment superiority by themselves
A threshold helps interpret what changed for an individual on a particular instrument. Evidence that CBD produced more such changes than placebo requires the corresponding controlled comparison. These thresholds are not percentages of symptom elimination or criteria for cannabis impairment. They should not be transferred unchanged to a different diagnosis or rating scale. Reporting the scale, direction of change and responder definition prevents a numerical improvement from being labeled a proven CBD treatment effect.
Related analysis: the trial’s main results and limitations.
