Table 1

Criteria for labelling presence and correctness of interpretation of an OR

A substantive effect size interpretation (interpretation), for the purpose of this study, is a statement about the magnitude of change in the outcome as a function of change in the predictor from the reported OR. Solely reporting the OR or its directionality was not considered a substantive interpretation. A correct substantive interpretation is an interpretation that accurately reflects the definition of the OR, as described in commentary by Davies et al6 and Norton et al.7 For example, if the OR were 1.5, the interpretation “50% increase in odds” would be correct, while “50% more likely” would be incorrect. Correct interpretations could also include other interpretations resulting from logistic regression that use the OR as an intermediary (eg, change in probability, marginal risk).
Example phrasesLabelReason
“… male sex was associated with seeking treatment (OR=2)…”No interpretationThe OR was presented as a parenthetical statement only.
“… was associated with decreased odds (OR=0.5) …”No interpretationOnly the direction of the association was reported.
“… were three times more likely (OR = 3) …”Incorrect interpretationAn interpretation was made by incorrectly expressing ‘odds’ as ‘likeliness’.
“… was associated with a 30% reduction in the log odds (OR=0.7) …”Incorrect interpretationAn interpretation was made by expressing the ratio of log odds but reported the OR.
“… were associated with a threefold increase in the odds (OR=3) …”Correct interpretationAn interpretation was made by expressing the ratio of odds.
“… each was associated with a 10% reduction in the odds of treatment failure (OR=0.90) …”Correct interpretationAn interpretation was made by expressing the ratio of odds.