Equations for converting scores between depression scales (MÅDRS, SRS, PHQ-9 and BDI-II): good statistical, but weak...

A variety of depression rating scales are currently used in clinical and research practice. As these scales are generally thought to correlate well, there may be some benefit in deriving equations to translate the score on one scale to that on another.

Using pairwise ratings, we compared the Montgomery-Åsberg Depression Rating Scale (MÅDRS), Beck Depression Inventory II (BDI-II), Patient Health Questionnaire (PHQ-9) and Zung Self Rated Scale (SRS). The primary comparisons of interest were as follows: PHQ-9 and MÅDRS, PHQ-9 and BDI-II, SRS and MÅDRS, and SRS and PHQ-9. One hundred and fifty pairs of ratings were collected for each of these four comparisons, and these were used in a regression analysis to generate a predictive equation between scale pairs. The predictive equations were then tested for accuracy by using novel data.

There was a reasonably strong correlation between scales. Simple regression equations describe the relationships between certain permutations of the BDI-II, PHQ-9, SRS and MÅDRS well enough to allow the score on one scale to be estimated from the score on another. On average, the equations work well. However, when used to predict individual cases, they often fail.

It does appear that conversion equations for depression scores are precise enough to be of use when applied to averages; these might aid comparison of data across different studies, meta-analysis or instrument selection for clinical trials. But regarding scale translation at the idiographic level, the variability we have observed is so great that severe doubt is cast on the validity of such an approach.