The baseline pain variability (BPV) has often been presented as positively correlating with the placebo response (PR) and associated with a lack of consistency in the subjects’ pain evaluation. Excluding high BPV subjects should then improve the precision of the treatment response. Another common method to increase the essay sensitivity is to adjust the analysis for covariates associated with the response. We aimed here at optimizing this essay sensitivity while combining both the exclusion of high BPV subjects with adjusted analysis.

Design:

We used the data from a randomized, placebo-controlled study with 171 osteoarthritis subjects. Up to 25% of the subjects with the highest BPV were excluded. In parallel, an adjustment for the Placebell placebo covariate was tested and the improvement in assay sensitivity was estimated.

Results:

The subjects’ exclusion didn’t significantly change the endpoints variance in the study, despite significant correlations between BPV and endpoints. On a positive note, this exclusion didn’t attenuate the performance of the adjustment which improve the essay sensitivity by up-to 37%.

Conclusion:

The high BPV subjects exclusion can be used in combination with covariate adjustment to improve the essay sensitivity. Surprisingly, however, this exclusion didn’t seem to have a strong impact on the endpoints’ variance in the study. Overall, the adjustment seemed to bring larger gains in precision and power than the high BPV subjects exclusion.