Skip to main content
We complement the authors for this wide-scale trial and all effort invested. We recently discussed this article in our journal club (made possible by the discussion tips, provided by Annals of Family Medicine). As a result, we have some points on which we would like to invite the authors to respond:
The authors conclude that the intervention did not have an effect on integration measured by the PIP score. But the PIP score (and some patient outcomes) improved more in the active practices that completed the intervention (workbooks) as compared to the active practices that did not complete the intervention.
One could say that this is a ‘per protocol’ analysis, instead of an intention to treat analysis.
The first remark: The authors compare practices that completed the intervention with active practices that did not complete the intervention. However, it might have been more appropriate to compare practices with completion of the intervention with the control group.
The second remark: From the introduction it seems that the intervention studied are implementation strategies which were found as best practice from previous studies. If this is indeed the case, including implementation outcomes and evaluating the perceived acceptability (agreeability of the intervention), adoption (intention to use the intervention) and appropriateness (perceived fit/ relevance of the intervention) of the healthcare professionals and fidelity of the strategy (toolkit) might have helped to understand the difference in effectiveness between the practices with or without completion of intervention.
A more general remark, also for future design of similar trials: Selection of below average practices (<75 PIP scores), lowers the power of the trial, because of regression to the mean. Regression to the mean implies that extreme observations tend to be less extreme, but closer to the population mean, when observed a second time.(1) This might be the reason of the improvement in PIP scores in the control group and could contribute to the non-significant effect of the intervention.
Sincerely,
W.J. van Dijk, MD,
M.E. Haas, PhD
R.C. Vos, PhD