Jennings JH, and Yessayan L. Response. Chest 2015; 148(3):e91-92.
We appreciate the comments by Dr Zaidi and colleagues on our recent article in CHEST.1 They ask whether the effect size in our study from 20% to 10% reduction in readmissions was overly optimistic. There is a paucity of data in the literature to guide us as to what an expected magnitude of reduction in 30-day readmission rates should be. Importantly, choosing an effect size should be based on what is “clinically meaningful.” Unfortunately, this can oftentimes be challenging and subjective. In the absence of a specific intervention, rates of 30-day readmissions for COPD range from 7% to 22%.2 We, therefore, chose an absolute reduction of 10%, given our baseline prestudy admission rate of 20% at Henry Ford.Would choosing a smaller effect size have shown us a difference between groups? Perhaps not. While the point estimate of the risk difference was −3.5%, the CI shows that true risk difference may very well be +8.8% in favor of no intervention. Nonetheless, had the larger sample size resulted in statistical significance for this small risk difference, one might question the clinical significance of an intervention as only marginally favorable.