Published in:
27-05-2022 | Visceral Ischemia | Correspondence
Effect modification in a clinical trial should be assessed through interaction terms, not prognostic modelling
Authors:
Roberta M. L. Roepke, Pedro V. Mendes, Luis C. M. Cardozo Junior, Marcelo Park, Bruno A. M. P. Besen
Published in:
Intensive Care Medicine
|
Issue 8/2022
Login to get access
Excerpt
We read with interest the secondary analysis of the NUTRIREA-2 trial published in Intensive Care Medicine [
1]. It has been discussed by colleagues because of the concern that it could lead to wrong conclusions and less utilization of enteral nutrition in patients on vasopressors, given a low risk of severe gastrointestinal adverse events [
2]. Despite the authors’ rebuttal [
3], we are still afraid a wrong message could be passed to the readers with a spin in the trial interpretation due to this secondary analysis. …