To the Editor.
—Dr Connors and the SUPPORT Investigators1 provide thought-provoking information regarding apparent harm from the use of the RHC. The accompanying Editorial2 provides thought-provoking reading, and the ripples from this article have already been felt around the world.3 However, the investigators emphasize that this is not a randomized controlled trial, and the comparisons made within it rely totally on the accuracy of the propensity score, derived to predict the likelihood of placement of an RHC. Sensitivity analysis showed that for a single missing covariate in the propensity score to have misrepresented a true beneficial effect as harmful it would have to increase estimates of risk of death and likelihood of RHC 6-fold. Is this possible?There does seem to have been a missing x factor. The 2016 patients in the study had almost identical propensity scores and disease categories, but only half received an RHC. This was unlikely
Boyd O. Effectiveness of Right Heart Catheterization: Time for a Randomized Trial. JAMA. 1997;277(2):110-111. doi:10.1001/jama.1997.03540260024019