In an article recently published in the Archives, Kazaure et al1 found that a patient's do not resuscitate (DNR) status was an independent risk factor for surgical mortality. This article demonstrates one of the key challenges when interpreting the results of statistical models: understanding when the presence of a risk factor truly causes the outcome. In this study, their finding that DNR status is an “independent risk factor” for mortality is reasonable based on their regression model output, but it makes little sense from a clinical perspective. Unfortunately, discrepancies between statistical modeling and clinical reality are common in the surgical literature, and readers should be cognizant of them.
Livingston EH, Dimick JB. Inferring Causation From Multiple Regression: Do Not Resuscitate? Arch Surg. 2011;146(8):898. doi:https://doi.org/10.1001/archsurg.2011.219
Customize your JAMA Network experience by selecting one or more topics from the list below.
Create a personal account or sign in to: