“…Investigating the association between ETT size and clinical outcomes is difficult, partly because randomizing patients to small or large ETT sizes is inappropriate and unethical. The propensity-matching nearest neighbor approach used in this study adjusted for potential confounders; but it cannot correct for unknown influences—including the judgment of clinicians whose intimate patient knowledge or inclinations can modify decisions in ways not accounted for by calculated scores. Uncaptured data on disease severity, airway abnormalities, organ failure, prior use of high-flow oxygen, or other unmodeled factors could systematically differ between groups.…”