Comment on "Can Quantitative Pupillometry be used to Screen for Elevated Intracranial Pressure? A Retrospective Cohort Study".

If one were to have walked into the room of a patient and flipped a fair coin instead of performing a pupillometry measurement, the negative predictive value of a coin toss test would be 93% - the overall rate of normal ICP in the study population. It's hard to imagine a brain injury population whose risk of elevated ICP is many times greater than these patients but for whom it would be appropriate to defer invasive ICP monitoring in favor of a screening test with 70% sensitivity. A diagnostic test that yields a 3.7% absolute improvement over a coin toss in differentiating elevated and normal ICP in the relevant population while producing manyfold false positive results for every true positive is not suitable for informing high-stake clinical decisions. [Extracted from the article]