I read with interest the Christakis editorial “Predictably Unhelpful: Why Clinicians Do Not Use Prediction Rules.”1 I agree with much of the editorial. However, I disagree that once a physician is concerned about a condition the degree of risk is not important.
A risk of 0.001% of a serious bacterial infection corresponds to a risk of 1 in 100 000. The risks associated with evaluating and treating 99 999 infants who do not have a serious bacterial infection to avoid missing the one who does are not negligible. There would be 99 999 well babies who would receive unnecessary venipuncture and likely unnecessary bladder catheterization, lumbar puncture, intravenous antibiotic administration, and hospitalization as well. The risk of adverse effects from any of these common procedures and medications is certainly very low, but each carries a risk that is likely higher than 1 in 100 000.2- 8 The combined risk is even higher.
Daymont C. Sometimes Zero Risk Is Not an Option. Arch Pediatr Adolesc Med. 2011;165(6):573-574. doi:10.1001/archpediatrics.2011.75