Recent work by Ellenhorn and coworkers tested the idea that preoperative scrubbing of skin with povidoneiodine soap adds no incremental protection against wound infection. Two cohorts (n 115 and n 119) were created by randomizing abdominal surgery patients to “paint-only” or “scrub-plus-paint” categories. If the authors were fastidious in their experimental execution, if randomization was “perfect” (ie, risk factors balanced), and if infection surveillance accuracy was the same for each cohort, comparing infection rates for the two experimental cohorts might support a generalization about the value of a povidone-iodine scrubbing step. The authors observed an incisional infection rate of 10% for each of their cohorts and essentially identical intraabdominal infection rates (3% and 2%) as well. I worry that certain nuances in their interpretation of these findings could unintentionally confuse fatigued readers who might skim the article’s abstract after full days of elective operating or long nights on emergency call. Whenever inferential statistics are used to compare frequencies of some defined outcomes in two clinical trial cohorts, a thought experiment is being performed in hopes of sharpening knowledge of reality. Essentially, such cohorts are viewed as samples taken from two populations or universes. In the Ellenhorn study, one imagined universe was “all abdominal surgery patients on Earth receiving paint and scrub” and the other imagined universe was “all abdominal surgery patients on Earth receiving only paint.” A statistical maneuver estimates the range of rate differences between universes that are compatible with some observed cohort rate difference, which might be zero, as measured in the present case. The authors used a method described by Rodary and colleagues and obtained an nonsignificant p value, concluded that the skin preparation routines are equivalent, and called for the elimination of povidone-iodine scrubbing. The meaning of equivalence is critical in digesting this article. The authors declared before commencing the trial that an absolute infection rate difference as great as 6% would be deemed compatible with equivalence of the skin preparation routines, but no clinical rationale was offered for selecting that value. In addition, hasty readers can misconstrue the adjective absolute. Unfortunately, the authors did not emphasize that their results cannot rule out a large advantage of adding povidoneiodine scrubbing. For example, their data are compatible with an actual infection frequency of 14.5% in the paintonly universe and 7% in the paint-plus-scrub universe. In clinical practice, this would represent an infection risk differential of 50% and few infection prevention protocols can discount any tactic having that potential impact. Using the authors’ raw data, computation of an exact 95% confidence interval (with continuity correction) for the difference in infected proportion between universes reveals a considerably wide range ( .085 to .089). We can confidently conclude only this: povidone-iodine scrubbing might add benefit, might have no effect at all, or might actually increase infection likelihood. These are the same three possible knowledge states that existed before the authors’ experiment was performed. Maybe we should discard preoperative povidone-iodine scrubbing. Maybe we should not.
Read full abstract