In this issue of CHEST (see page 890), Hurley1 performed a multilevel random effects analysis examining topical antibiotics (TAs) for the prevention of ventilator-associated pneumonia (VAP). Because TA use can confer herd protection in the ICU similar to vaccination programs in the community, contextual influences resulting from a population-based intervention cannot be estimated from a single trial. However, multilevel random effects analysis allows the estimation of contextual effects. Hurley1 found that the baseline incidence of VAP derived from observational studies was lower (23.7%; 95% CI, 20.6%-27.2%) than that in studies of TAs using concurrent control groups that either did or did not receive topical placebo (38% [95% CI, 29%-48%] vs 33% [95% CI, 20%-50%], respectively). This observed contextual influence could potentially inflate the apparent effect of TAs, especially within studies using topical placebo. The clinical importance of this observation is illustrated by investigations showing that TAs can promote the emergence of antimicrobial resistance and increase the burden of resistance genes in the gut biome of patients in the ICU.2,3 Without knowing the overall influence of TAs on antimicrobial resistance progression and clinical outcomes, their routine use cannot be endorsed, especially in areas where antibiotic resistance is already a clinically important problem.