Do surgical masks reduce transmission of respiratory viral infections when worn by the sick person (source control)?

Yes, surgical masks cut what a sick person breathes out, but the proof is in viral shedding and aerosol counts, not in trials measuring who actually gets infected.

The best direct evidence is a study of people with acute respiratory illness breathing into a sampler with and without a surgical mask [1]. Masks eliminated detectable coronavirus in both droplets and aerosols, and cut influenza in droplets, but the numbers were small and some results did not reach statistical significance [1]. Notably, influenza in fine aerosols and rhinovirus in both fractions were barely reduced, so the effect is virus-dependent, not universal [1].

A 2024 hospital study of 230 infected patients found viral RNA on nearly 8 in 10 of their masks, with the highest load at the inner nose site, showing masks catch virus but also that leakage around the mask is real [2]. In adults with cystic fibrosis coughing, surgical masks cut viable airborne bacteria at two meters by about 94%, far better than cough etiquette alone at about 53% [5]. On the mechanical side, surgical masks filter roughly 81% of particles in one bench test, and all eight tested mask brands let some aerosol pass [4][3].

What this retrieval cannot see: a randomized trial with actual transmission (who gets sick) as the endpoint. Everything here is a surrogate, virus caught in a sampler rather than infections prevented. That is a real gap, and it means the honest answer is "masks plausibly reduce source control, proven at the shedding level, unproven at the outcome level." No serious harm shows up in these records; the main cost is imperfect filtration and fit.

My call: yes, surgical masks worn by sick people substantially reduce expelled virus for some viruses, but the evidence stops at shedding and aerosol counts, not confirmed transmission reduction. Confidence: moderate.