Five Measures, Five Kinds of Evidence
The influenza PHSM recommendations look inconsistent if they are read as a single ranking: one measure is strongly recommended, another is conditional, and another is recommended against. They become more intelligible when each is judged by its target, timing, evidence certainty and burden. This chapter compares masks, ventilation, isolation, school measures and travel restrictions to show how one conditional model produces different conclusions without creating a universal effectiveness ladder.
Masks, ventilation and isolation answer different questions
Masks: a modest point estimate is not a universal percentage
The influenza evidence for masks is unusually useful because it contains a concrete pooled estimate and a clear boundary around that estimate. WHO’s review included 12 randomized controlled trials in community settings, involving more than 20,000 participants. The pooled point estimate suggested a 12% relative reduction in laboratory-confirmed influenza among people in the mask-wearing groups, expressed as a risk ratio of 0.88. The 95% confidence interval ranged from 0.61 to 1.27, and the review judged the evidence insufficient to exclude chance as an explanation for the apparent reduction.
What does that support? It supports describing a small-to-moderate possible effect under the trial comparisons, with moderate certainty overall and a plausible mechanism. It does not support saying that every mask policy reduces influenza by 12%, or that 12% of infections will be prevented in any given community. The estimate is relative, not an absolute number of infections avoided. An absolute result would require the baseline risk in the population being considered. The interval is also part of the evidence: it includes a range from a potentially meaningful reduction to little or no reduction in the comparison studied.
The use case matters. WHO conditionally recommends that symptomatic or positive individuals wear masks when in contact with others, and suggests community masking in crowded indoor settings during epidemics and pandemics of high or extraordinary severity. The first use is principally source control: the infectious person reduces the material reaching other people. The second may combine source control with some protection for the wearer, but its effect depends on how consistently people use masks, which masks are available, and how much transmission occurs through other routes.
The recommendation also acknowledges harms and unequal effects. Prolonged use can create discomfort or communication barriers, particularly for people who rely on facial cues or lip reading. Affordability and supply affect access, and adherence may vary by population and setting. These considerations do not disprove a mask effect. They explain why a study estimate cannot be detached from the people, conditions and purpose to which it applies.
Ventilation: strong guidance with low-certainty direct evidence
Ventilation illustrates a different evidence pattern. It acts on indoor air by increasing the dilution and removal of infectious particles. That mechanism is plausible and often low-disruption, but directly measuring its effect on influenza transmission across real buildings is difficult. WHO strongly recommends increasing mechanical or natural ventilation in households and non-residential buildings, while rating the evidence as low certainty.
The influenza evidence base included three studies. One compared residential buildings with different outdoor-air systems and observed fewer laboratory-confirmed respiratory-virus infections in the building with a dedicated outdoor-air supply. Another linked lower ventilation rates in university rooms with more self-reported respiratory illness.