Time to Review Air Quality Activity Guidelines
Sumil K. Thakrar
DOI: https://doi.org/10.1001/jamanetworkopen.2024.5259
2024-04-09
JAMA Network Open
Abstract:One strategy to reduce the health burden of air pollution is to not pollute so much; another is to limit how much pollution people breathe in. The US has pursued both emissions-related and exposure-related interventions. Twenty-five years ago, the US Environmental Protection Agency (EPA) first promulgated activity guidelines, an exposure-related intervention that advocates remaining indoors and avoiding vigorous exercise on days when air pollution levels are unhealthy. Even relatively clean air is associated with health harms, 1 but what counts as unhealthy is codified by the EPA Air Quality Index (AQI), a mapping of criteria pollutant concentrations to categories such as good, unhealthy, and hazardous, that is widely used for public health messaging. Over the last 25 years, the effectiveness of emissions-related interventions has received due scrutiny. The air in the US is far cleaner than it was 25 years ago, although the trend appears to be stagnating or reversing in recent years. There have been far fewer inquiries into the effectiveness of the AQI activity guidelines, of which the study by Brook et al 2 is in the minority. This lack of inquiries exists despite the guidelines tightening (lower pollution levels now count as unhealthy) and affecting more people (eg, the elderly population, who are subject to stricter guidelines, has grown). Brook et al 2 found that current AQI activity guidelines are very ineffective for all but the most extreme pollution events. It would require many thousands of people (and for some guidelines, millions) to adhere to the guidelines before even 1 adverse health event, such as a pulmonary or atherosclerotic cardiovascular disease (ASCVD) event, is avoided. Every avoided pulmonary or ASCVD event is a blessing, but interventions come with costs as well as benefits. For activity guidelines, 2 broad categories of costs are messaging-related (eg, reporting AQI and distributing alerts) and adherence-related (eg, health harms from avoided exercise 3 and lost work from staying indoors). To evaluate effectiveness, Brook et al 2 asked what number of people need to be treated (NNT, an informative, widely used public health measure) to reduce 1 adverse health outcome. The choice of treatment group warrants some care. Brook et al 2 straightforwardly defined the treatment group as those who are implicated in the activity guidelines (ie, those intended-to-treat), but not everybody who is implicated in the guidelines adheres to them. Effectiveness is sensitive to assumptions about adherence: if nobody adheres to the guidelines, the number treated approaches 0 and the NNT (as defined by Brook et al 2 ) skyrockets. However, if people self-select into adherence based on private information about their health status, 4 and if we take the treatment group to be those who adhere rather than those who are implicated, the NNT will be more muted. It is likely that adherence-related costs dominate messaging-related costs, in which case, the more appropriate treatment group is those who adhere rather than those who are implicated. In that case, if it is true that the only people who bother to adhere to the guidelines are the ones who really need them, this does not sound quite as ineffective as Brook et al 2 report. However, the NNT values presented in this study 2 were so high that it is hard to exonerate activity guidelines from the charge of ineffectiveness. Strategically targeted messaging may increase adherence among the highest-risk groups. Taking adherence seriously could strengthen the point Brook et al 2 make. To derive the NNT results, Brook et al 2 inferred changes in health risks from changes in activity, which raises a methodological concern. The associations of air pollution with health in the study 2 were derived from associations of ambient pollutant concentrations at a location with health outcomes of people living nearby who pursue all kinds of activities, including potentially compensatory activities, to mitigate exposure during high-pollution events. 4 We do not, in general, observe such activities, so it is troubling to reason counterfactually about them. A less arcane example illustrates this point. Suppose there is no association of how much it rains with how wet people get because the more it rains, the more people use umbrellas. If you observe the lack of association, but not umbrella use, you might erroneously conclude that umbrellas are unnecessary, because rain does not lead to wetness. The worry is that recommending lax activity guidance is similar to advocating against umbrellas. Admittedly, there is little evidence that the associations have changed over decades, even when we expect activities to have changed. In particular, except for some noteworthy studies, 4 there is li -Abstract Truncated-
medicine, general & internal