Is Air Pollution Killing South Africans?

Jacobus Erasmus, 19 June 2019


A study has been circulating on the news in South Africa that air pollution is dangerous for our health. For example, the Vaal Weekblad has an article titled “Air pollution by Eskom and Sasol is killing South Africans – Study”. What a scary title!

The study being circulated is from Andrew Gray, an American air pollution consultant. In the paper, Gray states that “Ambient PM2.5 pollution from the 14 facilities [in the Highveld] caused between 305 and 650 early deaths in the area in 2016.” The term “PM2.5” or “fine particulate matter” refers to tiny particles in the air that have a width smaller than 2.5 microns. These tiny particles can originate, for example, from vehicle exhausts, power plants, and burning wood or coal. Thus, according to Gray, the Eskom power plants are producing levels of PM2.5 that can and are killing people. Again, what a scary claim!

But is this claim true? How, exactly, does Gray know that the air pollution is causing these deaths? An examination of Gray’s methodology on pages 21-23 reveals that his methodology is flawed.

First, Gray uses a mathematical model (based on Burnett’s et. al. model) to predict or estimate the health problems that may be attributed to certain exposure to PM2.5. Setting aside the issue as to how reliable the model is, it is fallacious to argue that, if something has been estimated, then it will or has been actualised. Essentially, Gray’s argument has the following form:

(1) The Highveld air pollution might have caused X number of deaths.
(2) Therefore, the Highveld air pollution did cause X number of deaths.

But (2) does not follow from (1); Gray’s conclusion does not follow from his estimate. This would be similar to me arguing as follows:

(1) I predict that South Africa will win the cricket world cup.
(2) Therefore, South Africa will win the cricket world cup.

Or:

(1) I might win the national lotto.
(2) Therefore, I will win the national lotto.

Such reasoning is fallacious. Accordingly, the most that Gray could have concluded from his study is that ambient PM2.5 pollution from the 14 facilities in the Highveld might have caused between 305 and 650 early deaths in the area in 2016.

Second, Gray’s methodology cannot rule out the possibility that non-pollution factors are the primary causes of the deaths. How do we know for certain that, for example, those who died did not get ill from smoking, or inherent heart issues, or some other health issues unrelated to air pollution? Gray’s predictive model or methodology simply cannot tell us that air pollution did, in fact, cause the deaths. Indeed, Gray succumbs to a well known problem in observational studies about air pollution: such studies are almost completely circumstantial. As Joel Schwarts, an expert in air pollution, says:

An embarrassment for proponents of low-level air pollution as a cause of death is that the evidence is almost solely circumstantial, being based on statistical studies reporting small correlations between long- or short-term air pollution levels and risk of dying. These “observational” studies are not based on randomized trials, but on non-random data that inherently suffer from confounding by non-pollution factors with much larger effects on health than the purported effects of air pollution. Observational studies could be taken more seriously if they were supported by evidence from randomized, controlled studies that eliminate the possibility of con-founding by non-pollution factors. Such studies cannot, of course, be done with people, but they can be done with animals. However, researchers have been unable to kill animals with air pollution at levels anywhere near as low as the levels found in ambient air (Air Pollution and Health, 2006, p. 5).

Consequently, Gray’s estimates, in reality, do not tell us very much.

Third, Gray’s study assumes the following proposition:

(P) The inhalation of PM2.5 produced from air pollution might cause premature deaths.

This assumption has been used by many to claim that air pollution in the U.S. causes thousands of premature deaths to occur yearly. Now, Gray does not show that (P) is true but he simply assumes it to be true. However, current research renders (P) implausible. For example, in his article Scientific Distortions in Fine Particulate Matter Epidemiology, Dr James E. Enstrom explains:

The theoretical prevention of premature deaths from the inhalation of fine particulate matter is being used by the U.S. Environmental Protection Agency (EPA) to justify the National Ambient Air Quality Standard (NAAQS) and multibillion dollar regulations across the U.S., including the EPA Clean Power Plan and the California Air Resources Board (CARB) Truck and Bus Regulation. The epidemiology is severely flawed. Fine particulates probably make no significant contribution to premature mortality in the U.S. The publication of null findings has been blocked or marginalized and studies claiming excess mortality need to be reassessed.

Similarly, in their research paper titled Air quality and acute deaths in California, 2000-2012, Young et. al. write,

Many studies have shown an association between air quality and acute deaths, and such associations are widely interpreted as causal. Several factors call causation and even association into question, for example multiple testing and multiple modeling, publication bias and confirmation bias. Many published studies are difficult or impossible to reproduce because of lack of access to confidential data sources. Here we make publically available a dataset containing daily air quality levels, PM2.5 and ozone, daily temperature levels, minimum and maximum and daily maximum relative humidity levels for the eight most populous California air basins, thirteen years, >2M deaths, over 37,000 exposure days. The data are analyzed using standard time series analysis, and a sensitivity analysis is computed varying model parameters, locations and years. Our analysis finds little evidence for association between air quality and acute deaths. These results are consistent with those for the widely cited NMMAPS dataset when the latter are restricted to California. The daily death variability was mostly explained by time of year or weather variables; Neither PM2.5 nor ozone added appreciably to the prediction of daily deaths. These results call into question the widespread belief that association between air quality and acute deaths is causal/near-universal.

Likewise, Steven Hayward says,

Air pollution affects far fewer people, far less often, and with far less severity than is commonly believed. ... The air pollution-mortality claim deserves even greater skepticism. First, is is based on the same unreliable correlation moderns that have led medical authorities astray in other areas. Second, even though pollution was correlated with higher premature mortality on average, it seemed to protect against death in about one-third of cities. How could pollution kill people in some cities and save them in others? More likely, both results are chance correlations rather than real effects. Third, researchers have been unable to kill animals in laboratory experiments, even when they expose them to air pollution at levels many times greater than ever occur in the United States. This suggests that air pollution at today’s record-low levels doesn’t pose a risk, and current standards are health-protective with plenty of room to spare.

Furthermore, Warren Kindzierski, associate professor in The School of Public Health at the University of Alberta, calculates that "any epidemiology study claiming air pollution causes a disease will have a less than 18 per cent probability being true at a prevalence rate of 0.1". He continues:

When it comes to sharing results of these studies with the public, epidemiologists and too many reporters ignore another truth. They make bold claims that air pollution may cause a disease, but these claims are unlikely true. They seldom offer results about air pollution not causing a disease, even though these results are very likely true. This is a form of reporting bias (only offering selective results or suppressing results) that people should be aware of. Really, the true headline of air pollution-epidemiology studies is… Claims that present-day air pollution may cause [enter disease name here] are mostly false!

For these reasons, Gray’s claim that the Highveld air pollution did, in fact, cause the deaths of around 600 people in 2016 is irresponsible and cannot be taken seriously.

But what about the less severe health danger of asthma? Can air pollution cause asthma? Once again, we simply do not have enough reliable evidence to think that air pollution necessarily causes asthma. In their 2002 article Asthma in exercising children exposed to ozone: a cohort study, McConnell et. al. found that, in general, there is no greater risk of developing asthma for children in high polluted areas than in low polluted areas. In fact, when considering all the children in the study, the children in the high polluted areas were about 30% less likely to develop asthma:

Risk of developing asthma was not greater overall in children living in the six high pollution communities than children living in the six low pollution communities, after adjustment for stratified baseline hazards for age and sex, and for ethnic origin, irrespective of which pollutant was used to classify communities as high or low. ... Communities with high NO2 and associated pollutants, and communities with high ozone or daily maximum ozone were associated with a decreased risk of asthma; these associations were significant ... only for daily maximum ozone (p. 388).

Consequently, claims that air pollution cause asthma are often exaggerated.

Of course, this does not mean that air pollution does not have negative side affects. It is plausible that air pollution might have some negative short-term health affects, such as negative respiratory symptoms or blocked nose, but these are scarce and usually not serious (Schwartz, 2003, p. 27). But the above facts do show that the dangers of air pollution are often exaggerated. What is the reason for this exaggeration in the media? Joel Schwartz offers a plausible answer:

Studies that report harm from air pollution are more likely to be published than studies that do not. Regulatory agencies, whose power and budgets depend on the perception that air pollution is a serious health problem, are also major funders of the research intended to demonstrate the severity of the problem. Scientists who believe air pollution is a serious health threat and who report larger health effects are more likely to attract research funding. It is not a big leap to conclude that there is a great deal of selection bias in who does environmental health research, what questions they ask, and how they report their results. Journalists should be acting as a check on air pollution misinformation, but they are not. Media outlets face their own pressures to sensationalize stories. Good news does not sell newspapers or attract viewers. As a result, journalists and editors are more likely to cover studies claiming harm from air pollution, and to pass along these claims with little or no critical review (Air Pollution and Health, 2006, p. 8).