Does anyone die from indoor air pollution?

35 thoughts on “Does anyone die from indoor air pollution?”

1. Another very dangerous “indoor air pollutant” that is not mentioned is toxic mold.

2. Bill: I agree.

3. Nikhil: You use the example as smoke being a risk factor to COPD, but if the actual deaths due to COPD = 0, then the risk factor is irrelevant to attributing deaths (which is what this original post is all about, the WHO attributing millions of actual deaths to something).

   I do not disagree with the notion of risk factors, but you can’t use an inflammatory claim like 4 mil deaths. The implication is that if this risk had been removed, those 4 mil would still be alive which is false.

4. The use of dried animal dung for fuel is ubiquitous and timeless. It is a GOOD thing.

5. Dino: Your question was about the UN. You seem to know the answer. I don’t need to opine.

6. To NIKHIL: WHO reported it, therefore they support the BS. My question is valid regardless of the researcher, reporter, or your opinion.

7. To Dino: This is not WHO. It is merely reporting academic research. Much of it sponsored by USEPA and used in its regulatory justifications, agreed to by the Supreme Court so far. That is the law; the rest is sniping.

   To Kuhmkat: I need only repeat – “Correlations matter.” I leave academics to argue causality; they are experts in wasting time.

   To Allen Brooks: As far as “cause of death” goes, yes. To repeat, premature mortality or DALY is not an individual attribute. I am shocked at meth reference. Will check out. Thank you.

8. Perhaps the leading cause of death from indoor air pollution is CO. Radon daughters can also be threat and they are usually closely monitored. But the most subtle and dangerous cause of indoor air pollution is meth. We don’t hear much about it because it is not politically correct to do so. It would expose the extent of the abuse. Only recently a company formed in our city to test and decontaminate meth contaminated housing (meth busters)

9. Nikhil, thanks for clarifying your useless first post. Now, what is the evidence that COPD or ALRI causes death and is caused or made worse by the particulate?!?!

10. Is there anything from the United Nations that is true?

11. Re: Kuhnkat

    The question is not whether smoke makes anybody sick. But whether exposure to smoke is a risk factor for ALRI or COPD.

    Correlations matter.

12. Nikhil:

    “Smoke is sickening. It doesn’t take even grade school to know that.”

    I have been sick many times in my life. I am not dead. Something causing me to be sick is not a strong indication of an actual death or even possibility of a death. Correlation is not causation. Write that a few million times.

13. I seem to recall
    – In Asia many women are dying in their 30’s from cooking over open wood fires indoors.
    – In the US some days may be knocked off the end of some people’s live due to traffic particulate pollution.

    Any news report that uses raw deaths is junk, cos it is Quality Adjusted Life Days that should be counted.
    …cos death comes to 100% of humans (so far)

14. Re: Andrew Kerber

    Yes, that is correct. But it applies to actual disease and deaths. Premature mortality is about apportioning the diseases to risk factors, which may overlap, as is the case with CO and PM.

15. It would really be interesting to know how many of those deaths of children in the room where cooking occurred were due to CO poisoning, rather than particulates. Clearly in places where ventilation is poor, CO poisoning could be a problem.

16. Bill wrote: “life expectancy is rising which implies that there would have been MORE premature deaths in the past (sic) when levels were lower. So is 3.8 million deaths (assuming the number is anywhere remotely accurate) a bad thing or is a valid argument that all this PM2.5 actually reduces premature death (and explains the rise in life expectancy) >:)”

    You got this partly correct. Yes, life expectancy is rising, for whatever reason. Premature mortality is increasing because the number of deaths are increasing, including before the new life expectancy, because even if later, people have to die.

    When there is more premature mortality, and the prevalence of other risk factors or diseases declining, of course more premature mortality is attributed to solid fuel use. (Yes, something has to be blamed.) Read the Kirk Smith paper on how the numbers went from below 1 million to nearly 4 million.

    Bill also, “The LE/cause/risk factor methodology, for example, may have predicted things fairly well in the 1950s/60s/70s before mass migrations of populations and globalization became the norm but does that same method reliable now? ”

    Premature mortality is NOT a matter of prediction, only history, and only for parceling out numbers. The question is, are dose-response equations of the past applicable today as air has become much cleaner, people on average have a different nutrition and health profiles?

17. Bill wrote: “life expectancy is rising which implies that there would have been MORE premature deaths in the past (sic) when levels were lower. So is 3.8 million deaths (assuming the number is anywhere remotely accurate) a bad thing or is a valid argument that all this PM2.5 actually reduces premature death (and explains the rise in life expectancy) >:)”

    You got this partly correct. Yes, life expectancy is rising, for whatever reason. Premature mortality is increasing because the number of deaths are increasing, including before the new life expectancy, because even if later, people have to die.

    When there is more premature mortality, and the prevalence of other risk factors or diseases declining, of course more premature mortality is attributed to solid fuel use. (Yes, something has to be blamed.) Read the Kirk Smith paper on how the numbers went from below 1 million to nearly 4 million.

    Bill also, “The LE/cause/risk factor methodology, for example, may have predicted things fairly well in the 1950s/60s/70s before mass migrations of populations and globalization became the norm but does that same method reliable now? ”

    Premature mortality is NOT a matter of prediction, only history, and only for parceling out numbers.

18. @Allen Brooks, perhaps this is the reason why the Government is now focusing so much on PM2.5. With the drives to legalize marijuana this provides an additional avenue for them to regulate and provide kickbacks to winners? Perhaps there is a company already developing a method of measuring PM2.5 like a calorie count and it’s a way to make money through labeling requirements. /sarcasm off. Of course, it will likely just be a completely made up number like the fuel economy ratings on the new car window sticker.

19. Well, the other thing you have to consider with the statistical work is that the don’t know what side they are on. 3.8 million premature deaths, but life expectancy is rising which implies that there would have been MORE premature deaths in the past when levels were lower. So is 3.8 million deaths (assuming the number is anywhere remotely accurate) a bad thing or is a valid argument that all this PM2.5 actually reduces premature death (and explains the rise in life expectancy) >:)

    FYI, for perspective, i’m a computer scientist who works in large data set analytics and business intelligence…now “sold” as “Big Data” by the vendors selling tools and trying to reinvent the field because data warehousing got a bad name as a money sink to the same vendors. The naive often overlook questioning the validity of a dataset simply because a prior set of results fit the real worldand therefore any further questioning is avoided — ignoring that the dataset is evolving and not static. The LE/cause/risk factor methodology, for example, may have predicted things fairly well in the 1950s/60s/70s before mass migrations of populations and globalization became the norm but does that same method reliable now? We have the same discussions regarding economic policies but focusing on medicine: the WHO puts out “grades” of country medical systems and included in there is infant mortality which is one of the elements the policy makers pull out to lambast how bad the US medical system is. However, one could argue that infant mortality in the US is artificially high and in other countries artificially low simply due to migration. Just recently we had a South American woman with Zika who’s ultrasound showed that her fetus was defective. She traveled to the US where the baby was delivered and now being cared for in New Jersey. Should this child not survive they become part of the US statistic and not the statistic of their native country. The statistic used to demonstrate how poor the US system is actually demonstrates the inverse once the dataset is evaluated and not simply analyzed using statistical and meta analysis.

    Hopefully this illustrates why I am very skeptical of any calculated “deaths.”

20. I wrote “EPA assumes too many things in its computations.”

    But who am I to judge? EPA says scientific assessment of literature and review of the assessment by its advisory body underlie its assumptions. I haven’t checked. If the Supreme Court grants the Chevron deference, it doesn’t matter what others say. Or think.

21. Allen Brooks: That precisely is the point. There are NO autopsy reports listing “indoor air pollution” as the cause of death.

    Actual deaths have causes. The aggregate of premature mortality is computed and statistically allocated to “risk factors”.

    In a recent Lancet paper, “risk factors” for premature mortality among the youth (10-24) included unsafe water, alcohol use, drug use, handwashing, unsafe sex, high blood pressure and yes, household air pollution.

    These are “risk factors” that contribute to incidence of diseases that lead to premature deaths. Conversely, premature mortality is not caused by these risk factors but can be appoprtioned among these risk factors or can be said to be attributable to these risk factors.

    The methods, data, and interpretations require many assumptions. In any case, premature mortality attributable to household air pollution varies over time, not the least because incidence of other diseases (than those where household air pollution is a risk factor) may change, their treatment improves, or that nutrition and exercise profile changes.

    EPA assumes too many things in its computations.

22. nikhil: If “it has nothing to do with specific deaths” how can it be linked to “indoor air pollution? Where are the autopsy reports listing “indoor air pollution” as the cause of death? And yes tobacco smoke is sickening, but the same crowd that shouts this has no problem with pot, which is far worse.

23. Paper on China at http://www.pnas.org/content/early/2016/06/21/1604537113.
    A press release at http://phys.org/news/2016-06-household-fuels-major-contributor-beijing.html.

    My commentary:

    Kirk Smith and colleagues: “We show that due to uncontrolled and inefficient combustion of solid fuels in household devices, emission reductions from the residential sector may have greater air quality benefits in the North China Plain, including Beijing, than reductions from other sectors.”

    “Reduction of emissions from the residential sector via the replacement of solid fuels with other cleaner energy sources could substantially improve air quality in the BTH region of eastern China.”

    Amen. Substituting grid power (high share of coal) and electric cooking (rice cooker to kettle to stovetop and ovens) and heating for the next 25 years (while gas and renewables come up) would be good for the children of Mother Earth, even if not for the Mother. (And which mother would mind that?)

    Oh, wait a minute. Aerosols from coal power largely negate the radiative forcing due to their CO2. So coal power is cleaner both for health as well as climate.

    All emissions and all impacts considered – and discounted for time, uncertainty, and expert biases, modern energy transition for the poor is beneficial to them, their governments, and the planet.

    But some children of mother earth don’t care about their siblings.

24. This is in response to Bill who wrote about causes of death as reported on death certificates (DC).

    I wish the cooking up of risk factor allocations of premature mortality were that simple.

    Premature mortality is a population metric, not individual metric. Bill’s father’s premature death was from multiple “causes” but “causes” at individual level do not come into play. For every “population”, there are reported “causes”, as in the case of Bill’s father. But not all deaths are premature (roughly only a half), and life expectancies and premature mortality are specific only for a particular year.

    Say, the total male deaths in US in 2000 were 1.2 million, and life expectancy for males was 75 for that cohort (changes very slowly). Say, only 500,000 were before 75.

    Then there is a distribution from reported causes of death. (These have weaknesses, more so in poorer areas of the world. Lancet had an excellent, searing piece a year or two ago.) All the causes are aggregated, then distributed to “risk factors” – say, 100,000 due to CHF are given x% to obesity, y% to smoking, z% to something, etc. This allocation uses statistical methods and heroic assumptions in part based on specific epidemiological studies but very loosely; studies are not that consistent and the populations are not comparable. In any case, deaths any particular year have an unknown distribution of “confounding factors” – nutritional history, genetics – that is different from deaths for any other populations. Mislabeling of causes of individual deaths makes little difference.

    Why is this exercise done every few years, considering all the qualifiers that must be attached? To get an idea of “disease control priorities”.

    Thus, for instance, a recent paper in PNAS (I will post later) argued that in North China, air quality improvements from interventions in the residential sector are more promising than in other sectors. I think it is an important statement, even with qualifiers for exact numbers. It says, as VicB3 recommends above, that a coal power plant can help clean up the air from residential uncontrolled use of unprocessed solid fuels.

    To Milloy re: his opening note – Armstrong (1991) had ALRI as a more prominent cause because i) it is a cause, not a risk factor, and ii) it is for a year with a different distribution of deaths and life expectancies (and also when air pollution may ave been lower.) Read a paper on Kirk R. Smith website about how premature mortality allocation to Household Air Pollution shot up from 1 million to 4 million. It will also help understand the method behind the madness.

25. Sorry, I don’t mean this to come off sarcastic but I find it funny people think that the entries on the death certificate are accurate. My father had a heart attack and died. The DC said the cause was CHF; however, he had a previously diagnosed heart defect (diagnosed 20+ years prior). Now, he had fluid on his lungs, so everyone assumed CHF but the autopsy said the fluid was not consistent with CHF but instead consistent with infection. So now, we have a man who had a heart defect who had a lung infection which resulted in a heart attack. Death cert is heart attack, my doctors are all now trying to prescribe mega doses of every cardiac drug known to man because of my father’s death when the reality is he caught some sort of lung infection, went for 3+ months before getting it checked out. That 3 month long “walking pneumonia” then causes his heart to work harder causing it to swell (same as CHF) and when the defect was interjected — death.

    My point? What was the cause of his premature death? Was it even premature since he had a genetic heart defect? I’m sure he is included in the PM2.5 deaths somewhere simply because he had a heart attack so it must have been caused by PM2.5, not possibly an infection where the patient didn’t seek treatment until the damage had been done. The lung infection be damned (there was even early speculation of lung cancer but he passed before those tests had been conducted — his physician was shotgunning ideas of what was causing the fluid since it wasn’t responding to antibiotics or digitalis).

    Analysis conducted from death certificate causes is flawed IMO.

26. Allen Brooks says, “If there is any form of “indoor air pollution” and someone dies, (regardless of the cause, even automobile accidents or murder) they will use the numbers to prove their case.”

    This is simply not true. Premature mortality is a population concept, and isk factors a matter of allocating these deaths. It has nothing to do with specific deaths.

    Smoke is sickening. It doesn’t take even grade school to know that.

27. 1 – Indoor pollution isn’t the problem (if indeed it is), it’s the lack of ventilation. Tell somebody to open the damned window.

    2 – Build a few power plants and install some natural gas lines, add some modern appliances, and the problem will go away. In other words, let loose the market driven economy and the problem will solve itself quickly as people acquire wealth. Except for the odd outdoor BBQ, camping trip or romantic evening, nobody willingly burns wood or lights candles for cooking, lighting, heating or whatever if they don’t have to.

    Just a thought.

    VicB3

28. If there is any form of “indoor air pollution” and someone dies, (regardless of the cause, even automobile accidents or murder) they will use the numbers to prove their case.

29. If indoor air pollution was deadly, humanity would have been practically made extinct due to smoke from open fires in caves.

30. Stan B gets it right – “We have to blame something.” Grants are needed and authorized, but the whole question of “risk factors” is, what is important and ow much, so we can prioritize various options for better health?

    Science helps only in setting priorities. But priorities cannot be set in isolation. The air pollution community obsesses over PM2.5 and the unknown dose-response relationships to allocate the burden of disease, while ignoring the most dangerous air pollutants among the PM2.5! Air quality monitors do not measure all PM2.5 at all levels. Walk around any poor area in Asia or Africa and you’d see fires, chemical spills, open waste, and noxious gases, close to the ground level. This pollution is not measured.

    Spurious data, unsupported assumptions, unvalidated models. An open marriage of regulators, advocacy groups, and academia.

31. Has anybody but me noticed that longevity has become a commercially-motivated “competition?” Apparently, he who lives the longest, sucking up the most disease-care dollars, “wins.” Otherwise, why the ongoing obsession with the “risk factors” of everyone else’s “health?” Most of American “primary care” no longer is treating actual DISEASE–it is treating statistical constructs of dubious merit at enormous expense and misery to individuals. Why play the game? Anything coming out of the WHO is already suspect given their ridiculously unattainable definition of “wellness.”

32. We have to blame something, or the grant money dries up!

33. Exposure to air pollution in general – and to air pollution from solid fuel use in household cooking and heating – is a “risk factor” in the diseases – COPD and other – that are in turn implicated in actual deaths. Risk factors do not cause a disease or a death. Premature deaths are not individual deaths. They are a construct – deaths before a life expectancy appropriate to a particular population (in itself a debatable construct *), which changes with every 10-year cohort – and the total of those premature deaths are allocated, via the use of debatable methods and assumptions, to different “risk factors”. (Nobody bothers about the post-mature mortality, which has almost as many people though fewer years; arguably, the “risk factors” for premature mortality should be netted out against their contribution to post-mature mortality – e.g., by the number of years cigarette smokers die BEYOND life expectancy.)

    The question is not whether 4 million premature deaths are allocated to household solid fuel pollutants but rather, if not, which risk factors should they be ascribed to?

    WHO statements that you cite should be corrected as “premature deaths attributed to”, leaving aside the issue of whether the methods of attribution are valid. The assumption is, if experts agree that methods are valid, then the numbers are valid for experts; others don’t have to agree with experts. (Group consensus has its advantages and disadvantages.)

    So,

    Over 4 million premature deaths are from illness attributed to.. These include “More than 50% of premature deaths due to pneumonia among children under 5 ..” and “3.8 million premature deaths annually from noncommunicable diseases”

    That is, from the total number of premature deaths from non-communicable diseases such as “stroke, ischaemic heart disease, chronic obstructive pulmonary disease (COPD) and lung cancer”, 3.8 million “are attributed to exposure to household air pollution.”

    I am not sure an individual actually dies from just one of these particular diseases, though only one cause may be listed for cause. (I suspect more than a half of the deaths now still do not have a validated cause of death. Not many deaths occur in hospitals and death certificates are iffy.) Each death may have multiple illnesses and each illness may have multiple risk factors.

    Attribution is an allocation game. No claim is made for biological causality. Allocation in turn comes from

    * I am an Indian immigrant to the US. Which life expectancy should my death be compared to? Indian males of 1950s or 2030s? American males of those 2010s?

34. This reminds me of the argument over the health effects of marijuana.

    The pro-pot crowd claims there are no deaths linked to pot use, yet there is no easy way for those filling out death certificates to say that pot use was, or was not a factor without performing tests that cost money and may not be conclusive.

    But no one will raise an eyebrow if they check the box for “tobacco”.

    If enough people “in the system” accept a premise, it is not easy to refute it.

35. Let’s pick a number and see if it sticks.