UK air pollution causes 50,000 early deaths a year, say MPs

Minute particles from burning fuel can shorten lives by up to nine years, according to the environment audit committee

  • Article history
Air pollution : The Houses of Parliament are shrouded in early morning mist
In pollution hotspots such as London, air particles could be cutting vulnerable people's lives short by nine years, says the committee. Photograph: Oli Scarff/Getty Images

Fifty-thousand people a year may be dying prematurely because of air pollution, a influential committee of MPs has reported after a six-month investigation.

According to the environment audit committee, minute sooty particles, emitted largely from the burning of diesel and other fuels and inhaled deeply into the lungs, shortens lives by seven to eight months. In pollution hotspots like areas of central London and other cities, the particles could be cutting vulnerable people's lives short by as much as nine years.

The committee, which took evidence from government ministers as well as medical experts, said it was shocked so little was being done to address the problem, despite the health evidence and threats from Europe to take Britain to court and fine it millions of pounds a year until improvements are made.

In a series of Commons meetings, the committee heard the government had been breaking EU air quality laws for more than a decade and also extracted new figures from the Department for Environment Food and Rural Affairs. These showed that the scale and seriousness of the long-term problem had been known for many years but had been repressed.

Long-term air pollution, from the sooty particles known as PM10s, nitrogen dioxide (NO2) and nitrogen oxides (NOx), it was told, makes asthma worse and exacerbates heart disease and respiratory illness.

"Despite these considerable impacts on public health, very little effort is being put into reducing air pollution levels, compared with efforts to tackle smoking, alcohol misuse and obesity. Much more needs to be done to save lives and reduce the enormous burden air pollution is placing on the NHS," said the committee chair, Tim Yeo.

"The large EU fines we face, if we don't get to grips with this problem, should now focus ministers' minds", he said.

"Air pollution from road vehicles causes the most damage to health. A dramatic shift in transport policy is required if air quality is to be improved," concluded the report. "This means removing the most polluting vehicles from the road, cleaning up the vehicles that remain and encouraging smarter choices about transport. Many of the policies needed to reduce transport emissions have the added benefits of tackling climate change by reducing carbon dioxide emissions."

The air pollution problem is particularly acute in London where there has been a political standoff between the mayor, Boris Johnson, and central government, with each blaming the other for inaction.

Simon Birkett, director of Clean Air for London (Cal), a small group which has led efforts to expose the full extent of air pollution and the official reluctance to address the problem, said the inquiry confirmed Cal's estimates of premature deaths.

"This report shames Britain. The government should respond immediately by giving Mayor Johnson full responsibility for complying with limit values for dangerous airborne particles (PM10s) in London; publishing its plans for complying with legal standards for nitrogen dioxide (NO2) and oxides of nitrogen (NOx); and communicating clearly estimates for the number of premature deaths due to poor air quality," said Birkett.

"Why did it take an inquiry by one of parliament's most powerful select committees to get a government minister to refer, for the very first time, to the possibility of 35,000 premature deaths in the UK in a year due to air pollution?" he asked. "Why has the government never published an estimate for the number of premature deaths due to long-term exposure to dangerous airborne particles? Before the inquiry, the highest government number we had heard before was "up to 24,000 deaths per year"."

Environment groups said the report showed the folly of trying to expand Heathrow airport with a third runway. Geraldine Nicholson, from the No third runway action group, said: "The air quality limits have, in recent years, been consistently exceeded in the area around Heathrow airport and will not be met in 2010 or the foreseeable future."

"There are still no measures in place to help improve our air quality as it is today. The addition of a third runway and sixth terminal at Heathrow are proof that the UK government is, once again, swimming against the tide and is putting economic and industrial interests over the health and wellbeing of local residents."

Local authority chiefs also called for action. "The risk to people's health and the threat of fines from the EU leave no excuse for inaction and we need a co-ordinated national approach to policy and improving public awareness," said the LACORS chairman, councillor Paul Bettison.

"The full impacts on Londoners' health need to be spelled out clearly by the mayor and the government in their official documents. It appears that official statements made in previous years have completely underestimated the serious nature of the impacts on human health and the mayor needs to come clean," said Darren Johnson, Green party London assembly member.

A Defra spokesperson said: "We take improving air quality and meeting EU targets very seriously and have already made significant achievements – since 1990 we have succeeded in reducing sulphur dioxide emissions by 86% and have nearly halved particulates."

He added: "We are working across government to reduce emissions further, including transport and electricity generation. But we recognise that there is more to do and will consider the EAC report carefully."

A spokesperson for the Mayor Johnson said: "We welcome this influential Committee's findings and its recommendations [but] London cannot act alone in solving this problem. Additional funding and action is required from government ... if EU targets are to be met in the capital. The Mayor will shortly be publishing an air quality strategy for public consultation."

Your IP address will be logged


39 comments, displaying oldest first

  • This symbol indicates that that person is The Guardian's staffStaff
  • This symbol indicates that that person is a contributorContributor
  • Silvanus

    22 March 2010 11:36AM

    Please note the Metropolitan context of the seriousness of the problem. Meaning youse Bummies, Scouses, Geordies,et al, will have to reconcile yourselves to dying in increasingly large numbers until we get handle on the problem at the London end.

  • Owenbevt

    22 March 2010 11:52AM

    I?ve just moved to London and noticed my tissues are always black now which makes me worried for my lungs.
    Also having visited Copenhagen recently it feels completely insane that anyone would try to put large roads through major public spaces like the Piccadilly or Parliament Sq.s

  • ColinG

    22 March 2010 12:02PM

    Despite these considerable impacts on public health, very little effort is being put into reducing air pollution levels, compared with efforts to tackle smoking, alcohol misuse and obesity.

    Public ambivalence towards fossil fuel pollution never ceases to amaze me.

    People will happily protest against wind farms and nuclear powerstations, but these are utterly benign compared to the public health impacts of burning fossil fuel (or, for that matter, biomass/fuel).

    The sooner we adopt electric transport the better.

  • leadballoon

    22 March 2010 12:07PM

    A pretty strong argument for tidal/wave/wind/nuclear power and electric vehicles untainted by the fetish of CO2 counting. All 'green' strategies are not equal if diesel buses and taxis are a major source of pm10 pollution in the city.

    I'm a little suspicious of policy by 'limits' since they are inevitably guided by, on the one hand, pollution levels that can reliably measured and, on the other, shifts in life expectancy that can be distinguished from background. Reducing 'hotspots' by shifting pollution to a wider area where levels fall just under limits could be redistributing, say, one premature death by a year to a month for a dozen others. The strategy could be just making the effect harder to detect; better to reduce total pollution emission.

    The quotes from 'no third runway action group' are rather missing the point. Most of the pollution at ground level around an airport is from the vehicles serving it. Refusing to expand existing airports just shifts the traffic to someone else's backyard. Passenger bus and cargo lorry adds to and spreads the pollution between airports via the road links. Better to keep it in one place and why Heathrow in particular should be developed into a major transport interchange with High Speed Rail links in addition to another runway.

  • theonlynameleft

    22 March 2010 12:11PM

    At least we don't live in L.A or Mexico city. Their climate, topography and landscape makes it much worse.

    And at least i don't live in London.

    I agree with ColinG on the electric transport thing, but the electricity has to come from somewhere.

  • BwanaNdegi

    22 March 2010 12:40PM

    A simple filter on diesel engines (as used in other countries) to remove carbon particles would go a long way to help - or would that be too much effort for our esteemed leaders to legislate in?

  • Drcrinum

    22 March 2010 12:55PM


    I intended to post reference to the same study after reading the above article when I noted you already had done so. Here is another link:

    The major problem with medical and public health studies involving statistics is lack of statistical expertise -- actually it applies to most fields of science where statistics is employed. The flaw is usually lack of understanding of Baseyian Methodology. Here is a good discussion.,_Its_Wrong

    Any science studies where statistical analysis is employed should be mandated to have an independent review by a statistics expert as part of the peer review process.

  • Bovary

    22 March 2010 1:31PM

    Our esteemed leaders are only happy that there are fewer of us to deceive and worry about. Cynical, but true.

  • bojimbo26

    22 March 2010 2:08PM

    I thought it was supposed to be Global Warming .

  • RacktipOddling

    22 March 2010 2:23PM

    Those class of persons now programmed to deny any atmospheric science, except checking the morning sky to see if an umbrella is necessary, please now join in the chorus of "I'm Not Dead Yet, So It's Not Really Happening".

  • nevado

    22 March 2010 2:37PM

    The committee, which took evidence from government ministers as well as medical experts, said it was shocked so little was being done to address the problem, despite the health evidence and threats from Europe to take Britain to court and fine it millions of pounds a year until improvements are made.

    What do they expect? The government want to avoid spending money at all costs, and they dont really give a shit about the general health of the populace. If they did they would tax harmful stuff like tobacco out of existence. They are happier raking in the taxes on harmful and polluting substances, which they keep at (just about) affordable levels, from which they can easily pay any European fines. They would rather watch people drop dead than spend money fixing things.

    Bunch of moronic useless incompetents.

  • ColinG

    22 March 2010 2:42PM


    That is a surprising result. The COMEAP report published last year concluded that the evidence base regarding the effects of long-term exposure to air pollutants on health has strengthened. Mortality from PM2.5 particulate pollution is approximately 6 times higher than previously thought in 2001.

  • Plutonium

    22 March 2010 3:03PM

    At a Los Angeles conference on sub-micron particulates a few years ago, it was pointed out that the greatest sub-micron exposure is while driving on the freeway. It would seem that someone that owned more cars over their life would have driven more on the freeway. More exposure as a result. Death certificates could be searched for age at death and cause of death. Vehicle ownership records could be searched to see how many cars each person owned in their life. Also check to see if the last car owned was salvaged, meaning the person died in a crash. All this data is probably available for multiple regions.
    See what kinds of numbers fall out before spending huge money. Lead was removed from gasoline 1975 mainly because it poisoned catalytic converters, not because of its health effects on humans. MTBE was removed from gasoline because of supposed ground water contamination. Ethanol that replaced MTBE tends to make hydrocarbons emissions worse.
    There is also the additional issue that smoke particulates sequester radon daughters. By some modeling, second hand smoke reduces radiation exposure in lungs by 20%, relative to very clean air. Publications on this issue are somewhat obscure. Homes in cold areas with basements have higheer indoor radon concentration. If radiation is bad for you, second hand smoke may be good for you. Politically, you can not have it both ways. This does not distract from the issue that smoke stinks and is ugly. It is not easy being "green."

  • jennybates

    22 March 2010 4:09PM

    This report shows how crucial it is that the Mayor sets out adequate policies to bring the whole of London within EU legal limits on air pollution.

    Instead of relying on unspecified and unfunded measures to cut air pollution, the Mayor must do more to encourage people out of their cars by making London a safer place to walk and cycle and improving public transport.

    By removing the Western Extension to the Congestion Charge, dropping traffic reduction targets, proposing large new road crossings in East London and supporting a 50 per cent increase in flights from City airport, Boris Johnson will add to traffic and air pollution problems and increase carbon emissions.

    Friends of the Earth's London Campaigner, Jenny Bates

  • ianfalconer

    22 March 2010 5:22PM

    ascammer & drcrinium
    and for indoors specifically

    Now reading the cited paper by Koop et al it looks all very good and professional, except for the line at the bottom of page 10
    "Observations [of air quality indicators] are monthly means of daily means"

    Essentially they have averaged away any correlation between air quality indicator variation and morbidity before they have even done a regression.
    How can you possibly tell whether a 2 day TSP or ozone spike correlates to excess hospital admissions if your data is working on a monthly average ?

  • Leslie2

    22 March 2010 6:30PM

    about "a government minister to refer, for the very first time, to the possibility of 35,000 premature deaths in the UK in a year due to air pollution...."

    No worries they plan to lower the alcohol driving limit to save 65 people a year.
    Saving 35,000 people is too big a problem for the Government to consider.....

  • Drcrinum

    22 March 2010 7:27PM


    They address your question in the article. It is standard practice in that particular type of epidemiology to use monthly averages. The diseases studies were not toxic exposures leading to immediate reactions.

    The respiratory diseases listed in the ICD codes in the study are both acute and chronic diseases. Combining these with pollution levels doesn't usually lead to a one for one approach; i.e., exposure to SO2 today may exacerbate chronic bronchitis and eventually led to hospitalization 3-4 days later. In other words, there are usually delays between insult and hospitalization. Also, pollution levels tend to rise and fall over extended periods of time, days to weeks, and it may be an accumulation of the effects of the insult over a period of time that leads to the hospitalization. At least we have been led to believe this was the way the case .... now it looks like it is time to re-evaluate the effects of air pollution.

  • ianfalconer

    22 March 2010 8:58PM


    I'm afraid that your answer seems to agree that air pollution has a negative effect on health, which is not what your original post implied since the paper that you approved of says exactly the opposite.

    I apologise if I have misunderstood your comments and agree that a re-evaluation of the role of airborne pollutants, as a motivator for positive changes in air quality, is required as it has been swamped by other issues in recent years.

  • kennymac825

    22 March 2010 9:16PM

    The WHO estimates 15 million people worldwide die from starvation every year. 50 thousand premature deaths per year due to pollution?. That is a drop in the bucket. Precious money should be directed to where it will do the most good. Preventing death due to starvation. Not air pollution in Britain.

  • Guffet

    22 March 2010 9:25PM

    I have recently research the issue of waste incineration pollution and I am alarmed. The emission of dioxins and nano particles will be deadly. Not only has this government been concealing the premature deaths of 50,000 per year, failing to meet European limits by doing little to address the problem, but are now encouraging the network of vaste waste incinerators, which will kill more people, not to mention pump our CO2 in huge quantities 24/7 for the next 25 years. Of course, this Government is not exceeding the European limits on nano particles, because there are no European limits in existence. God help our children and grandchildren unless someone in Government gets a grip.

  • Jacksavage

    22 March 2010 10:10PM

    A report to bounce us into the expense of compliance to our EU masters.

    Would the committee chair be the same Tim Yeo mentioned in Wikipedia ?

    "Tim Yeo is chairman of Univent plc and non-executive chairman of Eco City Vehicles plc and AFC Energy plc."

    AFC Energy plc is all about hydrogen fuel cells. I wonder what Eco City Vehicles plc is all about. Univent plc seems to be about retirement homes. Is Tim worried we will all not live long enough to patronise one of his establishments?

    Air pollution is an ugly thing,but I believe the overall direction is already one of improvement,even taking Defra's figures with a pinch of salt.

    Doubtless Clean Air for London is a fakecharity funded by the EU or WHO or both....

    Let's have a few more know you want them.

    If it were true that all my vile habits and unhealthy lifestyle have shortened my life as much as scientific reports said they should, I would have been dead at 30.

    It is all so depressing....

  • Drcrinum

    23 March 2010 3:10AM


    Sorry if my answer was confusing.

    Medical science has taught for years that air pollution was bad for your health; i.e., it caused certain respiratory illnesses and exacerbated certain chronic respiratory diseases. Various epidemiology studies supported this concept. Public Health officials have touted these concepts for years, which is the message in this feature article by the Guardian.

    The study by Koop, McKitrick & Tole however found no association between air pollution and the respiratory diseases tabulated in their study. This recent study contradicts volumes of previous studies between the association of air pollution and respiratory diseases. Koop et al used Baseyian Methodology in their statistical analysis, and the latter accounts for the astounding conclusions in their study.

    I have a close colleague who is a statistical expert -- he regularly employs Baseyian statistics in his research. He has been telling me for years that most medical studies reported in the literature that involve statistical analyses are flawed -- because the statistical methods used are erroneous. If you read the article by Siegfried in Science News that I referenced above, he says the same thing: most medical studies which used statistical analyses are erroneous because Baseyian Methodologies were not used.

    So the study by Koop et al is most probably accurate; they specifically attribute their conclusions to using Baseyian Methodology. Hence, this original article in the Guardian is bogus since the studies forming the basis for the harmful effects of air pollution were not analyzed using Baseyian Methodology.

  • ViewFromAlberta

    23 March 2010 4:27AM

    Climate Action Now!

  • mannin

    23 March 2010 6:46AM

    I'm afraid I no longer believe Guardian summaries of environmental reports. Can we have a link to the data ?

  • Monkeybiz

    23 March 2010 7:14AM


    Air Pollution, Economic Activity and Respiratory Illness: Evidence from Canadian Cities, 1974-1994

    Guelph economist Ross McKitrick, along with Gary Koop of the University of Strathclyde in Glasgow and Lise Tole of the University of Edinburgh, analyzed a new database from 11 Canadian cities over a 20-year period. Unlike most earlier studies, this one included controls for effects of smoking and income.

    They found no evidence that air pollution affected either hospital admission rates or time spent in hospitals. However, they did determine that both smoking and income levels directly affect respiratory health. Their findings appear this week in the journal Environmental Modelling and Software

    More proof, if you needed it, that McKitrick, by way of his associations with The Frazer Institute, is continuing to spread corporate propaganda that fossil fuels=good; science to protect the biosphere=bad.

    BTW the cited paper is old, not published "this week" as the post implies, and in a journal so obscure that a leading air pollution expert has never heard of it, and on examining the paper found it to be a very weak analysis with lots of flaws. Most robust literature shows quite profound effects on health from PM10 and PM2.5s. Interested readers are referred to the work of Hedley and Wong, and Anderson for better non-partisan science.

  • Monkeybiz

    23 March 2010 7:30AM

    @DrCrinium -

    This recent study contradicts volumes of previous studies between the association of air pollution and respiratory diseases. Koop et al used Baseyian Methodology in their statistical analysis, and the latter accounts for the astounding conclusions in their study.

    Ah, Nobel prizes all round. One study by McKitrick contradicts hundreds of others because it used Bayesian methods. Do you think it is the only one to do so?

    "I have a friend who says..." arguments are not persuasive DrC.

    I have many statistically highly knowledgable colleagues who publish on air pollution and they would not be accepted in the leading journals they are if their methods were unsound.

    Cue argument 2: "The problem with the peer review process..." and we have the same set of unfounded assertions about air pollution studies as we saw for the past few months about climate science. Come on!

    As a botanist you should know that there are many ways to analyse data. The conviction that air pollution is harmful comes from a variety of studies performed in numerous sites, using different analytic methods, and the vast majority of which, like climate science coincidentally, show harmful effects associated with exposures to PM10s and PM2.5s specifically endothelial dysfunction and cardiac disease as well as lung disease.

    exposure to SO2 today may exacerbate chronic bronchitis and eventually led to hospitalization 3-4 days later. In other words, there are usually delays between insult and hospitalization.

    Yes, and that is why it is important to allow for this effect in the time series. It is also important to allow for harvesting effects, where an acute pulse of pollution is associated with an acute increase in deaths a day or two later, but then a lowering of deaths over the subsequent week as the susceptibles have been removed by the acute pulse, until more susceptibles accumulate. Di Koop consider this. Nope. Just one of the errors that happens to creep in when economists start trying to disprove science they do not fully understand. As with the climate science papers, so with the air pollution paper.

  • roverdc

    23 March 2010 7:46AM

    So why do we not start with the most practical vehicles for running on electricity? The old trolley busses were quiet pollution free at point of use and the one weakness that the booms fell off can now be eliminated because self seeking pantograph pickups using something no more complex than a washing machine control circuit board is, unlike electric car dreamland impractical, but twenty years ago technology.
    Better still a convertion to a hybrid trolley bus and battery electric where the charging is done just at bus stops would be still a very short development project compared to developing an electric car that would be bought by any but a real Eco freak with loads of money.
    However I think we have also heard MP's say they have cleaned up after the expenses scandal.

  • roverdc

    23 March 2010 7:52AM

    22 Mar 2010, 12:55PM
    Any science studies where statistical analysis is employed should be mandated to have an independent review by a statistics expert as part of the peer review process.

    I fed the climate data through a market research program for trend analysis and this certainly bears out your assertion. The trend analysis it produced was the opposite to the one claimed by the climate scientists. That is even using their loaded data.

  • vanderlay1

    23 March 2010 8:55AM

    Far from being repressed, all the data and information pointing towards these conclusions hae been available for years on the Defra website. Sceptics please go and read to your hearts' content.

    It is disingenuous for Defra to claim success on sulphur dioxide and halving particulate levels. These changes have been the result of EU policy (Euro-sceptics please note).

    The failure we see today is the result of two-fold complacency. 1) reliance on European policies which are insufficient to manage pollution levels in our own urban areas and 2) institutional resistance to change policy in government transport and development planning departments.

    By the way, all of the solutions cited by previous posters have been previously considered and rejected on cost grounds. Should the public purse fund the replacement of polluting vehicles? If not then millions of families just got another big bill.

  • thesnufkin

    23 March 2010 10:00AM

    I think this is a very big issue that green groups really shoulg give a higher priority too.

    Climate Change may be the biggest threat we face, but the timescale of the problme means the best we can get from ministers is commitments to action years or decades in the future that may well turn out to be worthless.

    Air pollution though is a current problem and actions should be taken now. And of course by reducing emmissions from vehicles we would also be helping deal with Climate Change.

  • ianfalconer

    23 March 2010 10:11AM


    OK, so you mis-wrote and are sticking to your Koop et al story.

    As someone who has carried out quantitative air quality research (I used a piece of equipment called a QEM-SCAN to identify and characterise particulates in the range PM1-PM100) I can tell you, from first hand experience, that averaging urban air quality time-series data over a month is most definitely NOT standard practice.

    Take a look at the UK's air quality archive and you will see that some data is on a 15min resolution, with data series going up to a 24 hour mean dependent on the pollutant. Particulates are actually quite difficult to collect and analyse automatically, hence the longer sample time than the gaseous pollutants.

    Your obsession with Baysian methodologies is endearing, but you should really read the first page of every statistics text book that says 'garbage in = garbage out, irrespective of the methodology'. Removing resolution from time-series data reduces, or in this case removes, the validity of any time-dependent findings based on analysis of that data. Koop et al have done the wrong thing with their data before they even put it into their Baysian analysis. They have taken (what I assume to be) perfectly good data and screwed it up before throwing into their Baysian dustbin. It is not surprising that they found no correlations between events that are shorter than a month and their effects.

    BTW I agree with MonkeyBiz that omission of ischemic heart failure in Koop's paper is also a major deficiency.

    chronic exposure to particulates is a major cause of premature death across the developing world as well as affluent urban areas, mainly for women and children, who spend more time indoors next to the cooking fire/stove. The WHO estimate that 2 million premature deaths a year are caused by air quality issues, so its not an insubstantial problem. Not on the scale of poverty, but integral and one of the directly attributable causes of death due to poverty. We do need to understand and address it.

  • brian1w

    23 March 2010 11:18AM

    In order to protect air quality surely we must apply due diligence in energy source decision making especially with regard to combustion. We are madly rushing into biomass systems ignoring the impact to health and plant life. Official reports from Sep 06 comparative emissions impact detailed biomass dirtier than oil and far dirtier than gas, report Jul 06 details NHS cost burden of biomass combustion thro to DEFRA air quality strategy doc again detailing hazards of biomass combustion and confirming all NOX and PM pollution is hazardous to health and environment.
    We know biomass combustion creates NOX burden per unit of energy 3.5 to 11 times higher than alternatives with PM creation hundreds of times higher. We know a small biomass energy plant creates NOX pollution equal to exhaust emissions from 200 million car/km per year.The bizarre situation is the intention to instal a number of large biomass plants in UK requiring import of millions of tonnes of low grade biomass knowing 48% of energy content will be consumed in processing and transport. Principal transport shipping which is known to be a major polluter.The Gov detail anticipated operating efficiency 30% which means 70% converted to pollution and direct climate change. Small biomass energy plants operate in 15 to 23% efficiency range which means even greater waste of resources and more pollution per unit of power output.
    We need renewable energy but surely it should maximise efficiency, minimise pollution and make best use of resources. Can anyone see logic in applying massive subsidies to import biomass with low energy density when there is an abundance of clean burning LNG available worldwide to bridge the energy gap while we develop truly clean,green energy sources. I must have missed the plot because scratching through the greenwash I find it impossible to find any logic in the biomass biomess other than a money making exercise at the expense of the gullible taxpayer/ consumer both financially and healthwise or am I just being cynical.
    Kind Regards
    Brian Wilson

  • ianfalconer

    23 March 2010 11:59AM


    You need to be careful when considering air pollution attributable to biomass combustion. There is a vast difference between simply burning wood in the grate of your fire at home and constructing a power plant to produce heat and electricity.

    If you are quoting figures from the recent UK consultation on the Renewable Heat Initiative ( with respect to feed-in-tariffs, then sorry, you are missing one of the points of large scale biomass use. The large plants are subject to NOx regulations, just the same as any thermal power station, so must employ scrubbers if their atmospheric emissions are excessive under EU and UK law.

    30% efficiency sounds about right for a biomass plant that doesn't capture waste heat and use it. Modern gas power stations are roughly 45% efficient. Coal is about the same as biomass. What each and every one needs is to capture its waste heat for use in industry. Doing that ups systemic efficiency to well over 60%.

    You should also take a look at Wales' plans for a National Forest economy. It takes time to get there but local forestry providing combined heat and power is actually a really efficient way to do things.

    I happen to agree that the incentives for small, household-scale biomass boilers are excessive and highly selective socially since they require a large space to run and are extremely expensive to install.

  • Amedeo

    23 March 2010 2:10PM

    These are interesting statistics but we cant have it all our own way - 70 years ago my Great Uncle died of a respiritory related illness (intollerable working conditions of the factories of Birmingham) his brother died of Daibetis casued through a poor diet and no NHS - Chimnies belched out smoke everywere, London had fataliites through the infamous London Smog - We now live in a Contry thats has good healthcare, people have enough to eat. I am an asmatic and the air quality needs to be improved but this obsession with the missery of environment is a Joke compaired to the way most people once lived.

  • Drcrinum

    23 March 2010 3:22PM

    Monkeybiz & ianfalconer ,
    I am not a botanist, although botany has been a lifelong hobby. I am a physician, a pathologist -- death investigation/analysis has been in my focus for >30 years.

    What we are discussing is epidemiological data and how it is analyzed. It is a very complex issue to address all of the variables involved in these studies, let alone just identify what variables need to be included in the study. It is virtually impossible to obtain and evaluate medical epidemiological data other than on a monthly basis. In general It is too complex an issue if the proper variables are to be included, and only retroactive studies can be accomplished. It also makes a difference where the studies are conducted; for instance, in Europe, every effort is made to keep patients out of the hospital emergency room, whereas in America, acute treatments focus upon rapid transport to the emergency room. There are also all sorts of issues involving decisions about first going to a family doctor versus the emergency room, what treatments a patient is currently receiving or has available at home, what is the patient's prior medical experience and disease history, what other diseases the patient has, etc.

    Many medical studies in the literature are flawed for 2 reasons: 1) All of the variables were not addressed, and 2) The statistical analyses are flawed. For instance, salt was identified years ago in a huge study as being adversely related to high blood pressure disease -- the findings were completely accepted and the results taught for many years. Now it turns out that the study was flawed because all of the variables were not addressed in the analysis and the statistics methodology was wrong. So now there is controversy.

    You need to understand that in these kind of epidemiological studies, individual patients are not scrutinized equally or in detail. Actual medical charts are not reviewed; the investigators never see or examine a patient. The details are extracted from patient summaries provided for each record, often reduced to numerical disease classifications/codes and simplified medical histories, usually summarized as predisposing conditions. Many times a patient has more than one disease process or is hospitalized for a complication of a disease rather than the prime manifestation of the disease, and then the pre-existing medications and compliance enter the equation (often overlooked or not addressed). The number of variables can become enormous, and if a significant variable is overlooked and not included, then the study becomes meaningless. So what you need to understand is that it is a very complicated game, yet usually the game is played in a very simplified and abbreviated way.

    To give you a classic example of why the statistics are often flawed in medical studies -- there have been thousands of studies where the treatment for Disease X was studied using Drug A and Drug B. Specified numbers of patients with Disease X were treated with either Drug A or a placebo, and the same number were treated with either Drug B or a placebo. The usual statistical results found that (for example) Drug A produced better results, so Drug A then became the front line treatment. However, had the study been performed where patients received either Drug A, Drug B or a placebo, then no statistical difference was found between Drug A and Drug B. Do you see the difference in the analysis?

    Now concerning the Koop et al study -- it only addressed respiratory diseases; you can't criticize it for not including heart disease. Almost all previous epidemiological studies have shown an association between air pollution and respiratory diseases. Why? Koop et al are saying that it is the statistical methodology. The latter is an issue that needs to be addressed. That is why I referenced the Siegfried discussion about statistics. (I could also throw in that statistical training is not part of the medical school curriculum, and the people who end up in public health only receive a cursory education in statistics -- real statistical expertise is sorely lacking in medicine.)

    So who is correct? The only answers will eventually come via replication studies over time. Even then there will likely be some controversy; for example, my wife accomplished a major study on hospital admission practices; she consulted my statistical colleague who advised her to use certain statistical methodology, but her professor insisted that a different methodology be used -- there were significant differences in outcomes between the two methods. So which one was correct?

    But real statistical expertise needs to be included ... somehow.

  • ianfalconer

    23 March 2010 9:12PM


    Science is difficult and so is statistics.
    I agree 100%, but if you manipulate data to make it easier you also loose that data's power to predict future outcomes.

    If the problem is that epidemiological data is in too low a time resolution to correlate with high-res air quality data then a researcher should gather it at a sufficiently high time resolution, not de-resolve the high quality data to match the low. That is lazy and poor execution of experimental and statistical method.

    In the past I have seen papers that that workers already have collected high-res medical data, hence the body of evidence linking airborne pollutants with detrimental health effects and will point again to the relevant WHO publication on the topic, which speaks of 3-5 day lags in response to exposure (as you stated in a previous posting), which can only be evidenced if daily data is available.

    I also agree that veracity is a measure of scientific truth. The veracity of Koop et al falls by the wayside at the first inspection of their statisitcal rigour.

    If I spotted a mistake in my child's maths homework, say 5+5=100, should I tell them that 'its fine, its a good effort, I especially like the glitter that you used to pick out the plus sign'. No, the answer is wrong and letting it pass is wrong because my child would learn nothing except not to trust Daddy to check their maths homework. In Koop's case 'Daddy' was the peer-review panel of Environmental Modelling and Software. I suspect that Daddy had a bad night before he checked this particular piece of homework, otherwise he would have picked up the really fundamental statistical error in the methodology - the de-resolution of the time-series air quality data.

    Just because they use a novel methodology, it doesn't mean that its right, irrespective of the field.

    Interdisciplinary work is doubly difficult because you have to keep two or more fields happy, but if you dip your toe into a pool of knowledge that you don't know well, you'd better learn how to swim quick ! And a good physical scientist, one who knows the limitations of experimental data and statistical manipulation used in their own methodology, would not cry out on the basis of a single paper that 50 years of evidence is wrong. That to quote Koop "In the most general
    specifications we find the here-observed health effects of air pollution are very small and insignificant, with signs that are typically opposite to conventional expectations." In other words their study finds that exposure to particulates and other air pollution is good for you. Well then, let's all suck a tail pipe next time we have a cold ! I'm sorry but this is my last post on this topic. Its soul destroying.

    For what its worth, my own research as a physical scientist wasn't worth publishing. The data was incontravertable because I was rigorous about its treatment. It just didn't say anything that interesting. Lots of science is like that. The answer to the particular question that I asked was 'I don't know, the data doesn't say one way or the other'. So I never published and moved on to pastures new, taking my statistical rigour and desire for veracity with me (along with a whole head full of useful information about dust).

    My suggestion is that Koop et al do the same (only they should learn how to do science before they amscray).

    I have made and will make no comments whatsoever on your good self or your wife's research, my gripe is with Koop being cited as a source for anything apart from a comedic routine. That paper is flawed from its foundations and in extremis.

  • Monkeybiz

    24 March 2010 4:12AM

    @DrC. I don't disagree with much of what you say, though I do think the Koop paper is flawed in all sorts of ways. It remains an outlier and we have to ask why. It MIGHT be that everyone else was wrong, but experience tells me that is seldom the case. We're not talking about one study changing the whole paradigm here like the dietary sodium paper did, but the reliance on the Koop study suggests that we should do exactly this. That is what I object to.

    I do not agree with the following statement you made:

    (I could also throw in that statistical training is not part of the medical school curriculum, and the people who end up in public health only receive a cursory education in statistics -- real statistical expertise is sorely lacking in medicine.)

    First, I'm not sure when you went to Med school, perhaps then it wasn't usual to teach stats but in the med school I'm most familiar with, it has been taught in one form or another for over 25 years to undergrad medics. Second, I would argue that for a department of public health to be able to survive academically these days they would need to have a robust command of, and skills in statistics. Many employ cohorts of statisticians and mathematicians because the analyses using large cohorts and data sets are so complex, as you point out. In my experience, it is the other departments in the medical schools who lack the statistical expertise and think that they can answer the question about the efficacy of a drug with a sample size of 30 and funding from the drug manufacturer, and thereby obtain an unbiased estimate of the true benefit.

Comments on this page are now closed.

Bestsellers from the Guardian shop

Guardian Bookshop

This week's bestsellers

  1. 1.  London's Lost Rivers

    by Paul Talling £9.99

  2. 2.  Atlantic

    by Simon Winchester £9.99

  3. 3.  Teach Yourself Volcanoes, Earthquakes and Tsunamis

    by David Rothery £10.99

  4. 4.  Cloudspotter's Guide

    by Gavin Pretor-Pinney £9.99

  5. 5.  Cloud Collector's Handbook

    by Gavin Pretor-Pinney £10.00