Particulate Matter Air Pollution: Science, Controversy, and Impact on Public Health C. Arden Pope III, PhD Mary Lou Fulton Professor of Economics Presented at the Western States Air Resources Council (WESTAR) Denver, Colorado November 7, 2013
Illustration for Charles Dickens Sketches by Boz, by George Cruikshank, 1838 The chimney, smoke Stack, and smoke and fog (smog) have long played important iconic roles in the art and literature of London.
Coffee Stall by Gustave Doré, In a description of a representative London Fog day (April 6, 1870) Doré s friend Blanchard Jerrold say s of an experience with Doré: We were well into the morning and it was as dark as the darkest midnight... It was choking: it made the eyes ache. From: London: A pilgrimage, 1872 and Doré s London: All 180 illustrations from London, A Pilgrimage, 2004
Claude Monet, Houses of Parliament, London, 1903
Houses of Parliament (Palace of Westminster), London
Claude Monet, Houses of Parliament, London, 1904
Claude Monet, Waterloo Bridge, London, 1903
Waterloo Bridge, London, July 2006 (St Paul s Cathedral, Tower 42, 30 St Mary Axe)
Meuse Valley, Belgium Smog, December 1-5, 1930 --60 deaths, 10 time expected
Donora, PA (Severe episode, Oct. 27-31, 1948) From, Public Health Service, Bulletin No. 306, 1949
Danora, PA. Noon, Oct 29, 1948. http://www.weather.com/multimedia/videoplayer.html?clip=12603&collection=257&fro
Donora, PA administering oxygen --Approx. 20 deaths, 6 times expected
London, England (Dec. 5-9, 1952)
London Fog Episode, Dec. 1952 From: Brimblecombe P. The Big Smoke, Methuen 1987
The killer episodes spurred air pollution policies in the U.S. Killer Smog (and related research) 1930: Meuse Valley, Belgium 1948: Donora, PA 1952: London, England Public Policy 1955-1967: Air pollution control acts 1967, 1970: Clean Air Act, Major Amendments, National Environmental Policy Act 1977 & 1990: Major Amendments to Clean Air Act Results 1980 s: Elimination of extreme Killer Smog episodes & Improved air quality By the 1980 s, many thought the air pollution problem had been solved
Geneva Steel, Utah Valley, 1989 (PM 10 = 150 mg/m 3 )
Utah Valley, 1989, (PM10 = 220 mg/m 3 )
Particulate matter air pollution is a mixture of particles suspended in the air. These particles vary in: -Size, -Shape, -Surface area, -Chemical composition, -Solubility, and -Origin. Electron Microscope images of soot particles From: Park K, Cao F, Kittelson DB, McMurry PH. Environ Sci Technol 2003
How small are fine particles? Human Hair (60 mm diameter) PM 10 (10 mm) PM 2.5 (2.5 mm)
Magnified ambient particles (www.nasa.gov/vision/earth/environment)
Utah Valley, 1980s Winter inversions trap local pollution Natural test chamber Local Steel mill contributed ~50% PM 2.5 Shut down July 1986-August 1987 Natural Experiment
When the steel mill was open, total children s hospital admissions for respiratory conditions approx. doubled. mg/m 3 /Numbers of Admissions 300 250 200 150 100 50 mg/m 3 /Numbers of Admissions 300 Children's Respiratory Hospital Admissions Fall and Winter Months, Utah Valley 250 200 150 100 Mean PM 10 levels for Months Included 50 Children's Respiratory Hospital Admissions Fall and Winter Months, Utah Valley Mean High PM 10 levels for Months Included Mean PM 10 levels for Months Included Mean High PM 10 levels for Months Included Bronchitis and Asthma Pneumonia and Pleurisy Bronchitis and Asthma Total Pneumonia and Pleurisy Total Mill Open Mill Closed 0 0 PM 10 concentrations Children's respiratory hospital admissions PM 10 concentrations Children's respiratory hospital admissions Sources: Pope. Am J Pub Health.1989; Pope. Arch Environ Health. 1991 Sources: Pope. Am J Pub Health.1989; Pope. Arch Environ Health. 1991
Health studies take advantage of highly variable air pollution levels that result from inversions. 100 90 Utah Valley (Lindon Monitor) 80 70 g/m 3 60 50 40 30 20 10 0 100 90 Salt Lake Valley (Hawthorn Monitor) 98 99 00 01 02 03 04 05 06 07 08 09 10 Salt Lake Valley (Hawthorn Monitor) 80 70
# of Deaths Daily changes in air pollution daily death counts Utah Valley Time (days)
Poisson Regression Count data (non-negative integer values). Counts of independent and random occurrences classically modeled as being generated by a Poisson process with a Poisson distribution: Prob (Y = r) = e (-λ) λ r r! Note: λ = mean and variance. If λ is constant across time, we have a stationary Poisson process. If λ changes over time due to changes in pollution (P), time trends, temperature, etc., this non-stationary Poisson process can model as: ln λ t = α + β(w 0 P t + w 1 P t-1 + w 2 P t-2 +...) + s 1 (t) + s 2 (temp t ) +... Modeling controversies How to construct the lag structure? (MA, PDL, etc.) How aggressive do you fit time? (harmonics vs GAMs, df, span, loess, cubic spline, etc.) How to control for weather? (smooths of temp & RH, synoptic weather, etc.) Also: How to combine or integrate information from multiple cities
Daily time-series studies ***of over 200 cities*** % increase in mortality 3 2 1 29 cities (Levy et al. 2000) Estimates from meta analysis GAM-based studies (Stieb et al. 2002, 2003) Non GAM-based studies (Stieb et al. 2002, 2003) Unadjusted (Anderson et al. 2005) Publication bias adjusted (Anderson et al. 2005) 18 Latin Am. studies (PAHO 2005) 6 U.S. cities (Klemm and Mason 2003) Estimates from Multicity studies 8 Canadian cities (Burnett and Goldberg 2003) 9 Californian cities (Ostro et al. 2006) 10 U.S cities (Schwartz 2000, 2003) 14 U.S cities, case-crossover (Schwartz 2004) NMMAPS, 20-100 U.S. cities (Dominici et al. 2003) APHEA-2, 15-29 European cities (Katsouyanni et al. 2003) 9 French cities (Le Tertre et al. 2002) 7 Korean cities (Lee et al. 2000) 13 Japanese cities (Omori et al. 2003) Estimates from meta analysis from Asian Lit 0 20 g/m 3 PM 10 20 g/m 3 PM 10 20 g/m 3 PM 10 20 g/m 3 PM 10 20 g/m 3 PM 10 20 g/m 3 PM 10 10 g/m 3 PM 2.5 10 g/m 3 PM 2.5 10 g/m 3 PM 2.5 20 g/m 3 PM 10 20 g/m 3 PM 10 20 g/m 3 PM 10 20 g/m 3 PM 10 20 g/m 3 BS 40 g/m 3 TSP 20 g/m 3 SPM Review of Asian Lit.--8 studies (HEI Report, Table TS2) 20 g/m 3 PM 10 20 g/m 3 PM 10 20 g/m 3 PM 10 PAPA Studies--4 studies (HEI Report, Table TS2) Asian Lit. incorporating PAPA studies (HEI Report, Table TS2) 10 mg/m 3 PM 2.5 or 20 mg/m 3 PM 10 0.4% to 1.5% increase in relative risk of mortality Small but remarkably consistent across meta-analyses and multi-city studies.
2006;114:2443-48 Jeffrey Anderson Methods: Case-crossover study of acute ischemic coronary events (heart attacks and unstable angina) in 12,865 well-defined and followed up cardiac patients who lived on Utah s Wasatch Front and who underwent coronary angiography
Binary Data, classic time-series y t 1 0 t Binary Data, case-crossover y t 1 Conditional Logistic Reg. 0 t Each subject serves as his/her own control. Control for subject-specific effects, day of week, season, time-trends, etc. by matching
Conditional logistic regression: Prob (Y t = 1) 1 - Prob (Y t = 1) ln ( ) = α 1 + α 2 + α 3 +... + α 12,865 + β(w 0 P t + w 1 P t-1 + w 2 P t-2 +...) Control by matching for: All cross-subject differences (in this case, 12,865 subject-level fixed effects), Season and/or month of year, Time trends, Day of week Modeling controversies: How to select control or referent periods. Time stratified referent selection approach (avoids bias that can occur due to time trends in exposure) (Holly Janes, Lianne Sheppard, Thomas Lumley Statistics in Medicine and Epidemiology 2005)
15.00 PM 2.5 % 10.00 5.00 Concurrent day 1-day lag 2-day lag 3-day lag 2-day moving av. 3-day moving av. 4-day moving av. Concurrent day 1-day lag 2-day lag PM 10 3-day lag 2-day moving av. 3-day moving av. 4-day moving av. 0.00 Figure 1. Percent increase in risk (and 95% CI) of acute coronary events associated with 10 g/m 3 of PM 2.5, or PM 10 for different lag structures.
Short-term PM exposures contributed to acute coronary events, especially among patients with underlying coronary artery disease. % 20.00 15.00 10.00 5.00 All acute coronary Unstable Angina Index MI Subsequent MI Age<65 Age>=65 Female Male Non Smoking Smoking BMI<30 BMI>=30 CHF,no CHF,yes Hypertension,yes Hypertension,no Hyperlipidemia,no Hyperlipidemia,yes Diabetes,yes Diabetes,no Family history,yes Family history,no # of Risk Factors 4+ 3 0 1 2 # of Diseased Vessels 2 0 1 3 0.00-5.00-10.00 Figure 2. Percent increase in risk (and 95% CI) of acute coronary events associated with 10 mg/m 3 of PM 2.5, stratified by various characteristics.
Short-term changes in air pollution exposure are associated with: Daily death counts (respiratory and cardiovascular) Hospitalizations Lung function Symptoms of respiratory illness School absences Ischemic heart disease Etc.
Longer-term air pollution exposure has been linked to even substantially larger health effects.
Age-, sex-, and race- adjusted populationbased mortality rates in U.S. cities for 1980 plotted over various indices of particulate air pollution (From Pope 2000). Adjusted Mortality for 1980 (Deaths/Yr/100,000) 1000 950 900 850 800 750 700 650 600 8 10 12 14 16 18 20 22 24 26 28 30 32 34 Median PM 2.5 for aprox. 1980
An Association Between Air Pollution and Mortality in Six U.S. Cities 1993 Dockery DW, Pope CA III, Xu X, Spengler JD, Ware JH, Fay ME, Ferris BG Jr, Speizer FE. Methods: 14-16 yr prospective follow-up of 8,111adults living in six U.S. cities. Monitoring of TSP PM 10, PM 2.5, SO 4, H +, SO 2, NO 2, O 3. Data analyzed using survival analysis, including Cox Proportional Hazards Models. Controlled for individual differences in: age, sex, smoking, BMI, education, occupational exposure.
Clean cities Average Polluted cities Highly Polluted cities
Cox Proportional Hazards Survival Model Cohort studies of outdoor air pollution have commonly used the CPH Model to relate survival experience to exposure while simultaneously controlling for other well known mortality risk factors. The model has the form ( l) ( t) ( l) ( t) exp T x ( l) ( t) i 0 i Hazard function or instantaneous probability of death for the i th subject in the l th strata. Baseline hazard function, common to all subjects within a strata. Regression equation that modulates the baseline hazard. The vector X i (l) contains the risk factor information related to the hazard function by the regression vector β which can vary in time.
Adjusted risk ratios (and 95% CIs) for cigarette smoking and PM 2.5 Cause of Death Current Smoker, 25 Pack years All 2.00 (1.51-2.65) Most vs. Least Polluted City 1.26 (1.08-1.47) Lung Cancer Cardiopulmonary All other 8.00 (2.97-21.6) 2.30 (1.56-3.41) 1.46 (0.89-2.39) 1.37 (0.81-2.31) 1.37 (1.11-1.68) 1.01 (0.79-1.30)
Particulate Air Pollution as a Predictor of Mortality in a Prospective Study of U.S. Adults Pope CA III, Thun MJ, Namboodiri MM, Dockery DW, Evans JS, Speizer FE, Heath CW Jr. Michael Thun 1995 Clark Heath Methods: Linked and analyzed ambient air pollution data from 51-151 U.S. metro areas with risk factor data for over 500,000 adults enrolled in the ACS- CPSII cohort.
Adjusted mortality risk ratios (and 95% CIs) for cigarette smoking the range of sulfates and fine particles Cause of Death Current Smoker All 2.07 Lung Cancer Cardio- Pulmonary (1.75-2.43) 9.73 (5.96-15.9) 2.28 (1.79-2.91) All other 1.54 (1.19-1.99) Sulfates 1.15 (1.09-1.22) 1.36 (1.11-1.66) 1.26 (1.16-1.37) 1.01 (0.92-1.11) Fine Particles 1.17 (1.09-1.26) 1.03 (0.80-1.33) 1.31 (1.17-1.46) 1.07 (0.92-1.24)
25 July 1997
Dan Krewski Rick Burnett Mark Goldberg and 28 others
Legal uncertainty largely resolved with 2001 unanimous ruling by the U.S. Supreme Court.
Mortality Risk Ratio Reduction in fine particulate air pollution: Extended follow-up of the Harvard Six Cities Study (Laden, Schwartz, Speizer, Dockery. AJRCCM 2006) Francine Laden Joel Schwartz 1.4 1.3 1.2 1.1 1.0 0.9 0.8 0.7 Kingston Steubenville Topeka Portage St. Louis Watertown 0 5 10 15 20 25 30 35 PM 2.5 (mg/m 3 )
2004. John Godleski RR (95% CI) 1.40 1.35 1.30 1.25 1.20 1.15 1.10 1.05 1.00 0.95 0.90 All Cardiovascular plus Diabetes Ischemic heart disease Dysrhythmias,Heart failure, Cardiac arrest Hypertensive disease Other Atherosclerosis, aortic aneurysms Cerebrovascular Diabetes Other Cardiovascular Respiratory Diseases COPD and allied conditions Pneumonia, Influenza All other respiratory 0.85 0.80 0.75 0.70 0.65
Pope and Dockery, JAWMA 2006. PM Inhalation Heart Altered cardiac autonomic function Increased dysrhythmic susceptibility Altered cardiac repolarization Increased myocardial Ischemia Heart failure exacerbation Vasculature Atherosclerosis, accelerated progression of and destabilization of plaques Endothelial dysfunction Vasoconstriction and Hypertension Lungs Inflammation Oxidative stress Accelerated progression and exacerbation of COPD Increased respiratory symptoms Effected pulmonary reflexes Reduced lung function Systemic Inflammation Oxidative Stress Increased CRP Proinflammatory mediators Leukocyte & platelet activation Blood Altered rheology Increased coagulability Translocated particles Peripheral thrombosis Reduced oxygen saturation Brain Increased cerebrovascular ischemia
Breathing Contaminants The Global Burden of Disease 2010
The Global Burden of Disease 2010 Breathing contaminates contributes to global burden of disease (GBD) Number of attributable deaths Disability adjusted life-years (DALYs) Tobacco Smoking 5.7 mil. 5.7% Second Hand Smoke 0.6 mil. 0.6% Household air pollution from solid fuels 3.5 mil. 4.5% Ambient PM air pollution 3.2 mil. 3.1% Ambient Ozone 0.2 mil. 0.1%
So, an obvious question Has reducing air pollution resulted in substantial and measurable improvements in human health?
Do cities with bigger improvements in air quality have bigger improvements in health, measured by life expectancy? Fine-Particulate Air Pollution and Life Expectancy in the United States January 22, 2009 C. Arden Pope, III, Ph.D., Majid Ezzati, Ph.D., and Douglas W. Dockery, Sc.D. - Matching PM 2.5 data for 1979-1983 and 1999-2000 in 51 Metro Areas - Life Expectancy data for 1978-1982 and 1997-2001 in 211 counties in 51 Metro areas - Evaluate changes in Life Expectancy with changes in PM 2.5 for the 2-decade period of approximately 1980-2000.
A 10 µg/m 3 decrease in PM 2.5 was associated with a 7.3 (± 2.4) month increase in life expectancy. This increase in life expectancy persisted even after controlling for socio-economic, demographic, or smoking variables
Finally We seem to be making major progress understanding the health effects of air pollution. In the U.S. reducing air pollution has resulted in substantial and measurable improvements in human health. Good News, Right?
Smith and Stewart Subpoena EPA s Secret Science Aug 2, 2013 Press Release (http://stewart.house.gov/) Washington, D.C. Science, Space, and Technology Committee Chairman Lamar Smith (R-Texas) and Environment Subcommittee Chairman Chris Stewart (R-Utah) issued a subpoena to the EPA, forcing the agency to release the secret science it uses as the basis for costly air regulations Will House Science Panel Need an Ethical Review? Kelly Servick, Aug. 12, 2013 Eddie Bernice Johnson and Lamar Smith Chris Stewart Two Utahns have stake in pollution health data fight Environment» Is congressional inquiry into pollution-health studies after truth or partisan gain? By Judy Fahys The Salt Lake Tribune Aug 27 2013 House Subpoenas Personal Medical Information in Continued Assault on Clean Air Policies by Sam Abbott, 8/6/2013 Citizen Health & Safety, Safeguarding Public Health and the Environment, Setting and Enforcing Regulations, Fostering Scientific Integrity, Environmental Protection Agency (EPA)
Working Draft, Aug. 2, 2013, C. Arden Pope III, P Harvard Six-Cities and ACS CPS-II Cohort Studies of Air Pollution and Mortality: Initial, Independent, Extended, and Replicative Analyses Harvard Six-Cities Study Dockery et al. New England Journal of Medicine (NEJM), 1993 ACS CPS-II Cohort Study Pope, et al. Am J Respir Crit Care Med (AJRCCM), 1995 Independent Re-analyses of Harvard Six-Cities and ACS CPS-II Studies Krewski et al. HEI Special Report, 2000; J Tox Enviro Health, Special Issue, 2003 3-yr reanalysis by a team of 31 independent researchers with oversight from a 9-member expert panel and peer review by a special panel of the HEI Health Review Committee. Included full data access that insured the privacy and confidentiality of research participants. Reanalyses include data audits, full replication and validation, and extensive sensitivity analyses. Extended analyses of Harvard Six-Cities study Laden et al. AJRCCM, 2006 Schwartz et al. EHP, 2008 Lepeule et al. EHP, 2012 Extended analyses of ACS CPS-II study Pope et al. JAMA, 2002; Pope et al. Circulation, 2004; Jerrett et al. Epidemiology, 2005; Krewski et al. HEI Rep. 2009 Jerrett et al. NEJM, 2009; Smith et al. Lancet, 2009 Turner et al. AJRCCM, 2011; Jerrett et al. AJRCCM, 2013 Replicative studies in other cohorts: Miller et al. (Women s Health Initiative) NEJM, 2007; Beelen et al. (Netherlands) EHP, 2008; Zeger et al. (U.S. National Medicare) EHP, 2008; Puett et al. (Nurses Health Study) EHP, 2009; Puett et al. (Health Professionals) EHP, 2011; Hart et al. (U.S. Truckers) AJRCCM, 2011; Lipsett et al. (California Teachers) AJRCCM, 2011; Crouse et al. (Canadian ) EHP, 2012; Cesaroni et al. (Rome) EHP, 2013 **See recent meta-analytic review by Hoek et al. Environ Health, 2013**
A Broader View of the PM Air Pollution and Human Health Scientific Literature Episode studies of morbidity and mortality Meuse Valley, Belgium, 1930 Donora, PA, 1948 London Smog, 1952 Intervention/Natural Experiment studies -Intermittent operation of a steel mill. Pope AJPH, 1989, Ghio JAM, 2004 -US Life Expectancy and pollution reduction. Pope et al. NEJM, 2009 -Ireland coal ban. Dockery et al. HEI, 2013 -China Huai river RD. Chen et al. PNAS, 2013 --and others Cardio- and cerebrovascular disease events studies Peters et al. Circulation 2001 Pope et al. Circulation 2006 Wellenius, et al. Arch Intern Med, 2012 --and many others Population-based cross-sectional mortality studies Lave, Seskin. Science, 1970 Evans et al. Environ Int, 1984 Ozkaynak, Thurston. Risk Anal, 1987 --and others Cohort-based Mortality Studies Time-series and case-crossover daily mortality studies Dominici et al. (NMMAPS) HEI, 2003 Zanobetti et al. (112 US cities) EHP, 2009 Analitis et al. (15-29 Euro cities) Epi, 2006 Anderson et al. (meta anal) Epi. 2005 --and many, many others Time-series and case-crossover hospitalization studies Schwartz (8 cities). Epi, 1999 Dominici et al. (204 US counties) JAMA, 2006 --and many others Lung function/resp. symptoms studies Dockery et al. JAWMA, 1982 Pope et al. ARRD, 1991 Hoek et al. Eur Respir J, 1998 Gauderman et al. NEJM, 2004 --and many others Subclinical markers of cardiovascular health studies Including systemic inflammation, oxidative stress, endothelial cell activation, thrombosis and coagulation, vascular dysfunction and atherosclerosis, blood pressure, altered cardiac autonomic function/hrv, etc.) For overview see Brook et al. Circulation 2010 Working Draft, Aug. 2, 2013, C. Arden Pope III, PhD Controlled human exposure and animal toxicological studies For overview see Brook et al. Circulation 2010 Others The literature is large, complex, and growing.
The assertion that the Harvard Six-Cities and the ACS cohort studies are secret is simply false on its face. The methodology, protocol, and results of these studies are openly published in peer-reviewed scientific journals. As is common in medical and public health studies, these studies include analysis of some private data. Releasing private data is generally unethical and often in violation of confidentiality agreements with research participants, other data providers, and mandates of Institutional Review Boards for human subjects. For the Harvard Six-Cities and the ACS cohort studies, full data access has been provided for independent data audits, validation, and reanalysis under conditions that protected the privacy and confidentiality of research subjects. For approximately two decades, multiple research teams have been actively involved in open, collaborative, extended analyses. Results of the studies have been reproduced, replicated, extended, and documented in multiple peer-reviewed publications.
Biggest criticisms regarding the overall results: 1. The effects aren t big enough to be compelling (need RR > 2.0) 2. The effects are too large to be biologically plausible based on an extrapolation of smoking literature.
Pack-a-day smoker: RR ~ 2 Daily inhaled dose ~ 240 mg 2.0 Adjusted Relative Risk 1.5 Live in polluted city or With smoking spouse RR ~ 1.15 1.35 Daily inhaled dose ~ 0.2 1.0 mg 1.0 0 60 120 180 240 estimated daily dose of PM 2.5, mg 18-22 cigs/day
Pope, Burnett, Krewski, et al. 2009. 2.5 Adjusted Relative Risk 2.0 1.5 1.0 <3 cigs/day 4-7 cigs/day 8-12 cigs/day 13-17 cigs/day 0 60 120 180 240 300 estimated daily dose of PM 2.5, mg 18-22 cigs/day 23+ cigs/day Figure 1. Adjusted relative risks (and 95% CIs) of IHD (light gray), CVD (dark gray), and CPD (black) mortality plotted over estimated daily dose of PM 2.5 from different increments of current cigarette smoking. Diamonds represent comparable mortality risk estimates for PM 2.5 from air pollution. Stars represent comparable pooled relative risk estimates associated with SHS exposure from the 2006 Surgeon General s report and from the INTERHEART study.
Adjusted Relative Risk 2.5 2.0 1.5 Exposure from Second hand cigarette smoke: Stars, from 2006 Surgeon General Report and INTERHEART study And air pollution: Hex, from Womens Health Initiative cohort Diamonds, from ACS cohort Triangles, Harvard Six Cities cohort Exposure from smoking <3, 4-7, 8-12, 13-17, 18-22, and 23+ cigarettes/day Figure 2. Adjusted relative risks (and 95% CIs) of ischemic heart disease (light gray), cardiovascular (dark gray), and cardiopulmonary (black) mortality plotted over baseline estimated daily dose (using a log scale) of PM 2.5 from current cigarette smoking (relative to never smokers), SHS, and air pollution. 1.0 0.1 1.0 10.0 100.0 estimated daily dose of PM 2.5, mg
40 Lung Cancer 35 Adjusted Relative Risk 30 25 20 15 1.50 10 1.25 5 1.00 0.0 0.5 1.0 3.0 Ischemic heart (light gray) Cardiovascular (dark gray) Cardiopulmonary (black) Adjusted Relative Risk 2.5 2.0 1.50 1.5 1.25 1.0 1.00 0.0 0.5 1.0 0 60 120 180 240 300 360 420 480 540 (<3) (4-7) (8-12) (13-17) (18-22) (23-27) (28-32) (33-37) (38-42) (>42) Estimated daily dose of PM 2.5, mg (cigarettes smoked per day)
Marginal Costs of Abatement Marginal Health Costs of Pollution $ Pollution Abatement C* C H Ambient Concentrations ( g/m 3 ) C NC
Marginal Costs of Abatement Marginal Health Costs of Pollution $ C* C H Ambient Concentrations ( g/m 3 ) C NC
Marginal Costs of Abatement Marginal Health Costs of Pollution $ C* C H Ambient Concentrations ( g/m 3 ) C NC