Air Pollution and Health Sheryl Magzamen MESA-Air Pollution

Air Pollution and Health Sheryl Magzamen MESA-Air Pollution Study Presentation for Environmental Sciences 100 University of Washington February 26, 2010

Today’s Presentation Air pollution and health: overview n Nuts and bolts of epidemiology n Asthma and Air Pollution n n Studies in schools n Correlation v. Causation n Three “natural” experiments Exposure Assessment n Putting it all together n Environmental Studies 100 February 26, 2010

Health Effects: Air Pollution Map courtesy of WHO Environmental Studies 100 February 26, 2010

Health Effects: Air Pollution n WHO estimates 3 million premature deaths/yr attributable to air pollution n n Developing world, ~90% attributable to indoor air pollution 800, 000 deaths attributed to outdoor air pollution (Cohen et al. 2005) n n United States: 20, 000 premature deaths/year attributable to air pollution n 1. 2% of all deaths globally Aggravated respiratory disease Exacerbates cardiovascular disease Increases in hospital admissions, emergency department visits, and restricted activity days Environmental Studies 100 February 26, 2010

Sources of Air Pollution Source: Vallero 2008 Environmental Studies 100 February 26, 2010

Sources of Air Pollution Source: EPA Environmental Studies 100 Annual VMT, United States (Source: Wall Street Journal) February 26, 2010

Traffic Related Air Pollution 4 -stroke cycle 2 -stroke cycle Diesel HC HC CO SOx CO CO 2 PM CO 2 NOx Environmental Studies 100 NOx PM CO 2 NOx Source: Vallero 2008 February 26, 2010

Traffic Related Air Pollution Image courtesy of the EPA Environmental Studies 100 February 26, 2010

Traffic Related Air Pollution Environmental Studies 100 February 26, 2010

What is epidemiology? 1. 2. 3. 4. 5. The branch of philosophy concerned with the nature and scope of knowledge The study of the history of words, their form and meaning The scientific study of insects The study of the distribution and determinants of disease in a population Branch of medicine dealing with skin and its diseases

General Epidemiologic Model (analytic) (descriptive) Exposure Health Outcome Environmental Studies 100 February 26, 2010

Air Pollution Epidemiology Where did I put that donut ? Epidemiology focuses on the presence and magnitude of the exposure effects on health outcomes Environmental Studies 100 February 26, 2010

Epidemiologic Outcomes Morbidity n Symptoms n Exacerbations n Health care events (i. e. hospitalizations, emergency department visits) n Limitations in activity n Changes in physiologic measurements Environmental Studies 100 Mortality Death February 26, 2010

Asthma Outcomes Morbidity Mortality n Wheeze, trouble sleeping n Use of inhaler n Unplanned health care Very rare (~5000 use cases/year US) n School/work absence n Impacts on quality of life n Lung function Subjective measures (patient reports), objective measures (physiologic measures, administrative data) Environmental Studies 100 February 26, 2010

Epidemiologic Measures Incidence Prevalence Incidence is used to understand etiology, or why the disease occurs; prevalence is used often for planning Environmental Studies 100 February 26, 2010

Epidemiologic Measures: Air pollution In the Mc. Connell et al. paper, what health outcome was measured? 1. 2. 3. 4. Prevalence of asthma cases Incidence of asthma hospitalizations Prevalence of heart attacks

Oakland Kicks Asthma • Asthma surveillance and intervention program • Students complete a brief, in-class asthma survey • Students identified as “current asthma” are eligible for intervention(s) Environmental Studies 100 February 26, 2010

Air Pollution and Asthma Table 1. Demographic characteristics of cohort by quartile of NO 2 exposure Q 1 Q 2 Q 3 Q 4 6. 13 - 16. 14 – 17. 14 – 18. 14 – μg∙m-3 17. 13 μg∙m-3 18. 13 μg∙m-3 24. 10 μg∙m -3 (n= 186) (n=202) (n=229) (n=188) chisq Race/Ethnicity A/PI 10. 2 8. 9 13. 5 12. 8 0. 41 Black 41. 9 49. 0 42. 4 46. 3 0. 42 Latino 16. 1 19. 8 14. 0 16. 5 0. 44 White 10. 2 0. 5 2. 6 4. 3 <0. 0001 Nat. Am 0 0. 5 0 0 0. 39 Mixed 9. 1 7. 9 9. 2 9. 0 0. 96 Other 4. 8 8. 4 8. 3 3. 7 0. 12 Environmental Studies 100 February 26, 2010

Air Pollution and Asthma Table 2. Logistic Regression Results – High v. Low Quartile Q 4 v. Q 1 OR (crude) Night sx (prev mo) Q 4 v. Q 1 95% CI (crude) (1. 01 – 2. 57) ED visit 4. 52 (2. 69 – 7. 60) Asthma rx 1. 35 (0. 77 – 2. 35) White 0. 68 (0. 25 – 1. 86) 1. 57 (1. 01 – 2. 43) ED visit 2. 73 (1. 64 – 4. 56) Asthma rx 1. 30 (0. 78 – 2. 17) White 0. 56 (0. 21 – 1. 50) Environmental Studies 100 1. 68 (1. 11 – 2. 64) Q 4 v. Q 1 OR (adjusted) 1. 61 Can’t complete sentence (prev yr) 1. 71 Q 4 v. Q 1 95% CI (crude) (1. 11 – 2. 56) February 26, 2010

Observation v. Experiment How do we distinguish correlation from causation? Environmental Studies 100 February 26, 2010

Correlation v. Causation Football and the stock market: If an old AFL (now AFC) wins the Superbowl, the stock market will be down. But…if winner is from the old NFL (now NFC), the market will be up. Has been on the money 33 of 41 years (80% success rate) Between 1967 and 1997, it was right 28 out of 31 times (better than 90%!) Environmental Studies 100 February 26, 2010

Causality in Controlled Experiments Randomize Adolescents w/ Asthma dx to be… <500 ft from Freeway? YES Emergency Care Utilization? Y N NO Emergency Care Utilization? Y N

Causality in Controlled Experiments Image courtesy of Environmental Health Perspectives Environmental Studies 100 February 26, 2010

Observation v. Experiment Acute Air Pollution Episodes 1930 Meuse Valley Smog (photo courtesy of Fluoride Action Network) Environmental Studies 100 1948 Donora Smog 1952 London Smog (photo courtesy of The Allegheny Front) (photo courtesy of BBC) November 23, 2009

Observation v. Experiment Environmental Studies 100 February 26, 2010

Causal Thinking: Natural Experiments Environmental Studies 100 February 26, 2010

1996 Summer Olympics Friedman et al. (2001): • Compared number of childhood asthma acute care events in the Atlanta Metro Area: • 17 days of the 1996 Summer Olympic Games • 4 weeks directly prior to and after the Games Rationale: Increase in alternative transportation options for the Olympics, including: 24 hour public transportation, addition of 1, 000 buses, and encouragement of telecommuting.

1996 Summer Olympics n n 46% decrease in acute asthma events among Medicaid children during Summer Games compared to baseline periods. Non-asthma acute events decreased 3% during Summer Games compared to baseline periods. Table 1. Changes in Pollutant Levels from Baseline to Summer Games 1 -hr Peak O 3 -27. 9%* CO -18. 5%* NO 2 -6. 8% PM 10 -16. 1%* SO 2 +22. 1*

1996 Summer Olympics Table 2. Changes in Traffic from Baseline to Summer Games Weekday 1 -hr morning peak traffic counts -22. 5%* Weekend morning peak traffic counts -9. 7%* Public transportation ridership 217%* Conclusion: Alternative transportation options for the 1996 Summer Olympics reduced air pollution and traffic volume Follow-Up: Similar findings for the relation between traffic mitigation and asthma hospitalizations for the 2002 Summer Asian Games in Busan, South Korea (Lee et al. 2007) Environmental Studies 100 February 26, 2010

Summer Camp Studies n n n Children tend to spend more time outside over the summer Due to meteorological conditions, ozone levels are generally highest in summer months Change in health with greater ozone exposure? Environmental Studies 100 February 26, 2010

Summer Camp Studies n Series of six studies conducted in US and Canada that examined lung function and ozone exposure (Kinney et al. 1996) n n n Tested lung function on non-asthmatic children every afternoon in various US/Canadian summer camps Obtained daily ozone readings for each day lung function was measured Negative relationship between daily ozone measures and FEV 1 Environmental Studies 100 February 26, 2010

German Reunification n Major improvements in air pollution levels in East Germany 1 – 3 years post reunification (1990) Findings: • Higher prevalence of non-asthmatic respiratory illness for children growing up in areas of higher of TSP and SO 2 • Reductions in prevalence of bronchitis, colds and infections decreased in parallel with reduction of TSP and SO 2 in 1992 – 93 and 1995 – 96. Environmental Studies 100 Heinrich et al. 2000 February 26, 2010

What are other considerations? My house ¼ mi What are important considerations in estimation? Environmental Studies 100 February 26, 2010

Exposure Assessment Many studies use proxy variables to estimate exposure to ambient air pollution <=150 m = “Exposed” >150 m = “Not Exposed” Environmental Studies 100 February 26, 2010

Exposure Assessment Environmental Studies 100 February 26, 2010

Exposure Assessment Environmental Studies 100 February 26, 2010

Exposure Assessment Environmental Studies 100 February 26, 2010

Exposure Assessment What are other important considerations when trying to estimate exposure? n Meteorology Courtesy of NASA Environmental Studies 100 February 26, 2010

Exposure Assessment What are other important considerations when trying to estimate exposure? n Time activity patterns Environmental Studies 100 February 26, 2010

Does Epidemiology Have All the Answers? Taubes: Do We Really Know What Makes Us Healthy? Epidemiology: Do We Really Know What Makes Us Unhealthy? Environmental Studies 100 February 26, 2010

Observation v. Experiment In the Taubes article, what was identified as the type of study that can get us closest to an answer on causation? A) Cohort study B) Case-control study C) Randomized-controlled trial D) Nurses’ Health Study Environmental Studies 100 February 26, 2010

Putting it all together Why does the US Men’s National Soccer team have difficulty winning in Mexico? A) Really, really intimidating crowd B) Playing at altitude (Mexico City: 7349 ft. above sea level) C) Pollution levels in Mexico City D) It’s hot there over the summer E) We are just an inferior soccer nation Environmental Studies 100 February 26, 2010

Putting it all together n In science, we often take a reductionist viewpoint n n Air pollution is not one exposure! n n n Looking for the “smoking gun” Many different compounds, concentrations that co-exist Remember that ozone = sunlight + NOx + VOC Disease = interaction between many factors: n n n Environment Genetics Behavior Environmental Studies 100 February 26, 2010

Conclusion Epidemiology is the study of the distribution and determinants of disease in human populations n Air pollution is implicated in both chronic disease morbidity and mortality n Air pollution epidemiology has several important challenges n n n How do we capture the relevant exposure? For whom do we want to make scientific inference? No one study will give us all the answers: look at the body of research as a whole Environmental Studies 100 February 26, 2010

Thank You Feel free to contact me at: slm 1@uw. edu Environmental Studies 100 February 26, 2010