Research on Air Pollution and Socioeconomic Conditions: Status and Future

1
Research on Air Pollution and Socioeconomic
Conditions Status and Future

• Marie ONeill (Boston, MA, USA)
• Alfesio Braga (Sao Paulo, Brazil)
• Luis Cifuentes (Santiago, Chile)
• September 23, 2003 Perth, Australia

2
Outline of our Remarks

• Problem
• Methodologies
• Results
• Goals and guiding questions

3
THE PROBLEM
Why are we interested in studying air pollution
and socioeconomic conditions?

• Widespread exposure
• Serious and costly consequences
• Moral dimension
• Policy focus, funding opportunities

4
Widespread Exposures
THE PROBLEM

• Socioeconomic Conditions
• 3 billion live in poverty (World Bank, 2003),
  mostly
• in developing countries
• Socioeconomic gradients (inequalities) exist
• everywhere
• Air Pollution
• 1.5 billion live in polluted cities (WHO, 2000)
• By 2025, 65 of worlds people to live in cities
  (Ibid)
• 40 of kids in developing world cities (Davis
  Saldiva, 1999)
• High indoor and outdoor exposures from biomass
  fuel (Smith et al, 2000 Arbex et al, 2000)

5
Serious and Costly Consequences
THE PROBLEM

• Socioeconomic inequalities (individual and
  societal)
• Gradients observed for
• infant/adult mortality
• infectious/chronic disease
• psychiatric disorders
• (Haan et al, 1987 Krieger et al, 1997
• Marmot, 2001, Kawachi et al 1997)

6
Serious and Costly Consequences
THE PROBLEM

• Air pollution
• Increased mortality, hospital admissions
• Reduced life expectancy, asthma
• Adverse birth outcomes (fetal deaths, defects,
  prematurity, low birth weight, etc)
• Many effects appear to be without threshold
• (Pereira et al, 1998 Brunekreef and Holgate,
  2002 Clancy et al, 2002 McConnell et al, 2002
  Ritz et al, 2002 Schwartz Zanobetti, 2000)

7
Moral Dimensions
THE PROBLEM

• Environmental justice concerns
• Right to clean environment
• Higher exposures documented among lower
  socioeconomic groups (Bullard and Wright, 1993)
• Environmental exposures may partly explain
  observed health outcome gradients by SEP (Evans
  and Kantrowitz, 2002)
• But, even if exposures are NOT higher, special
  responsiblility to protect people already
  disadvantaged

8
Policy Focus, Demanding Actions
THE PROBLEM

• Interaction becoming a policy concern
• Air pollution and poverty both priority areas for
  public health intervention
• Ezzati M, Lopez AD, Rodgers A, Vander Hoorn S,
  Murray CJ, Comparative Risk Assessment
  Collaborating Group. Selected major risk factors
  and global and regional burden of
  disease.comment. Lancet. 20023601347-60.

9
Other reasons

• Funding sources (e.g., NIH, HEI) interested in
  social disparities research

10
Methodologies
11
Socioeconomic factors in air pollution
epidemiology

• Can be associated with both air pollution
  exposure and health, and confound effect
  estimates in cross-sectional/cohort studies
• Air pollution exposures may affect people
  diffferently depending on socioeconomic position
  (effect modifiers)

12
Confounder
A third factor (C) associated with exposure
(E) and with disease (D), independent of
exposure If not controlled for in analysis,
can result in biased estimate of true association
Observed truth


13
Effect modifier

• Third factor for which, within different levels,
  the association between exposure and disease
  differs
• AGE O3 effect on pulmonary function could be
  among children than adults
• A factor can be both a confounder and an effect
  modifier, either, or neither
• Our main interest is effect modifiers

14
Socioeconomic Level and Air Pollution Exposure

• Exposures can be distributed differentially by
  socioeconomic position
• proximity to roadways
• particle/NO2 exposure,
• O3 exposure (scavenging)
• settlement patterns
• center city vs. Suburbs
• topography/meteorology
• indoor and occupational exposures
• Literature is sparse location dependent

15
Socioeconomic Level and Air Pollution Exposure in
São Paulo, Brazil
16

• South, East and North borders
• Air Pollution distribution

17
Districts X Origin-Destination Zones
18
Socioeconomic level and air pollution exposure in
Chile
19
PM2.5 and educational level are correlated in
Santiago, Chile

• PM2.5 in 1993

Mean Educational level by municipality
20
Hypothesis
Being in lower socioeconomic position is related
to and/or confers increased susceptibility to air
pollution health effects
21
Terminology

• Socioeconomic position
• Rank of a persons relative social and economic
  level in a stratified society
• Other terms
• Socioeconomic status (Weber)
• Social class (Marx)
• Deprivation (commonly used in U.K.)

22
Distribution of susceptibility factors may differ
by socio-economic position

• Medical conditions (diabetes, CVD, asthma)
• Age structure of population
• Life choices smoking, diet, exercise
• Life course experiences hygiene hypothesis in
  asthma
• early infection and exposure to endotoxins is
  protective

23
Low socioeconomic level may confer susceptibility

• Less nutritious diet (fewer anti-oxidants)
• More infections (crowding, sanitation)
• Higher exposure of environmental pollutants from
  multiple sources
• Occupation
• Location of residence
• Transportation habits
• Indoor pollution sources/activities

24
Potential Pathways for Socioeconomic Position to
Increase Susceptibility Exposure
Low Educational Attainment
Low wage job (potential occupational exposures)
Residence in low-income neighborhood

• Higher levels of indoor pollutants (e.g., NO2
  from gas stove)
• Proximity to outdoor pollutant sources (e.g.
  diesel exhaust)
• Few nearby supermarkets wtih fresh
  fruits/vegetables
• Crowded living conditions (higher infection
  level)
• Low quality health care (access limited,
  prevention lacking)
• Violence/insecurity (in home or neighborhood)
  leading to stress

25
Measuring socioeconomic position

• Individual level
• education, occupation, income
• Population/geographic level
• Composition ( people w/ college degree)
• Contextual features of the community or society
  (parks, availability of food stores, crime rate,
  social support networks)

26
Individual level indicator
50 year old males with Ph.D.
Small health effect
Air pollution exposure
50 year old males with 8th grade education
Big health effect
27
Geographic/population level (compositional)
50 year old males living in county where 90 of
people have 12 years education
Small health effect
Air pollution exposure
50 year old males living in county where 40 of
people have 12 years education
Big health effect
28
Studying group and individual socioeconomic level
together

• Could evaluate relative importance of
  individuals education and the people in
  persons neighborhood/county with a given
  education level, on air pollution effect
  (multilevel model)
• International studies have shown negative effect
  of income inequality in society on health,
  independently of individuals income

29
RESULTS
30
ResultsEffect modification by socioeconomic
position A few illustrative examples

• Health outcome mortality
• Exposure Airborne particles
• Geographic Brazil, United States (U.S.)
• Individual (education) Netherlands, U.S.

31
Geographic effect modification

• In press (JECH) study
• Sao Paulo, Brazil (Martins et al)
• Time series PM10 and daily mortality
• Hypothesis PM10/mortality associations will be
  higher in districts or counties ranked lower by
  geographic socioeconomic variables

32
Socieconomic Variables (Census data and
Origin-Destination Survey in area)

• São Paulo, Brazil
• education
• income
• percent people in slums

33
Modeling approach

• Exposure Daily PM10 from city monitors
• Poisson regressions, outcome is daily deaths at
  each region
• Control for long-term mortality trends with
  nonlinear terms for day of study, temperature,
  other mortality predictors
• logE(daily deaths) a (b1PM10)
  (b2ns(time)) ...)
• exp (10b1) mortality rate ratio (RR)
  associated with of 10 mg/m3 PM10, controlling
  for other predictors

34
Results

• Higher associations seen in
  districts with
• lower income, education
  and more slums

35
(No Transcript)
36
Results from 20 U.S. cities

• No substantial differences in the
  PM10/mortality association by socio-economic
  indicators measured at county level ( poverty,
  attained education, others)
• Samet JM, Dominici F, Curriero FC, Coursac
  I, Zeger SL. Fine particulate air pollution and
  mortality in 20 U.S. cities, 1987-1994. New
  England Journal of Medicine 20003431742-9

37
Concerns regarding geographic indicators

• Resolution
• Size of counties or districts may be too great
  heterogeneity within them
• Interpretation
• Contextual/compositional variables do not
  measure individual characteristics (ecologic
  fallacy)
• Not lesser than individual-level variables,
  such as medical conditions but of intrinsic
  interest as the effect of neighborhood or
  community characteristics on health

38
Individual level effect modification Education

• Two cohort studies of air pollution and mortality
• Traffic pollution in the Netherlands
• PM10 in the United States

39
The Netherlands

• 5,000 elderly people 1986-1994
• Long term black smoke (BS) exposure
• Mortality rate ratios and 95 confidence intervals

• Hoek G, Brunekreef B, Goldbohm S, Fischer P,
  van den Brandt PA. Association between mortality
  and indicators of traffic-related air pollution
  in the Netherlands a cohort study. The Lancet
  2002360 1203-09

40
United States

• American Cancer Society 500,000 adults,
  1982-1998
• Education level Less than high school (HS), HS,
  HS
• All-cause, lung cancer, and cardiopulmonary
  mortality
• Relative risks expressed per 24.5 ug/m3 in
  fine particles
• Health Effects Institute. Reanalysis of the
  Harvard Six Cities Study and the American Cancer
  Society Study of Particulate Air Pollution and
  Mortality (A special report of the Institute's
  Particle Epidemiology Reanalysis Project).
  Cambridge, MA Health Effects Institute, 2000.

41
(No Transcript)
42
Concerns regarding individual indicators (esp.,
education)

• May be recorded inaccurately
• May not reflect socioeconomic situation
• e.g., person with no public health insurance may
  pay private insurance
• People without traditional education can have
  high socioeconomic level
• Example Bill Gates (Harvard dropout)

43
Conclusions

• Growing recognition that socioeconomic and
  environmental conditions both affect human health
• Understanding their interaction requires
  increasingly sophisticated methods
• Collaboration with social epidemiologists,
  sociologists and others will enhance progress

44
Goals for workshop participants

• Better understand methodological approaches and
  challenges
• Learn more about combining social and
  environmental epidemiology methods
• Develop new ideas for research proposals and
  funding opportunities
• Meet potential collaborators and make new friends!

45
Guiding Questions for Afternoon

• Health outcomes
• Metholodogies
• Geographical Areas
• Pollutant sources and pollutants
• Interest by communities/policy makers
• New questions? (audience generated)

46
Afternoon format

• 4 working groups of 25 people
• 4 Moderators and 4 Scribes
• Maybe a Rapporteur?
• Goal Each group addresses 2-3 of the guiding
  questions
• Produce statements on what we know, gaps,
  recommendations, comments

47
Expected Workshop Outcomes

• Workshop summary report
• Identify follow-on activities
• Working groups
• Research proposals
• Funding opportunities
• Research agenda with priorities