Docstoc

Indoor Air Pollution

Document Sample
Indoor Air Pollution Powered By Docstoc
					Occupational and Indoor Air
  Exposure Assessment
    Mike Flynn, Sc.D., C.I.H.
        Dr. Joel Schwartz
  Harvard School of Public Health
• “This isn’t just pure science – it’s science
  for public policy. If you’re not willing to rub
  people’s noses in the fact that you’ve
  identified a problem, and that something
  needs to be done about it, then have you
  improved public health? You need to be
  willing to take the next step – and to take a
  lot of criticism of your findings from people
  who have a monetary interest in their not
  being true.”
              The flip side
• If you rant and rave about a statistically
  significant finding that indicates a public
  health problem when in fact its trivial or
  non-existent you may have done more
  harm than good to the goal of improving
  public health.
• Correlation is not causality; statistical
  significance is not enough for conclusive,
  effective action, beware of bias and
  respect the unknown.
                The Issues
• To what extent do airborne pollutants within
  indoor environments (home, school, and work)
  represent risk factors for significant health
  impairment?
• How do we differentiate an observed association
  between exposure and disease from exposure
  as a cause of the disease?
• When should we act vs. require additional
  study? What are effective control interventions?
Causal Inference – establishing
exposure as the “cause” of the
            disease
         Scientific Inference
• Deductive
  – Conclusions follow necessarily from premises
  – Logical, Mathematical
  – Physics – first principles
• Inductive
  – Reasoning from the specific to the general
  – Observational studies - epidemiology
  – Statistical evidence
      References for Causality
•   Hill, A.B. (1965).The environment and disease:
    association or causation? Proc. Roy. Soc.
    Med. 58, 295-300.
•   Stellman S.D. (2003). Issues of causality in the
    history of occupational epidemiology. History
    of epidemiology.
•   Rothman K & Greenland, S (2005). Causation
    and Causal Inference in Epidemiology. AJPH
    (95) Sup 1, S144-S150
  Hill on the definition of causation and the
         goals of preventive medicine
• “….the decisive question is whether the
  frequency of the undesirable event B will be
  influenced by a change in the environmental
  feature A. How such a change exerts that
  influence may call for a great deal of research.
  However, before deducing causation and taking
  action we shall not invariably have to sit around
  awaiting the results of that research. The whole
  chain may have to be unraveled or a few links
  may suffice.”
       A.B. Hill - perspective
• Somewhat of disdain for p-values as they
  cannot account for non-random errors
• A practical approach to data analysis
• “All scientific work is incomplete – whether
  it be observational or experimental…. That
  does not confer upon us a freedom to
  ignore the knowledge we already have, or
  to postpone the action that it appears to
  demand at a given time.”
        A.B. Hill - Causality
• Strength of association (e.g., Relative
  Risk)
• Consistency (Effect observed repeatedly
  by others)
• Specificity (mesothelioma – asbestos; VC
  and angiosarcoma)
• Temporality (does the cause (exposure)
  precede the effect (disease))
• Biological Gradient e.g., Dose Response
    A.B. Hill - Causality cont’d
• Plausibility (consistent with current
  understanding e.g. mechanistic models -
  PBPK)
• Coherence ( not in conflict with “generally
  known facts”)
• Experiment (Successful Interventions e.g.,
  LEV in granite industry and TB)
• Analogy (previous similar patterns – e.g.,
  ceramic fibers and asbestos?)
               Annual TB Mortality
From Ventilation for Control of the Work Environment by Burgess et al
  TB Deaths for various exposures
From Ventilation for Control of the Work Environment by Burgess et al
A - pneumatic tools; B-Surface machines: C-plant dust; D-Ventilation
    Effect of Ventilation dust levels
From Ventilation for Control of the Work Environment by Burgess et al
  General Background
For Indoor Air Exposures
              Basic Principles:
             Exposure and Risk
• Risk is often assumed to be approximately
  proportional to exposure (Haber’s Rule).
• Acute effects (e.g. ammonia and irritation, CO
  and death) = short term exposure limits
• Chronic outcomes (asbestos and cancer, metals
  and neurological disease), = longer time limits
• Exposure is the product of time and
  concentration. It is quantified as a time weighted
  average (TWA) concentration:
Time Weighted Average
Concentration (Exposure)

                         N


          1
             T           C t    i i
 C twa     C bz dt    i 1
                            N

                          t
          T 0
                                 i
                          i 1
Consequences of the Definition of
Exposure – the importance of time!
• We spend 80-90% of our time indoors!
• About 25% of our time at work (often indoors)
• Thus indoor and workplace airborne
  concentrations may often represent a major part
  of the dose received from inhalation even if the
  concentrations are lower.
• Note indoor air is directly influenced by outdoor
  levels (and in some cases vice-versa) so a
  decoupling is not always possible
Concentration = Generation Rate
divided by the effective ventilation


                G
            C 
                Q
    Other factors affecting risk &
             exposure
• Interaction with & Proximity to sources of
  pollutants
• Individual Susceptibility
  – Health Status
  – Age (Children, Fetus)
  – Sex
  – Genetic makeup
Historical Background
         Historical Perspectives
•   400BC    Hippocrates- Pb toxicity in miners
•   100AD    Galen- toxic acid mists in Cu miners
•   1556     G. Agricola - De Re Metallica
•   1500’s   Paracelsus- Hg Poisoning in smelters
•   1700     Ramazzini - De Morbis Artificum –
             silicosis, Occup. as a risk factor!
• 1788       P. Pott - Cancer in Chimney Sweeps
• 1833       The English Factory Acts – first
             workers comp, & Factory Inspectorate
    Historical Perspectives, cont’d
• 1900     Alice Hamilton - Exploring the
           Dangerous Trades (1943).
•   1910   US B.O.M & PHS - Steel & mining
•   1913   NY and MI- first state IH Depts.
•   1936   Walsh Healy Act
•   1969   Coal Mine Health and Safety Act
 Historical Perspectives, cont’d
• 1970    Williams-Steiger Act – OSHA
         and NIOSH created;
• 1970    EPA created
• 1973    Oil Embargo
• 1978    First International Indoor Climate
          Symposium
• 1980    Supreme Court: Benzene Case
          (Risk Assessment is born)
• 1980    Supreme Court: Cotton Dust Case
          (Cost-Benefit Analysis rejected)
  Primary Regulatory Authority for
    Indoor Air Pollution - OSHA
• OSHA – PEL’s for the workplace only
• EPA – Some back-door authority via
  Clean Air Act and TSCA e.g., controlling
  VOC content in some products
• Federal and State Laws Regulating E.T.S.
  and Pb, Hg, As in some products
• Other than the workplace there appears to
  be little regulatory authority for Indoor Air
       The major force for OSHA
       legislation = Labor Unions
• It was only through the efforts of organized labor
  that the 1970 OSH Act was passed. This law
  had a profound influence by creating not only an
  enforcement agency (with ability to promulgate
  administrative laws) but also research and
  education capabilities via NIOSH.
• The growth of industrial hygiene, occupational
  medicine and nursing, epidemiology and safety
  and ergonomics accelerated dramatically!
                             AIHA Membership

          14,000
          12,000
          10,000
Members




           8,000
                                                                 Series1
           6,000
           4,000
           2,000
               0
               1920   1940    1960          1980   2000   2020
                                     Year
    OSHA Permissible Exposure
         Levels = PELS
• 29 CFR 1910.1000 Tables Z1; Z2 and Z3
  = TWA’s for 8hrs, Ceilings for 15 minutes,
  various peaks
• Special Standards – Lead, Cotton Dust
• Personal Sampling –Worst case
  (compliance) sampling not random
• Legal Recognition of Engineering and
  Administrative Controls over PPE
• Cancer standards e.g. Vinyl Chloride
Philosophy of Control Interventions
• Hierarchy of Controls
  – Engineering
  – Administrative
  – Personal Protection
  – Work Practices
• ALARA
  Usually for radiation
 ACGIH Threshold Limit Values =
            TLV’s
• Recommendation’s only - they lack the
  force of law
• Extensive documentation; under constant
  revision
• Published annually - TLV booklet
• Castleman and Ziem, Rappaport and
  Roach – undo influence by industry in the
  setting of the TLV’s
 NIOSH Recommended Exposure
        Level’s REL’s
• Also lacks the force of law
• NIOSH Pocket Guide to Chemical
  Hazards
• IDLH Levels
• HHE’s = Health Hazard Evaluations
• Criteria Documents
• Control Technology Assessments
• http://www.cdc.gov/niosh
    A comparison of EPA and OSHA
     Particulate Material Standards
• PM2.5 and PM10 24hr TWA values are 35
  and 150 μg/m3 respectively.
• OSHA’s respirable nuisance dust 8 hr PEL
  is 5 mg/m3 or 5000 μg/m3 !
• If you worked 8hrs at the OSHA PEL and
  had no other exposure then your 24hr
  TWA would be 1666 μg/m3 or over 40
  times the PM2.5 level or 11 times PM10!
Major Legal Cases
    The Benzene Case (1980)
• AFL-CIO vs. API
• The Court upheld the 5th Circuits decision
  to vacate the standard that lowered the
  Benzene PEL from 10ppm to 1ppm
• The Court rejected the notion of regulating
  carcinogens to the lowest level feasible. It
  required that a “significant risk of material
  impairment” be demonstrated (quantitative
  risk assessment)
    The Benzene Case cont’d
• In addition it required OSHA to show that
  the proposed standard (lower PEL) would
  reduce the risk
• The issue of cost-benefit was not decided
  in this case but deferred to the Cotton
  Dust case
Magnitude of Occupational Disease
   (injuries not included here)
• 49,000 deaths estimated 1997 (Steenland
  et al Am.J.Ind.Med. 43:461-482 (2003)
• Estimated to be the 8th leading cause of
  death ahead of motor vehicle deaths, but
  less than diabetes
• Respiratory disease, cancer,
  cardiovascular disease, chronic renal
  failure and hepatitis were the major
  contributors.
Asbestos Deaths
Looking at effects beyond mortality
              (WHO)
• The DALY – disability adjusted life year
• Includes years of life lost YLL plus years
  lost to disability YLD
• DALY=YLL + YLD
• YLL=N*Le; N=# of Deaths; Le = life
  expectancy at age of death
• YLD=I*DW*Ld; I=# of cases; DW =
  weighting; Ld=average duration till death
  or remission
  WHO 2004 ESTIMATED DALY's
   HIGH INCOME COUNTRIES


      6%
           9%
                        Communicable, Maternal,
                        Nutritional
                        Injuries

                        Non Communicable

85%
WHO 2004 ESTIMATED DALYs        Communicable, Maternal,
                                Nutritional
 HIGH INCOME COUNTRIES
                                Injuries
          8%   6%
     4%              9%         Malignant Neoplasms
 4%
                                Neuropsychiatric
6%
                          15%
                                Cardio vasc
8%
                                Sense organs


 15%                            Respiratory
                    25%
                                Musculo-skeletal

                                Digestive
WHO 2004 DALY's Neuropsychiatric Conditions

                                       Depression
          18%                          Alzheimers
                       32%             Alcohol
    2%
                                       Drug use
    5%
                                       Bipolar
    5%                                 Schizophrenia
     5%                                Migraine
                      14%              Parkinson's
         6%     13%
                                       Other
       What fraction of the
 Neuropsychiatric conditions have
     Environmental causes
• Metals: Lead, Manganese, Mercury,
• Pesticides -
• Solvents –

• Further research is needed here as a
  small attributable fraction could make a
  big difference -
          NIOSH-NORA
    Research items for Indoor Air
• Improved assessment of the economic consequences of
  unhealthful indoor environments
• Attention to reducing building-related respiratory
  infections
• Improved methods for measuring exposures
• Development of biomarkers indicating exposure and
  health effects
• Attention to mixed exposures (both biological and
  chemical) in the indoor environment in relation to
  adverse health effects
• Improved understanding of the effectiveness of indoor
  environment interventions in reducing exposures and
  alleviating associated health effects
Exposure Assessment

     ENVR-890-03
What are the questions we seek to
        answer and why?
• We want to assess exposures because we
  believe they represent a risk for disease
  – What are “safe” levels?
  – Do exposures exceed standards?
    (compliance)
  – Do controls reduce levels? (quasi-
    compliance)
  – Law suits – is an individuals disease the result
    of a given exposure (job)?
   What are the “safe” levels?
• Studies to assess long-term exposures
  and correlate with biomarkers of effect or
  disease/symptom rates (Hill’s dose
  response)
• Population based epidemiological studies
  with exposure assessment component
   Compliance: Enforcement and
            Controls
• Estimate worst-case and sample that –
  inference based on this being ok or not
• Not a random sampling approach
• But an effective use of resources
• Concept of action level i.e. ½ the TLV for
  intervention
                  Legal
• Reconstruction (documentation) of an
  individual’s exposure
• Knowledge of population studies is
  relevant, but we now have more specific
  info on the person, (e.g., health status,
  genetics, smoking history, etc.)
Assessing long term exposures
• For many of the major diseases of concern
  (e.g, cancer, neuro-degenerative,
  cardiovascular, COPD, etc) we want multi-
  year or lifetime exposures to correlate with
  disease.
• Practical realities make it necessary for us
  to infer these chronic exposures from a set
  of finite shorter-term exposures
         The basic problem
• Given a set of short term (lets say 8-hr
  TWA) measured exposures on an
  individual; estimate a longer term (lets say
  month or year long TWA)
• If the short term samples completely cover
  the long term interval there is no problem
  we get the answer as a single number,
  difficulties arise when only a small portion
  of the interval is covered (the usual case)
  The mathematical statement
Definition                                N

               1       T
                                          C t    i i
    CTWA                 C dt          i 1
                                             N

                                           t
               T     0
                                                  i
                                           i 1
               N
    where     t
              i 1
                       i   T

             1               ti
    and Ci 
             ti           0
                                  Ci dt
The real – world approximation
• sampling        N

               Ct           i i
     CTWA      i 1
                   N

                     t
                     i 1
                             i

              N
     where    t
              i 1
                      i      T
                Question
• How good can the real world
  approximation be? How should a sampling
  strategy be devised to estimate the long
  term number based on sparse data?
• The problem is one within the domain of
  time series i.e., the analysis of stochastic
  processes (exposure); e.g, see reading
  #11)
    The concept of stationarity
• A stationary stochastic process has a
  mean, variance, and covariance that do
  not change in time, i.e., there are no
  trends in the data.
• Most time series analysis hinges on the
  assumption of stationarity.
• Most exposures are unlikely to be
  stationary; particularly long-term ones of
  interest in health studies.
        Campaign Sampling
• The resources required to conduct
  sampling generally are significant and a
  common approach is to collect data over a
  given finite period – a campaign
• What are the consequences of this, or any
  other approach for assessing long-term
  exposures outside of the time period
  sampled?
Basic Sampling Strategy Concepts
• To estimate an individuals exposure:
  – random vs. stratified random sampling in time
  – campaign sampling
• To estimate a group exposure
  – Random individuals
  – Stratified in time
                        Worker 3 exposure to IPA

          100

           80
IPA ppm




           60
                                                        Series1
           40

           20

            0
                0   5        10         15    20   25
                                  Day
                      Worker 3 Exposure to IPA

            500
            400
IPA (ppm)




            300                                  Series1
            200                                  Poly. (Series1)
            100
             0
                  0      50              100
                          Day
            first 3 week   middle 3    stratified


Mean              66.776   262.2493    150.5787
Median             71.83      215.51        114.5
Std. Dev,      22.21495      118.371     106.271
Kurtosis       0.557688     -1.11776   5.866483
Skewness       -0.92276    0.475718    2.242699
Range              78.87      371.82          424
Minimum            15.17      101.24        49.06
Maximum            94.04      473.06      473.06
Count                 15          15            15
LCL            54.47377    196.6977    91.72777
UCL            79.07823     327.801    209.4296
    Worker 3 (20 week TWA)
• CTWA = 126.4 ppm (T=20 weeks)
• It is a single number (no variability)
• ONLY the stratified time sampling 95% CI
  contains the true mean!
             Basic Principles
• Exposures cannot reasonably be extrapolated to
  periods outside the interval sampled based on
  statistical considerations alone
• To avoid bias in estimating the exposure for a
  given interval, stratification in time is necessary
• Exposures correlated with contaminant
  generation rates and ventilation rates or their
  surrogates represents the only scientifically
  defensible way to do this (exposure
  determinants problem and retrospective
  assessments)

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:5
posted:9/13/2011
language:English
pages:62