Policies to Improve Air Quality and Health - Jessiman...–De-emphasize extreme metrics? •Long-term vs. peaks B.C. Lung 2017 Air Quality & Health Workshop 0 10 20 30 40 50 60 70

Policies to Improve Air Quality and Health

Summary and Synthesis

B.C. Lung 2017 Air Quality & Health Workshop

• Strength: A small association does not mean that there is not a causal effect

• Consistency: Consistent findings observed by different persons in different places.

• Specificity: Causation is likely if there is a very specific population at a specific site and disease with no other likely explanation.

• Temporality: The effect has to occur after the cause • Biological gradient: Greater exposure should generally lead to greater

incidence of the effect. • Plausibility: A plausible mechanism between cause and effect is helpful • Coherence: Coherence between epidemiological and laboratory findings

increases the likelihood of an effect. • Experiment: Occasionally it is possible to appeal to experimental

evidence. • Analogy: The effect of similar factors may be considered.

Establishing Causality for Risk Assessment (weight of evidence approach)


Establishing Causality for Risk Assessment (weight of evidence approach)

Relationship Description

Causal Relationship Other factors sufficiently ruled out; evidence in controlled human exposure studies; observational studies in which plausible alternatives have been addressed. Many studies.

Likely to be a Causal Relationship Above with important issues remaining e.g., co-pollutants and other factors; no or limited supporting studies; animal toxicology but limited human evidence. Many studies.

Suggestive but not sufficient to infer a Causal Relationship

A well conducted epidemiologic study shows effects but others are inconsistent; an animal toxicology study shows effects, but other evidence limited.

Inadequate to Infer a Causality Relationship

Quality, quantity and consistency are lacking in the database.

Not Likely to be a Causal Relationship Good quality studies at relevant exposure levels and for relevant susceptible populations consistently show no effect.


Adapted from U.S. EPA, ISA Preamble 2015


Air Quality Benefits Assessment Tool (AQBAT)


Benefits of reducing air pollution

Stieb et. al. 2015. Can J Pub Health 106(6)


Benefits of reducing air pollution

Wang et. al. 2017. EHP 125 (3).

Tier 3 Vehicle Regulations

http://www.gazette.gc.ca/rp-pr/p1/2014/2014-09-27/html/reg1-eng.php


Tier 3 Vehicle Regulations: Expected Benefits



Tier 3 Vehicle Regulations: Expected Benefits (millions 2013 dollars)

Monetized Impacts 2014 to 2020 2021 to 2025 2026-2030 Total

Health Benefits 600 2,500 4,100 7,200

Environmental Benefits 10 32 49 90

Total Benefits 610 2,532 4,149 7,200

Fuel Costs 791 131 117 1,038

Vehicle Technology Costs 393 607 587 1,587

Government Costs 2 0 0 2

Total Costs 1,186 738 704 2,628

Net Benefits (costs) (576) 1,793 3,444 4,662



Air Quality Management System (AQMS)

• CAAQS establishment

– PIA: Population Improvement Approach

– Non-threshold (management levels; pop exposure reduction goal)

– AirZones and non-linear CRFs

– De-emphasize extreme metrics?

• Long-term vs. peaks


0

10

20

30

40

50

60

70

80

90

100

pe

rce

nt

of

po

pu

lati

on

exc

ee

din

g co

nce

ntr

atio

n

µg/m3 PM2.5

Population Exposure 3-yr Rolling Annual Average PM2.5

2000-02

0

10

20

30

40

50

60

70

80

90

100

pe

rce

nt

of

po

pu

lati

on

exc

ee

din

g co

nce

ntr

atio

n

µg/m3 PM2.5


2001-03

0

10

20

30

40

50

60

70

80

90

100

pe

rce

nt

of

po

pu

lati

on

exc

ee

din

g co

nce

ntr

atio

n

µg/m3 PM2.5


2002-04

0

10

20

30

40

50

60

70

80

90

100

pe

rce

nt

of

po

pu

lati

on

exc

ee

din

g co

nce

ntr

atio

n

µg/m3 PM2.5


2003-05

0

10

20

30

40

50

60

70

80

90

100

pe

rce

nt

of

po

pu

lati

on

exc

ee

din

g co

nce

ntr

atio

n

µg/m3 PM2.5


2004-06

0

10

20

30

40

50

60

70

80

90

100

pe

rce

nt

of

po

pu

lati

on

exc

ee

din

g co

nce

ntr

atio

n

µg/m3 PM2.5


2005-07

0

10

20

30

40

50

60

70

80

90

100

pe

rce

nt

of

po

pu

lati

on

exc

ee

din

g co

nce

ntr

atio

n

µg/m3 PM2.5


2006-08

0

10

20

30

40

50

60

70

80

90

100

pe

rce

nt

of

po

pu

lati

on

exc

ee

din

g co

nce

ntr

atio

n

µg/m3 PM2.5


2007-09

Air Quality Standards Goal • Improve on this trend • 15% target for population

exposure improvement

CAAQS and Management Levels under the AQMS

Management Level

Management Actions

Air Management Threshold Levels

Ozone (ppb)

PM2.5 Annual (µg/m3)

PM2.5 24-hour (µg/m3)

2015 2020 2015 2020 2015 2020

Red Zone Actions for Achieving Air Zone CAAQS

Concentration 63 62 10 8.8 28 27

Orange Actions for Preventing CAAQS Exceedance

Concentration 56 6.4 19

Yellow Actions for Preventing AQ Deterioration

Concentration 50 4.0 10

Green Actions for Keeping Clean Areas Clean

12

B.C. Lung 2017 Air Quality & Health Workshop 13

Non-linear Concentration-response Functions (CRFs)

Recent Results from Canadian Studies

Fine Particulate Matter Nitrogen Dioxide Ground Level Ozone

Nasari et al., 2016, Air Quality, Atmosphere, and Health

B.C. Lung 2017 Air Quality & Health Workshop 14

Global Burden of Disease Analysis:

Impact of Canadian Cohort Studies

85% of

Canadians

Live below

This

Level

GBD 2015

Uncertain

Effect

Range

GBD 2010

Uncertain

effect

Range

2% live

below this

level


Air Quality Standard Metric

0

10

20

30

40

50

60

70

80

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

pp

b o

zon

e

Trends in Ozone Concentrations: Canada

Warm Season 4th highest 8 hr daily max

Warm Season average 8 hr daily max

Annual Average

What did we see today?

It’s complicated


Arden Pope’s Top Ten

• Led us through the history of air pollution

– Really bad air pollution was really bad

– But not a lot of comfort to be had from reduced air pollution levels, but reduction clearly beneficial

– The effects span an array of effects, probably to the lowest of levels

– The public health benefits are large and meaningful: regulatory impacts


Doug Dockery: how do you do that?

• You’d like to access Randomized Control Trials

– But the complications of real life limit this

• Natural Experiments are attractive

– But they’re rare and limited in time

• Evaluation of Policy

– Leads to long-term decline in air pollution

– But its complicated (life that is).

• You get your evidence where you can

– Combine with the basic knowledge of effects


Corwin Zigler: how could you do it?

• Described the current regulatory impact process

– The steps from emissions reductions to health benefits

• The counterfactual!

– The Base Case in our terms

• There is an additional approach that can provide extra perspectives on success – Causal Inference; how to build into a rigid regulatory system?


Marine Fuel Regulation and Air Quality

• Relatively ideal situation

– Little long-range transport; specific signal (SO2)

• Increasing ship population and high S fuel sparks concerns and speaks to need for regulation.

• Emission Control Area (ECA)

– S focus affects all ships immediately

– NOx affects only new engines (turnover overest.)

• Emissions ↓ leads to [conc] ↓ pop. exp ↓

• Halifax Harbour study; Arctic ECA consideration


Woodstove Intervention Study: Launceston

• Very large wintertime signal; uniquely woodburning

• Majority of homes used wood; replacement over time with electricity+

• Control population appropriate

• Incorporated aspects building on previous intervention studies

• Comprehensive intervention: regulation, education, incentives

• Large air pollution decline (≈ 40%): successful intervention

• Large if imprecise health signal improvement

• Utility of local information (same in Canada)


Buckley Valley & Lakes Woodstove Exchange

• Analagous to Launceston: prevalent woodstove use – Small communities, prevalence of wood heating, alternatives difficult

• Woodsmoke can dominate community air quality – Even large (Montreal: 30% of winter PM)

– Here apparent up to almost half the year

• Exchange program inspired by earlier (Libby)

• Detailed monitoring study: spatially and chemically

• About 10% of 4200 non-EPA stoves replaced?

• Overall decline in PM measures, but rebounding?

– Black market? Virtuous discontinuity? Weather related?

– Lack of response in biomarkers broadly consistent with air quality


Arden Pope: the business of air pollution

• For air pollution (and other risks) there is a Tension between improvement and costs (who pays?)

• Benefit Cost Assessment: its what we do.

• Other approaches (Pigovian) (Cap and Trade).

• Form can inform: but its not intuitive (supralinear)

• By any measure, it would be hard to go too low with air pollution


London Congestion Study

• Large urban agglomeration with complex air quality

• Air pollution measures associated with some measures of children’s respiratory health

• LEZ expected to produce some significant improvements in local air quality but did not do so

– Diesel fleet increase; failure of diesel regulation; other

– Not surprisingly, improvements in health were not found

• Did the LEZ prevent a deterioration of air quality?


California Children’s Health Study • Characterized by shear ambition

– Recruited and characterized large study population

– Follow-up sustained

• Initial findings of strong association of lung health parameters with air pollution

• Moving to lower air pollution improved lung health

– so improved air quality = !!!

• Air pollution was improving over a long and sustained period

– Concomitant and substantial improvement in lung health

• Dedicated study; D-R; signif. ↓ a.p.; long-term follow-up


AirCare and Mortality Avoidance

• Aircare provided substantial NOx reduction over and above ongoing reductions

– Expected to provide benefits based on TRAP conclusions

• Over long time period of program multiple and conflicting trends are occurring

• Can this one intervention’s effects be seen? – Examined from multiple time and statistical angle

– yes


Intervention Studies in Risk

• As for risk assessment, one study does not make the case

• Making the case for air quality improvement by itself is sufficient

• Showing health improvements as well is great but fraught with difficulties

• It’s the body of evidence, drawing from multiple locations and sources that will be convincing


• Where there is big harm there is opportunity to examine the finer details (but its complicated)

• Intervention studies offer a window into the examination of source-associated benefits (ibid)

• Intervention studies are a compliment to risk assessment findings that direct management (ibid)

• Intervention studies can provide some important clues in a multi-pollutant environment.

• As air pollution has decreased, the resistance to continuing the effort has increased

– All lines of evidence are needed to support and refine what we do on air pollution

Policies to Improve Air Quality and Health - Jessiman...–De-emphasize extreme metrics? •Long-term vs. peaks B.C. Lung 2017 Air Quality & Health Workshop 0 10 20 30 40 50 60 70

Documents