Health Risk Assessment

1

• Health Risk Assessment
• Tord Kjellstrom
• National Center for Epidemiology and Population
  Health
• Australian National University

2
Three parts to the lecture

• Principles of health risk assessment (HRA)
• Case study of HRA for air pollution in New
  Zealand
• Link between HRA and guidelines/standards setting
  at international level

3
Ultimate aim of health risk assessment (HRA)

• To provide the best possible scientific, social
  and practical information about the health risks,
  so that these can be discussed more broadly and
  the best decisions made as to what to do about
  them
• from Environmental Health Risk Assessment,
  EnHealth Council report, June 2002. Canberra,
  Commonwealth Depatment of Health and Aging
• www. health.gov.au/pubhlth/strateg/envhlth/risk/

4
Different settings for HRA

• Generic HRA for a particular pollutant
• What are the documented health hazards due to CO
  exposure?
• Location/time specific HRA for a particular
  pollutant
• What is the health risk(or impact) caused by this
  pollutant in a particular geographic area during
  a particular time?
• Location/time specific HRA (or HIA) for a
  combination of all pollutants from a particular
  source.
• What is the health impact of vehicle air
  emissions, or what improvement in health could be
  gained from a 25 reduction of driving?

5
Air pollution exposure
Population health effects
Future health impacts
Time
Past EPIDEMIOLOGY Now
HEALTH RISK ASSESSMENT Future
Does air quality pose a health risk here? Until
now? In the future?
6
Four steps in risk assessment

• RISK ASSESSMENT
• 1. Hazard identification
• 2. Dose-response assessment
• 3. Exposure assessment
• combine 1,2,3 into gtgtgt
• 4. Risk characterization
• What is the estimated attributable occurrence of
  the adverse effect in a given population ?

• The answer leads to gtgt
• RISK MANAGEMENT
• Risk evaluation
• Risk perception/communic.
• Exposure control
• risk monitoring

7
EnHealth document, Environmental HRA

• Adds an initial step Issue identification
• What is the concern?
• Why is the concern an issue?
• How was the concern raised?
• Is the issue amenable to risk assessment?
• This is similar to scoping the issues in
  Environmental impact assessment
• Otherwise the same approach is promoted by
  EnHealth as by WHO, USEPA, NSW EPA, as well as
  authorities in the UK, Canada, NZ, etc.

8
Hazard identification

• What types of adverse health effects may be
  caused by the hazard?
• How quickly may the effects be experienced and
  how long would their duration be?
• Toxicological and epidemiological data, review
  reports, generic health risk conclusions
• Information sources WHO Env Health Criteria,
  Australian govt review documents, Internet
  searches, handbooks on specific topics (e.g.
  metals)

9
Dose-response assessment

• Quantifying relationships between exposures and
  health effects
• Dose-response relationship
• Dose-effect relationship
• Public health impact
• Burden of disease and injury

10
Dose-response curves for effects of lead in
children
11
View over Christchurch, winter morning woodsmoke
12
Residual of model without pollutants against the
level of PM10 (Christchurch, whole year model,
average daily temp)
Each point is an average of 20 adjacent points
sorted by PM10
13
Poisson regression result (1988 - 1997)(for
summer only)
Health outcome Total deaths

• Maximum hourly temperature above 27.9 degree C
  had a significant effect on same day total
  mortality but no significant effect on total
  mortality on subsequent days.
• PM10 level had a significant effect on same day
  total mortality and also on total mortality on
  subsequent days. However, PM10 level of the same
  day had the strongest association with total
  daily mortality.
• CO and NO2 levels were not significant when PM10
  was in the model

14
Dose-Effect Relationship
15
Effect of Ozone on Pulmonary Function
16
Dose-response curves, speed and car crash
injuries
17
Interpretation issues in dose-response
relationships

• Short-term exposure and short-term effects Acute
  effects
• Longer term exposure Chronic effects
• Time-scale in air pollution epidemiology studies
• Daily, weekly, monthly, seasonal, annual
• Latency period, Time-lag
• Effect modification by age, or longer exposure
  and higher dose at higher age?
• Effect modification by temperature, weather,
  season
• Short lag in summer, longer lag in winter

18
Age-effect on weekly mortality/10,000 before and
during London Fog 1952
19
Christchurch non-external mortality increase ()
per degree, Rh or 10 ug/m3 lag 0 (1 or 2
similar) Figures in red, plt0.05
20
Christchurch mortality increase () per degree,
or 10 ug/m3 lag 0 (1 or 2 similar)figures in
red, plt0.05
21
Climate temperature and humidity

• Cold damp climate
• frostbite
• respiratory disease
• mold, housemites allergies, asthma
• Increased mortality in people with chronic heart
  and respiratory disease

• Hot humid climate
• Heat stroke, deaths
• Lower work ability
• Skin irritation from excessive sweating

22
Seasonality of hospital admissions, Auckland
23
Exposure assessment

• Air monitoring outdoors, representative levels of
  the whole populations exposure
• Personal monitoring or high density multiple
  monitoring stations outdoor and indoor air, in
  vehicle air
• Atmospheric models based on emissions inventories
• Validation of model comparing monitoring and
  model, sample surveys of individuals with
  personal monitors
• Personal time-location records exposure
  contribution from each microenvironment
• Seasonal variations in exposure sources

24
When and where do people get the highest
exposures?
25
Motor vehicle air pollution in Auckland
26
Overlaid grid cell NO2 concentrations and census
area boundaries in Auckland
27
Circulatory Respiratory Mortality
No. Days NO2 gt 33 µg/m3
28
Risk characterization

• Health risk exposure x dose-response coeff.
• Health impact (case number) risk x population
• Ideally age-specific differences in dose-response
  coeff. Should be taken into account
• Presenting impacts as numbers of cases, potential
  years of life lost, DALYs, etc.
• Calculating health risk with a metabolic model
• Risk evaluation comparing with acceptable
  risk, often purely by comparing predicted
  exposures with air quality standards
• and then Preventive policies and actions

29
Metabolic model for a chemical hazard
Assume intake Calculate level in critical
organ Use dose-response relationship at organ
level to estimate risk of damage Assess health
risk
30
Considering multiple health impacts

• A particular pollutant may cause different types
  of health effects e.g. PM causes increased
  acute mortality and also lung cancer
• A pollutant may be considered as a proxy for a
  combination of pollutants with different effects
  e.g. PM, CO and NO2 almost always occur together
  as pollutants from vehicle or combustion sources
• A pollutant source may be associated with quite
  different effects, that would be prevented if
  certain action is taken to reduce pollution from
  this source e.g. reducing vehicle use leads to
  collateral health gains

31
Estimating deaths from crashes and vehicle air
emissions in Auckland during commuting hours

• 80 deaths/year in car crashes 50 during weekday
  commuting hours 40 deaths/year
• recent study in Europe car air pollution
  contributes to 400 deaths/million people/10 ug/m3
  PM10
• if half of Aucklands million people are exposed
  to 10 ug/m3 from vehicles, and half of the
  pollution is created during weekday commuting
  hours
  .
  100 deaths/year

32
Deaths from physical in-activity among car drivers

• - lack of natural daily physical exercise
  one outcome of in the 300,000 car commuters of
  Auckland
• obesity prevalence in Auckland appr. 20 (60,000
  of car commuters)
• obesity increase among constant car users appr.
  1.2 (appr. 10,000 car-obese among commuters)
• at 0.8 background mortality per year and appr.
  50 increased mortality among obese people, the
  car commuter obesity would cause 0.4 x 10,000
• 
  40 deaths/year

33
Potentially reduced health costs with increased
public transport commuting and reduced
unnecessary car driving in Auckland

• Total road toll from crashes, air pollution and
  obesogenic environment 180 lives/year
• Bringing Auckland to Melbourne level of public
  transport use (10 reduced car commuting)
  possibly saves 18 lives/year
• Bringing Auckland to Stockholm level (25 reduced
  car commuting) possibly saves 45 lives/year

34
Comparing Collateral health gains with
Greenhouse gas gains

• 25 reduction of private car commuting would lead
  to a 20 reduction of CO2 emissions 380,000
  tons
• at US 21/ton CO2 equ.gtgtgt NZ 20
  million/yr
• Value of reduced mortality at 2 million/death
  
• 
  NZ 90 million/yr
• To this should be added other health costs
• Thus, CO2 value alone underestimates economic
  value by at least a factor of 5 ! ( 20 instead
  of 110 million)

35
Case study
  Health effects due to motor vehicle air
pollution in New Zealand.     Report to the
Ministry of Transport  20 March,
2002        G.W. Fisher1, K. A. Rolfe2, Prof. T.
Kjellstrom3, Prof. A. Woodward4, Dr S. Hales4,
Dr A. P. Sturman5, Dr S. Kingham5, J. Petersen1,
R. Shrestha3, D. King1.   1.   NIWA 2.   Kevin
Rolfe Associates Limited 3.   University of
Auckland 4.   Wellington Medical School 5.  
University of Canterbury http//www.transport.gov
t.nz/publications/niwa_report
36
Hazard identification and scoping section
4 The New Zealand Situation 4.1 Scope 4.2 Applicab
ility of overseas research 4.3 Validity of
comparisons between 'health effects' and 'road
toll effects' 4.4 Possible confounding
effects 4.5 Previous studies New Zealand studies
linking air quality and health effects
37
Difference in mortality impact between
non-external deaths and traffic crash deaths
38
Exposure assessment section
5 Air Pollution Exposure 5.1 Scope 5.2 Methodology
5.3 Data sources Measurement methods
Proportion due to vehicles 5.4 Concentration
results 5.5 Discussion Extreme days Natural
sources Seasonal variations Vehicle
proportion 5.6 Exposure results Total NZ
population Regional breakdown
39
Input from emission inventories
40
Summary of national exposuresNumber of people
(thousands) exposed to nine range categories of
annual PM10 exposure
Only people gt 30 years of age only vehicle
related pollution
41
Dose-response and risk characterization section
6 Health Effects 6.1 Scope 6.2 Calculation
methods 6.3 Dose-response relationships The
Künzli study Studies providing the dose-response
relationship for the Künzli study 6.4 Results
42
Basic calculation method Kunzli study
43
Estimated current number of deaths due to PM10
air pollution in NZ
44
Total mortality (all ages) due to road toll and
vehicle emissions
45
Applying in HRA in standards setting and policy
development

• An international example the WHO Air Quality
  Guidelines for Europe, 1999
• www. who.int/docstore/peh/air/guidelines
• Objectives
• to help countries derive their own national air
  quality standards
• technologically feasible consider social
  constraints
• eliminates, or reduces to a minimum, hazards to
  health

46
WHO Air quality guidelines

• Covers
• SPM (PM10, PM2.5) dose-response functions
• lead, gaseous pollutants (CO, NO2, SO2, ozone,
  etc.) maximum exposure concentrations
• Carcinogenic compunds estimates of carcinogenic
  potency (additional lifetime risk from exposure
  to 1 ug/m3)
• Adapting these to national needs requires
• Considering vulnerable populations
• Applying uncertainty factors
• Taking socio-economic issues into account

47
Dose-response functions for PM10 (WHO AQGs)
48
With current modes of economic development comes
the air pollution from vehicles ! How will the
global community deal with this challenge to
sustainability?
Manila, 2001 (PM10 300 ug/m3)