Transcript Formal Elicitation of Expert Judgment: Issues and Approaches

The Kuwait Oil Fires
John S. Evans, Sc.D.
Harvard School of Public Health
New England Chapter – Society for Risk Analysis
Boston, Massachusetts
28 May 2008
The Fires
Public Health Impacts of Kuwait Oil Fires – 28 May 2008

More Than 700 Fires

First Fires –

Air War ~ 17 January 1991

Ground War ~ 23 February 1991

Liberation ~ 28 February 1991

Last Fire - 6 November 1991

Oil Burned ~ 4 x 106 barrels per day

PM Emissions ~ 3 x 109 kg
The Question


Were there substantial public health impacts of
Iraq’s 1990 invasion and occupation of Kuwait
and the ensuing Gulf War?
If so, how large were the impacts? what were the
causes?
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
The Hazards & Prior Assessment

Mines and Ordnance
Environmental
Contaminants





Smoke from the Fires
Volatile Organic Compounds
from the Oil Lakes
PAHs and Metals from the
Fires, Oil Lakes and Oil Spills
Depleted Uranium from
Military Ordnance
Psychological Trauma
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
5.0
Kuwaitis
4.0
Deaths/1,000 p-yrs

3.0
2.0
Non Kuwaitis
1.0
0.0
1986
1988
1990
1992
1994
1996
1998
2000
Our Team

Risk Assessment & Valuation

Dr. George Gray
Dr. Andrew Wilson
Epidemiology & Medical
Monitoring
Dr. Abdulrahman Al-Muhailan
Prof. Jaafar Behbehani
Dr. David Mage
Louise Hanson
Dr. Josh Cohen
Prof. Jim Hammitt
Prof. Petros Koutrakis
Dr. Alan Eschenroeder
Prof. Douglas Dockery
Martha Fay
Prof. Debbie Bennett

Expert Judgment
Prof. David Christiani
Dr. Elpis Soteriades
Dr. Rosalind Wright
Prof. Roger Cooke
Dr. Jouni Tuomisto
Dr. Andrew Wilson
Oswaldo Morales
Prof. Allan Hill
Marko Tainio

Delft, KTL, INE, Earthtech
Public Health Impacts of Kuwait Oil Fires – 28 May 2008

Kadhema, Temple, KU, KISR
The People
.

About 600,000 Kuwaiti Nationals
100

Very Young Population
80
Raudhatain
Sabriyah
60

Bahra
~ 25% >= 30 Years Old
40
Many Outside of Kuwait During
Occupation
Northing (km)

20
0
-20
Minagish
-40

~50% in Kuwait in mid February
1991
Umm Guidar
-60
Kuwait Airport Met Station
-80
-120

~95% Had Returned by mid
October of 1991
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
Burgan
-100
-80
-60
-40
-20
Easting (km)
.
0
20
40
60
80
The Measurements

Time Series of PM at Mansouria
1800
Ambient PM10 (Measured)
1600
Ambient Average = 300

1400
3
PM concentration (mg/m )
2000
Equipment Had Been Stolen or
Destroyed by Iraq
No Measurements During First
Several Months of Fires
1200

1000
800
600
400
200

Typical Values ~ 300 mg/m3

High Days 1000 to 2000 mg/m3

Reflect ALL SOURCES of PM
0
be
m
ce
De
1
r-9
1
r-9
be
1
r-9
be
to
m
ve
No
Oc
1
r-9
be
em
pt
Se
91
lyJu
1
-9
1
1
-9
ne
Ju
ay
M
9
rilAp
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
PM10 Measurements at Three Sites
from mid-April through December
of 1991
The Model



US DOD Used HYSPLIT Model to
Estimate Exposure of Allied Troops
HSPH Obtained Results through
Freedom of Information Request
Values Vary Across Country and By
Day
Legend
Modeled TSP (ug/m3), Feb-Oct
1.9 - 8.5


Typical Values ~ 10 mg/m3 (in
populated areas of Kuwait)
High Days > 300
areas of Kuwait)
mg/m3
8.5 - 15
15 - 22
22 - 28
28 - 35
35 - 41
(in populated
41 - 48
48 - 54
54 - 61
61 - 68

Model is Poorly Documented
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
0
25
50
Miles
Why Worry About PM?
Time-Series Studies
5
4
3
% change in daily mor tality/10 ug/m
3
incr ease in PM 10

2
Following the London Fog of
December 1952 it was noticed that
~ 4000 excess deaths occurred.
During the period of the fog,
smoke levels reached 4500 mg/m3
and averaged 1600 mg/m3. The
population of London was ~ 10
million at the time.
1

0
-1
-2
- - - - - - - - - W orldwide literature - - - - - - - - - - - - -
Pooled
worldwide
estimate
-3
Public Health Impacts of Kuwait Oil Fires – 28 May 2008

In large cities around the world,
numbers of deaths are higher on
days with higher levels of PM, all
other things being equal.
Such studies were first done in
London and New York, but now
have been repeated in hundreds of
cities.
Why Worry About PM?
Cohort Studies



Mortality rates are higher in cities
with higher levels of PM, all other
things being equal.
The Six Cities study followed the
mortality experience of ~ 8000
adults living in 6 US cities with
different levels of PM for ~15
years.
The ACS study followed the
mortality of ~ 500,000 adults for
~ 7 years living in 151 US
metropolitan areas.
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
Back of the Envelope Risk Assessment

R ~ b C Mo P


Where b is the risk coefficient (0.4% per mg/m3 PM2.5 cohort (ACS) or 0.1%
per mg/m3 PM10 time series)

C is the concentration (300 mg/m3 PM10 or 10 mg/m3 PM2.5 due to fires)

Mo is the background mortality rate (3/1000 p-yr or 9/1000 adult-yr)

P is the population (600,000 total OR 150,000 adults)
“Back of the envelope” screening estimates:

35 deaths (2/10,000 risk among 150,000 adults) – using ACS cohort coefficient

116 deaths – using Six Cities cohort coefficient

0 deaths – if epidemiology does not reflect causal relationships
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
Key Issues

Exposure Level and Pattern -- Kuwait oil fire exposure level and pattern is different than those of
interest for most regulation, with background PM10 levels in Kuwait of 200 or 300 µg/m3, and PM2.5 increment
due to fires averaging 10 µg/m3 with spikes of several hundred µg/m3.


Should time-series or cohort studies be used to estimate risk?
Composition -- Oil fire smoke differs in composition from typical urban aerosols in the U.S. and
Europe.


Should an adjustment be made for differential toxicity?
Age-Structure of Population -- Kuwaiti population is far younger than US or European
populations.


Do relative risk estimates from US/European studies apply directly?
Causation – Epidemiological studies may reflect association rather than causation due to various biases,
failure to control for confounding, etc.
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
The Experts -European Elicitations

Bert Brunekreef, Ph.D.
Professor of Environmental Epidemiology, University of Utrecht, Netherlands

Annette Peters, M.D., Ph.D.
Assistant Professor, GSF National Research Center for Environment and Health, Germany

Nino Kuenzli, M.D., Ph.D.
Assistant Professor, University of Basel, Switzerland (now at USC)

H. Ross Anderson, M.D.
Professor of Medicine, University of London, England

Ken Donaldson, M.D.
Professor of Medicine, University of Edinburgh, Scotland

Juha Pekkanen, M.D., Ph.D.
Head of Environmental Epidemiology, National Public Health Institute (KTL), Finland
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
Rationale for Use of Expert Judgment




Scientific data may not speak for themselves; require
careful interpretation
Scientific data may seem conflicting or inconsistent;
require judgmental synthesis
In the absence of data, assumptions are necessary;
requires judgment about plausibility of assumptions
Choice of and/or construction of models may require
judgments that are beyond the expertise of the risk
assessor
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
The Question


How many deaths among the exposed Kuwaiti National
population are attributable to exposure to smoke from
Kuwait Oil Fires?
TOO HARD !

 Decompose… Warm Up… Start with Easy Questions

 Work Toward the REAL QUESTION.
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
Mortality Impact of Permanent Decrease
in PM2.5 of 1 µg/m3 in the US
Question
1
Setting
US
Exposure
(Effect Interval)
Long-term
Change
1 μg/m3
Pollutant
PM2.5
Composition
Ambient
Baseline
18 ug/m3
What is your estimate of the true, but unknown, percent change in the total annual, non-accidental mortality
rate in the adult U.S. population resulting from a permanent 1 μg/m3 reduction in long-term annual average
PM2.5 (from a population-weighted baseline concentration of 18 μg/m3) throughout the U.S.? To express the
uncertainty associated with the concentration-response relationship, please provide the 5th, 25th, 50th, 75th, and
95th percentiles of your estimate.
5% :____________ 25%:____________ 50% :____________ 75%:____________ 95%:____________
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
Linking Evidence & Answers

What evidence or theory suggests large values for this relationship?




What is the highest plausible value?
Tell us a little about your reasoning, the evidence, and theories that lead you to
this value.
Can you tell us of scenarios that would yield higher results?
What evidence or theory suggests small values?



What is the lowest plausible value?
Tell us a little about your reasoning, the evidence, and theories that lead you to
this value.
Can you tell us of scenarios that would yield lower results?
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
Mortality Impact of Permanent Decrease
6
Mortality Impact
3
)]PM
2.5e
% c[%
hachange
nge ininbbaseline
aseline mortality
mortalitper
y punit
er 1of exposure
g m3 an(1nµg/m
ual avPM
erag
2. 5
in PM2.5 of 1 µg/m3 in US (Green) or Europe (Blue)
5
4
3
2
1
0
US EU
Expert A
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
US EU
Expert B
US EU
Expert C
US EU
Expert D
US EU
Expert E
US EU
Expert F
The Sequence
Exposure
Duration
Effect
Window
US
(Baseline:
18 µg/m3)
Permanent
Long-term
Q1
One day
1 week
Q3
One day
3 months
Q6
Timing of Expression of Impact
Q7 & Q8
Differential Toxicity
Q9 & Q10
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
MCMA
(Baseline:
35 µg/m3)
Europe
(Baseline:
20 µg/m3)
Q2
Q4
Q5
Pre-Elicitation “Warm Up” Questions

What are the key properties of an ideal epidemiology
study for measuring long-term mortality impacts of PM
exposure?

Similarly, what are the key properties of an ideal
epidemiological study of short-term mortality impacts
of PM exposure?

What factors need to be considered to decide whether
epidemiology results should be viewed as causal?
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
The Answer
Deaths Attributable to Exposure to Smoke from the Kuwait Oil Fires
800
Deaths
600
400
200
0
Expert A
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
Expert B
Expert C
Expert E
Expert F
The Answer
Deaths Attributable to Exposure to Smoke from the Kuwait Oil Fires
Expert *
5%
50%
95%
Approach
E
6
13
21
TS
A
4
32
63
TS
C
<1
54
426
C
F
37
110
210
C
B
16
164
872
C
D
575
2874
11496
C
•Experts are listed in order of their median estimate of risk.
•Letter identifications are randomly assigned to experts.
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
CalPUFF Model

Harvard Used CalPUFF Model to Estimate
Exposure of Kuwaiti Citizens


.
Period Average Contour Plot for All Sources
February 10 - October 15, 1991
Computed Plume Rise From
Thermodynamics of Oil Fires
100
80
Used Improved Meteorological Data and
Finer Grid Spacing
60
40

Accounted for Coastal Effects and
Considered Plume Enhancement
20
0



Values Vary Across Country and By Day
-20
Typical Values ~ 40-50 mg/m3 (in populated
areas of Kuwait)
-40
-60
High Days 600 - 800 mg/m3 (in populated
areas of Kuwait)
-80
-120
.
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
-100
-80
-60
-40
-20
0
20
40
60
80
The Epidemiology

Survival curves after liberation by location
1.0

Out of Kuwait entire
time
Probability of Survival
0.9
In Kuwait entire
time
In and Out of
Kuwait
0.8

0.7
0.6
2/27/91

2/26/93
2/27/95
2/26/97
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
2/27/99
2/26/01
2/27/03
HSPH conducted a cohort study of
5000 Kuwaitis older than 50 on the eve
of Iraq’s invasion.
Information was gathered by
questionnaire on health status,
smoking, education, income, diet,
exercise, location during the fires and
exposure to trauma.
The survival of those who were in
Kuwait during the occupation was 2030 % lower than the survival of those
who were outside of Kuwait during
this same period.
This difference was statistically
significant and persisted after control
for age, smoking, income and
education.
Post Traumatic Stress Disorder



The Al-Riggae Center conducted a
study of ~2000 Kuwaitis in 1993
and a follow-up study of these
same individuals in 1998.
Information was gathered by
questionnaire on sociodemographics, exposure to trauma
and symptoms of anxiety,
depression and PTSD.
In 1993 among adults rates of
PTSD were higher among those
who had remained in Kuwait
during the occupation than among
those who were outside of Kuwait.
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
Prevalence of PTSD
By Location During Occupation
Always Out In and Out
Always In
14.4%
20.5%
25.2%
(28/194)
(168/819)
(234/927)
The Role of Trauma

2.5
Hazard Ratio
2.0
1.5

1.0
0.5
0.0
In Kuwait
In Hiding
Witnessed Attacked or
Violence
Arrested

2.0
Hazard Ratio
Top figure shows mortality rate of
participants in PHS by self-reported
exposure to trauma.
1.5
1.0

0.5
0.0
In Kuwait
In Hiding
Witnessed Attacked or
Violence
Arrested
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
Bottom figure shows coronary heart
disease morbidity rates of
participants in PHS by self-reported
exposure to trauma.
The elevations among individuals
attacked or arrested are statistically
significant and those among individuals
who witnessed violence are on the
margin of significance.
Both graphs present results after
control for age, smoking and
socioeconomic status.
Kuwait’s Public Health Claim
Deaths Attributable to Oil Fires


PAAC’s approach was to limit its claims to those which
were well supported by science and to be entirely open
in its presentation of the state of the relevant science.
On the order of 35 deaths among Kuwaitis were due to
exposure to smoke from the oil fires (based on US
DOD model and conventional regulatory risk
assessment).


This claim is likely to be quite conservative (i.e., understated) in view of the new
air pollution modeling results and the findings of the expert judgment study.
The social cost to Kuwait of these 35 deaths is ~ 200 million US$ (based on a
contingent valuation study done in Kuwait and extrapolation of VSL from the US
and Europe).
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
Daily Deaths in Kuwait -2 Aug 90 to 6 Nov 91


Daily Deaths Among Those In Kuwait
(data from PACI and MOF files).
4
Oil Fires Started – mid-February 1991
3

Death Rate

Before Oil Fires -- 0.28 deaths/day
2


During Oil Fires – 0.47 deaths/day
Correlation Between Deaths and Smoke
Concentration ~ 0.6 (Weekly Averages)


1
Not statistically significant
However, only 200,000 exposed people and
a relatively short period of 248 days.
0
0 1 J UL 1 9 9 0
0 1 SEP1 9 9 0
0 1 NOV 1 9 9 0
0 1 J A N1 9 9 1
0 1 MA R1 9 9 1
0 1 MA Y 1 9 9 1
Da t e
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
0 1 J UL 1 9 9 1
0 1 SEP1 9 9 1
0 1 NOV 1 9 9 1
0 1 J A N1 9 9 2
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
Current Ambient Air Pollution –
A Less Obvious, But Perhaps More Severe, Problem
Public Health Impacts of Kuwait Oil Fires – 28 May 2008

Fixed Site Monitoring

Khaldiyah & Um Al-Haiman
 PM10 ~ 100 mg/m3 annual mean

Khaldiyah, Um Al-Haiman & Um AlAish
 PM2.5 ~ 45 mg/m3 annual mean
 1/3 crustal

Values Similar to MEXICO CITY !
“Back of the Envelope” Risk Assessment


Elements of Risk Assessment

Adult Population ~ 300,000

Baseline Mortality Rate ~ 9/1000 p-yr

PM2.5 Level ~ 30 mg/m3

Risk Coefficient ~ 0.4 % per mg/m3
Results

~ 300 deaths per year
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
The Public Health Study
Phase I – Medical Monitoring & Screening

Hypertension
50%
40%
35.5%
34.7%
30%
19.6%

20%
10.2%
10%
0%
< 120/80
120-139 / 80-89
140-159 / 90-99
≥ 160/100
Resting Blood P ressure (mm/ Hg)
Cholesterol
HSPH conducted medical
monitoring and screening of about
5% of the participants in Phase I.
In addition to the questionnaire
information, the participants’
height, weight and resting blood
pressure were determined. Venous
blood samples were collected and
analyzed for cholesterol & blood
sugar and for the metals lead and
mercury.
60%
38.8%
40%
36.3%
24.9%

20%
0%
< 200
≥ Cholesterol
200 & < 240
Cholesterol Level (m g/dl)
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
≥ 240
The results indicate high levels of
both hypertension and total
cholesterol in this sample of elderly
Kuwaitis.
Blood Mercury Levels
Tentative Results from Phase I Medical Screening

1000.0

Lognormal distribution

15% <= LOD (1 mg/dL)

Median ~ 5 mg/dL

GSD ~ 2.5
WHO and US EPA criteria:
Blood Mercury (ug/L)
100.0

10.0
1.0

0.1
-3.00
-2.00
-1.00
0.00
Normal Z-Score
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
1.00
2.00
3.00

WHO PTWI = 0.5 mg/kg-d or ~ 25 mg/dL

EPA RfD = 0.1 mg/kg-d or ~ 5 mg/dL
Values regressed on fish consumption,
gender and age:

Women’s levels 3 mg/dL lower than men’s.

Local fish consumption important.
All participants notified of results and
provided with Arabic version of EPA/RfD
guidance on mercury. Those with values
above WHO criterion were notified in
person by Dr. Behbehani and offered free
re-testing of their blood.
Kuwait National Center for Environment and Health
Epidemiology
MOH
Vital Statistics
Exposure
PHS I & II
PM & VOCs
Mercury in Fish
Kuwait Longitudinal
Health Study
Chemical Body
Burden
KU
Faculty of Medicine
KU
Science &
PAAC
Engineering
EPA
Integration & Interpretation
Kuwait Burden of Disease Study
Environmental Risk & Decision Analysis
MOH
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
KISR
EPA
Citizens
NGOs
Industry
Descriptive Epidemiology:
Infant Mortality 1950-2005 – “No Obvious Reversal”
180
Infant mortality rate per 1000
160
Male s IMR
140
Fe male s IMR
120
Re giste re d IMR
100
80
60
40
20
0
1940
1950
1960
1970
1980
Year
1990
2000
2010
Descriptive Epidemiology
180
100%
160
90%
140
80%
70%
120
Male s IMR
60%
100
Fe male s IMR
80
Male s 45q15
Fe male s 45q15
60
50%
40%
30%
40
20%
20
10%
0
0%
1940
1950
1960
1970
Year
1980
1990
2000
2010
% surviving from 15 to 60
Infant mortality rate per 1000
Adult Survival: 1950 to 2005 – “A Different Picture”
Descriptive Epidemiology
Comparing Male & Female Mortality: 2001 to 2005
“Excess Mortality of Young Men”
1
0
1
5
10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85
Age groups
0.1
Log nMx
Males
Females
0.01
0.001
0.0001
Unique Natural Experiment

Historically, epidemiologic studies linking specific warrelated exposures (e.g., war-related trauma) to longterm physical morbidity have been challenging for a
number of reasons

Often difficult to examine direct linkages of the original
experiences of war-related events given that most populations
in this circumstance experience prolonged periods of armed
conflict

Even after the armed conflict, most endure prolonged economic
crisis, lack of stable social systems and ongoing deprivation
including displacement to other countries as refugees which
may continue to impact health
Unique Natural Experiment

The invasion of Kuwait by Iraq in 1990 resulted in the 7
month long Iraqi occupation of Kuwait

After liberation, Kuwaiti nationals were encouraged to
return and did not live in exile for long periods

The physical and social infrastructure was restored in a
comparatively short period of time

Health and psychiatric care were made available to all
Kuwaiti nationals
Mind-Body Paradigm
Webster, Tonnelli and Sternberg 2002
linking psychological stress and affective states to disruption of key physiological mechanisms (e.g.
neuro-endocrine and immune functioning, oxidative stress, autonomic response) and ultimately to
disease expression

Major Salient Conditions:
PHS I and PHS II
Male
Female
60%
60%
50%
50%
40%
40%
30%
30%
20%
PHS I
20%
PHS II
PHS I
PHS II
10%
10%
0%
0%
Number of Men and Women Reporting
Witnessing Specific War-related Events
800
600
400
200
0
Men
Women
Number of Men and Women Reporting
Being Victim of Specific War-related Events
600
500
400
300
200
100
0
Men
Summary Measure of War-Related Trauma
 Challenging to summarize
mulitple items
250
Men
Number of Subjects
200
Wome
n
150
100
50
0
-1 -0.5 0 0.5 1 1.5 2 2.5 3 3.5 4
Trauma Score
Adopted modeling approach
that places subjects on a
continuous scale with
higher scores indicating
more severe trauma (Rasch
modeling)
Divided this equally
 no trauma
 low trauma
 intermediate trauma
 high trauma
War-Related Stressor Score and
Asthma Incidence: Adjusted
ModelHazards Ratio
(95% Confidence Interval)
War-related Stressors
None
Low
Intermediate
High
Reference Group
1.0 (0.6, 1.8)
1.4 (0.8, 2.4)
2.3 (1.3, 3.9)
Adjusted for: age, sex, education, literacy, smoking, BMI, oil
fire pollution
Multivariate Hazard Ratios of a 1-SD Increase in the
War-Related Trauma Score
Hazard Ratio (95% CI)
Events (Person Years)
Non-Fatal Myocardial Infarction
1.21
(1.05 1.40) §
186 (15021)
Angina Pectoris
1.59
(1.13 2.22)§
33 (15875)
(1.09 1.44)§
194 (14975)
Adjusted for: age, sex, education, literacy, smoking, BMI, oil fire pollution
* Includes MI and angina pectoris
§P<0.01
Coronary Heart Disease *
1.25
PHS Phase II
Father
Mother
N~1800
N~1800
Child
Child
Child
0-3 yrs
4-7 yrs
8-11 yrs
N~1200
N~1200
N~1200
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
55
PHS Phase II -- Questionnaire




Individual Characteristics

Age, sex

Height and weight

Smoking

Diet and exercise
Medical history
Date of death

(Deceased only)








56
Current functioning
Depression and anxiety
Checklist
Gulf War Syndrome
Social network and support
Coping strategies
Traumatic events

Location

During invasion,
occupation and oil fires
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
Social/Psychological
Before and After occupation
During occupation
PHS Phase II -- Biologic Markers


Anthropometry
 Height, Weight, and Blood pressure
Clinical Measures (fingerstick, 89% participation)
 Immediate reading





Blood drops stored on specially treated filter paper
Hair sample (60% of possible)


HgbA1C (blood sugar control)
C-reactive protein (systemic inflammation)
Future genetic testing


Total cholesterol, HDL, LDL, triglycerides
Blood glucose (random)
Environmental metals (mercury)
Salivary cortisol (“Children” only, 37% participation)

Measure of stress response
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
57
PHS III
PHS II
PHS II
PHS I
Accelerated Lifecourse Research Design
Longitudinal
Follow-up
The childhood shows
the man, as the
morning shows the
day.
John Milton, Paradise Lost (1667)
PHS Phase II
& Possible Extensions
Father
Mother
Recruit Siblings
Interview Age
N~1800
N~1800
Child
Child
Child
Child
Child
<16 yrs
16-19 yrs
20-23 yrs
24-27 yrs
28+ yrs
N~1200
Recruit Children of
Second Generation
Public Health Impacts of Kuwait Oil Fires – 28 May 2008
N~1200
Grand
Child
N~1200
Grand
Child
60
Grand
Child
Grand
Child
Grand
Child
Public Health Impacts of Kuwait Oil Fires – 28 May 2008