Short term effects of air pollution on health. The association between daily air pollution levels and mortality from respiratory and cardio-vascular diseases. Rony Braunstein Department of Epidemiology and Preventive Medicine Sackler School of

1
Short term effects of air pollution on
health.The association between daily air
pollution levels andmortality from respiratory
and cardio-vascular diseases.Rony
BraunsteinDepartment of Epidemiology and
Preventive MedicineSackler School of
MedicineTel-Aviv University

• Under The Supervision Of
• Dr. Ayana Goren
• Department of Epidemiology and Preventive
  Medicine
• Sackler School of Medicine
• Prof. David Steinberg
• Department of Statistics and Operation Research
• Schreiber School of Mathematical Sciences

2
VERY IMPORTANT NOTICE !!!

• This Power Point Presentation is not yet ready.
• Please consider it as an unfinished work.
• I hope it will be finished ASAP.

3
Air and Air - Pollution
Air is essential for the survival of human beings
as well as other populations. The importance of
air can be illustrated by the fact that the
average adult requires about 15 kilograms of air
per day compared with less than 1.5 kilograms of
food and about 2 kilograms of water (source
Goldsmith, 1968). Compared with other necessities
of life, obligatory continuous consumption is a
unique property of air. The average adult can
live 5 weeks without food, 5 days without water
and only a few minutes without air. Air is
essential for life and the absence of clean air
can lead to severe health problems and even to
death.
Since air pollution seems to be constantly
increasing both in the industrialized and in the
developing parts of the world, many studies have
been conducted throughout the world to examine
its characteristics and its effects on human
health. There are several known reasons for the
increase in air pollution (see e.g. Stern, 1968),
among which are the increasing amount of cars,
factories and power stations, which emit various
pollutants typical of urban and industrial
sources.
4
Topics
1. Air Pollution - An Introduction 1.1
Definition 1.2 History 1.3 Types 1.4
Effects 1.5 Sources
2. The Current Research - Methods and
Results 2.1 Objectives 2.2 Time Frame, Region
and Population 2.3 Data Bases 24 Theoretical
Model and Practical Fitting 2.5 Results
Conclusions and Discussion 2.6 The Delayed
Effect 2.7 A Comparative Study 80s vs.
90s. 2.8 Estimating the Count of Deaths
Related to Air Pollution
5
Air Pollution - Definition
Normal air consists of a mixture of several gases
(see e.g. Tebbens, 1968). The air will be
considered as polluted whenever (i) Other gases
and or particle matters will be present OR (ii)
When the balance of the normal air mixture will
be changed. The concentrations of the various
gases are represented in the table in parts per
million (PPM).
Concentration Of Gases Comprising Dry
Air Gas Concentration (PPM) Nitrogen 780900 O
xygen 209400 Argon 9300 Carbon
dioxide 315 Neon 18 Helium 5.2 Methane 1.0-1
.2 Krypton 1.0 Nitrous oxide 0.5 Hydrogen 0.5
Xenon 0.08 Nitrogen dioxide 0.02 Ozone 0.01-
0.04 Source Kuiper 1952.
6
HISTORY OF AIR POLLUTION AND HEALTH STUDIES
1. Acute Effects Whenever a major air pollution
event occurred, it was noticed that the
population tended to suffer more frequently and
more severely from respiratory as well as from
cardio vascular diseases. Example The London
episode of air pollution (December 1952). A
tremendous increase in mortality and morbidity
from respiratory and cardio vascular diseases
(3500 - 4000 excess deaths).
2. Long Term Effects Later on, the research was
focused on the long term and cumulative effects
of air pollution on health.
3. Short Term Effects New research directions
deal with the short-term effects of changes in
air pollution concentrations, not necessarily
acute ones, on health. The research of short-term
effects of air pollution on health is considered
to be most important nowadays and epidemiologists
are trying to deal with this question in many
places in the developed parts of the world.
7
Types of Pollutants
1. Sulfur compounds. Combustion of sulfur
containing fuels contributes large amounts of SO2
and some SO3. Many industrial processes generate
H2S.
2. Nitrogen compounds. NO and NO2 are produced in
some industrial operations and in automobile
engines. The measurements of NO and NO2 are
combined together and called NOX.
3. Carbon compounds. Carbon dioxide and carbon
monoxide arise in huge amounts from the
combustion of carbonaceous fuels. In large
cities, most of the daily production of CO and
CO2 results from incomplete utilization of carbon
in gasoline in automobile engines, industrial
machineries and power plants.
4. Photochemical compounds. The photochemical
compounds are defined as Secondary Pollutants.
The process of creating photochemical compounds
needs bright sunlight in which nitrogen oxides,
hydrocarbons and oxygen interact chemically to
produce oxidants like ozone and peroxyacetyl
nitrate (PAN).
5. Halogen compounds. Some industrial and
metallurgical processes produce inorganic halogen
compounds such as HF and HCl.
6. Organic compounds. The organic compounds are
emitted principally as vapors and the less
volatile compounds may occur as liquid droplets
or solid particles. Their sources are (i)
Household products including paints, paint
strippers and other solvents, (ii) wood
preservatives, (iii) aerosol sprays and air
fresheners, (iv) insect repellents and
disinfectants, (v) stored fuels and automotive
products and (vi) dry-cleaned clothing.
7. Fine solids and coarse particles. The finer
aerosols include man-made particles of metal,
carbon, tar, oxides, nitrates, sulfates,
chlorides, fluorides and silicates on one side
and natural particles such as dust, resin,
pollen, fungi and bacteria on the other side. All
those fine solids are less than 10? in diameter.
8. Radioactive compounds. Atmospheric
radioactivity originates from natural as well as
artificial sources. The natural radioactivity
results from the presence of radionuclides, which
originate from radioactive minerals or from the
interaction of cosmic radiations with the gases
of the atmosphere. Radon (Rn) is an example for a
natural radioactivity source. The man-made
radioactivity is a by-product of the developed
nuclear activity in the last fifty years.
8
Air Pollution Effects
1. Visibility reduction The effect of air
pollution on visibility is currently the most
easily observed. The effect of reduction in
visibility is produced by the scattering of light
from the surface of airborne particles.
2. Material damage Air pollution causes damage
to certain materials. Direct damage is caused to
structural metals, surface coatings, fabrics and
other materials of commerce. The destruction is
related to many types of pollutants.
3. Agricultural damage A large number of food,
forage and ornamental crops have been shown to be
damaged by air pollution and in particular by
halogen compounds such as HF and HCl.
Photochemical compounds, sulfur compounds and
nitrogen compounds also contribute to
agricultural damage.
4. Psychological effects Fear is a recognizable
element in public reactions to air pollution and
the psychological aspects of the phenomenon
cannot be ignored. It is possible that
psychosomatic illnesses are related to air
pollution events.
5. Physiological and health effects There is a
lot of evidence that air pollution can be
destructive to human health and can even result
in death. Air pollution affects human health by
causing respiratory diseases, cardio-vascular
diseases, lung cancer and other diseases.
9
Health Effects

1. Acute sickness or death.
2. Insidious or chronic disease, shortening of life
   or impairment of growth.
3. Alteration of important physiological functions,
   such as ventilation of the lung transport of
   oxygen by hemoglobin or various functions of the
   nervous system.
4. Unpleasant odor, impairment of visibility or
   other effects sufficient to lead individuals to
   change residence or place of employment.
5. Air pollution is known to be a causal factor in
   chronic pulmonary diseases like lung cancer,
   bronchitis, emphysema, and asthma.
6. Aggravation of existing cardiovascular disease.
7. Eye and respiratory tract irritation, headache,
   dizziness, visual disorders, and memory
   impairment are among the immediate symptoms that
   people have experienced soon after exposure to
   some organic compounds.

10
Air Pollution Sources
1. Natural Sources 1.1 Eruptions of volcanoes
pollute the air with dust and ash. 1.2 Dust
storms can fill the air with particles. 1.3
Forest fires create black smoke and emit noxious
gases. 1.4 Radioactive minerals and cosmic
radiations are the causes for atmospheric
radioactivity.
2. Artificial Sources 2.1 Energy sources like
wood and coal emitting black smoke and produce
noxious gases. 2.2 Extensive use of petroleum in
industry and transportation. 2.3 Industrial
processes produce pollutants as side effects of
their industrial process. 2.4 Nuclear power
stations and nuclear experiments (and accidents)
contribute to radioactive pollution.
11
APHEA2 - 32 European Cities (1996 2001)Air
Pollution and Health a European Approach
K. Katsouyanni (Greece), J. Schwartz (USA)
B. Kriz, M. Celko, J. Danova (Prague, Czech
Republic) F. Kotesovec, J. Skorkovski (Teplice,
Czech Republic) J. Pekkanen, P. Tittanen
(Finland) S. Medina, A. Le Tertre, P. Quenel, L.
Pascale, A. Boumghar (Paris, France) D. Zmirou,
F. Balducci (Grenoble, France) C. Spix, A.
Zanobetti, H.E. Wichmann (Germany)
K.Katsouyanni, G. Touloumi, E. Samoli, A.
Gryparis, Y. Monopolis, E. Aga, D. Panagiotakos
(Greece, Coordinating center) A. Paldy, J.
Bobvos, A. Vamos, G. Nador, I. Vincze, P. Rudnai,
A. Pinter (Hungary) Clancy, P. Goodman
(Ireland) A. Goren, R. Braunstein (Israel) MA.
Vigotti, G. Rossi, E. Cadum, G. Costa, L. Albano,
D. Mirabelli, P. Natale, L. Bisanti, A. Bellini,
M. Baccini, A. Biggeri, P. Michelozzi, F.
Forastiere (Italy) J. Schouten, J. Vonk (The
Netherlands) B. Wojtyniak, D. Rabczenko, K.
Szafraniek (Poland) E. Niciu, V. Frunza, V.
Bunda, (Romania) M. Macarol-Hitti, P. Otorepec
(Slovenia) J. Sunyer, M. Saez, F. Ballester, S.
Perez-Hoyos, E. Alonso, K. Kambra, A. Arangues,
A. Gandarillas (Spain) B. Forsberg, B.
Segerstedt, (Sweden) C. Schindler (Switzerland)
Z. Dortbudak (Turkey) HR Anderson, R. Atkinson,
J. Ayres (U.K.) .
12
Research Objectives
General ObjectiveThe general aim of this
research is to examine the short-term effects of
air pollution on mortality from respiratory and
cardio-vascular diseases.
The first objective is to fit a
Mathematical/Statistical model to the data set in
order to be able to define and measure the
short-term effects caused by various
air-pollutants on mortality from several related
diseases.
The second objective is to examine the delayed
effect of air pollution on mortality from cardio
vascular and respiratory diseases.
The third objective is to conduct a comparative
study on air-pollution in the Tel-Aviv metropolis
area over time. In this study air-pollution
concentrations in the early 1980s are compared
to those in the early 1990s.
13
Study Design
This study was based on data sets that were
collected in an earlier stage. The design of the
research was a historical prospective cohort
since (i) we used data that were already
collected in the past (ii) we examined the
count of deaths (the outcome) while levels of air
pollution (the exposure) were given.
. In this study the individual sample unit is a
single day for which we had (i) the count
of deaths from the relevant diseases (ii)
the exposure levels of air pollution (iii)
the measurements of several confounding variables.
14
THE RESEARCH TIME FRAME 1/1/1991
31/12/1996 Six Years 2192 Days
THE RESEARCH AREA Metropolitan Tel-Aviv
THE RESEARCH POPULATION Includes all the
people living in the wider Tel-Aviv Area. No
limitations regarding age or sex or any other
condition.
Population Size in the Tel-Aviv Area (In
Thousands) YEAR COUNT PCT of
Change 1991 1164.1 0.00 1992 1175.4 0.97
1993 1173.9 0.84 1994 1174.5 0.89 1995 11
77.5 1.15 1996 1177.2 1.12
15
Monitoring Stations in Tel-Aviv 1991-1996
Table Key 1 - SO2 , 2 - NOX , 3 - O3 , 4 -
PM10 Source Israeli Electrical Company
16
The Tel-Aviv Metropolis Map
IEC Monitoring Stations Network 1991-1996
17
THE RESEARCH VARIABLES
1. Dependent Variables 1.1 Daily Count of
Deaths from Cardio-Vascular Diseases ( 11.5
cases ) 1.2 Daily Count of Deaths from
Respiratory Diseases ( 1.5 cases )
2. Independent Variables 2.1 Daily Average of
SO2 ( 22.18 ?g/m3 ) 2.2 Daily Average of
NOX ( 139.87 ?g/m3 ) 2.3 Daily Average of
O3 ( 36.67 ?g/m3 ) 2.4 Daily Average of
PM10 ( 50.21 ?g/m3 )
3. Confounding Variables 3.1 Daily Amount of
Rain ( 14.69 mm ) 3.2 Daily Relative
Humidity ( 65.96 ) 3.3 Daily Average
Temperature ( 20.38? Centigrade ) 3.4 Day
of the Week 3.5 Holidays 3.6 Sequence Date
Time
18
Data Bases
1. The mortality data file The source for this
file was the Department of information and
computation in the Ministry of Health. This file
is based on death certificates, which are coded
in a data file and are kept in the Ministry of
Health data base. Each record in the mortality
data file contains the daily count of deaths from
respiratory and cardio vascular diseases as
defined by the ICD9.
2. The air pollution data file The source for
the air pollution data file is the Israeli
Electrical Company (IEC). The IEC has a network
of monitoring stations in Tel-Aviv where the
pollutants SO2 , NOX , O3 and PM10 are
measured every day. Each record in the air
pollution data file contains the daily average
concentration for each one of the pollutants.
3. The meteorological data file The source for
the meteorological data file are the
Meteorological Services. The meteorological
services have a meteorological station in
Tel-Aviv in which several meteorological factors
such as temperature, amount of rain, relative
humidity and cloudiness are measured daily. Each
record in the meteorological data file contains
the daily meteorological measurements.
19
FITTING THE MODEL
Poisson Regression Model Time Series Model
(not AR) Generalized Additive Model
WHERE C1Cc - are the confounders Si - is a
smooth function (loess) for the ith
confounder X1XK - are the pollutants gj - is an
appropriate transformation for the jth pollutant
Steps 1. Adjusting the confounding effects 2.
Adding the explanatory variables
20
Base Model Fitting
Quasi-likelihood model. The estimated dispersion
parameter 1.09
Cardio Vascular Mortality Model Fitting and the
Residuals
AIC 2301
Cardio Vascular Mortality Smooth Functions
(Loess) for the confounding effects
21
Results (Cardio Vascular Mortality)
Cardio Vascular Mortality RR of the Four
Pollutants as Non-Linear Terms
22
Results (Respiratory Mortality)
Respiratory Mortality RR of the Four Pollutants
as Non-Linear Terms
23
More ResultsCardio Vascular Diseases

Respiratory Diseases
24
THE DELAYED EFFECT OF THE POLLUTANTS
Lag Model for a Single Pollutant
Lag Model for Multiple Pollutants
Unstable Model Coefficients gt Unstable RRs
25
THE POLYNOMIAL DISTRIBUTED LAG MODELAlmon S.
(1965)
P is the polynomial degree l is the lag of
the coefficient (l0L) Ai - coefficients
describing the polynomial shape of the lag
coefficients
Matrix Notation
26
Example
27
PDLM - A Numerical Example CVD and SO2
28
CI for the Lagged RR
29
PDLM Final Results - CVD Respiratory
30
PDLM SUMMARY CONCLUSIONS
31
Changes in Air-Pollution Levels with TimeThe
1980s vs. The 1990s A Comparative Study
32
Changes in Air-Pollution Levels in the 1990s
Example Changes In SO2 Daily Concentrations
1/1/1991 31/12/1996
SO2 Crude Data
SO2 Crude Data In Log Scale
SO2 Lowess Smoother 90 Fraction
SO2 Time Series Plot
SO2 Lowess Smoother 10
33
Changes in Air-Pollution Levels in the 1990s
34
Changes in Mortality Levels in the 1990s
35
Estimation of the Count of Saved or Lost Lives
Single-Pollutant Model
Confidence interval for LS
Multi-Pollutant Model
Confidence interval for Multiple LS
SE2 in Matrix Notation
36
Estimation of Saved or Lost Lives 1980s vs.
1990s
Cardio Vascular Diseases 102 excess deaths per
year in the 90s as compared to the 80s the 95
CI..-125 , 343.
Respiratory Diseases 4 excess deaths per year in
the 90s as compared to the 80s the 95
CI..-71 , 91.
37
Estimation of Saved or Lost Lives - 1993 vs. 1996
Cardio Vascular Diseases 29 more deaths in 1996
as compared to 1993 the 95 CI..-66 , 128.
Respiratory Diseases 38 excess lives in 1996 as
compared to 1993 the 95 CI..5 , 70.
38
Estimation of Saved or Lost Lives Sensitivity
Analysis
39
Thats All Folks
Thank You Until Next Time !!!!