2. It is not easy to establish links between
environmental exposures and health outcomes
using individual level data
• Problems for obtaining measures of dose levels for
individuals
• There is also often the difficulty to deal with long
latency times between exposure and diagnosis. (eg
cancer)
Ecological study designs provide one way of
avoiding these difficulties and studying the
effects of environmental exposures on different
health outcomes.
3. Definition: observational study with
group (instead of individual) being unit
of observation (unit of analysis)
Types:
• Exploratory study
• Multiple group comparison study (geographical
study)
• Time series study
4. The simplest ecological study
Observing differences in rates of disease in
different geographical areas
No exposures are measured and, generally, no
formal data analysis is used.
The objective is to search for spatial patterns
that might suggest an environmental etiology
or a special etiologic hypothesis.
5. • Ex1: Age-adjusted cancer mortality rates in the US by
county for the period 1950-69;
• Difference statistically significant for oral cancer in
geographic patterns by sex
- Men: mortality rates >in the urban Northeast
tobacco smoking and alcohol consumption
- Women: mortality rates >in the Southeast
hypothesis: snuff dipping
• confirmed by a case-control study
• [Morgenstern H., AJPH, 1982,72,12, 1336-1344]
6. Observe the association between the
average exposure level and the disease rate
among several groups.
Eg:
• Emile Durkheim, who investigated suicide in
western Europe during the 19th century.
• The data on four groups (religion and suicide rate)
of Prussian provinces between 1883 and 1890
studied.
• Result: Protestants were 7½ times as likely to
commit suicide as were other residents (most of
whom were Catholic)
7.
8. Observe the relationship between the change in
the average exposure level (or intervention) and
the change in the disease rate for a single
population
With time trend studies involving a sudden
change in exposure (e.g: the start of an
intervention program):
• compare the slope in the disease trend before and after
the intervention.
With
time trend studies involving a gradual
change in average exposure level:
• must compare trends in both variables.
9.
10.
11. Frequently used in environmental epidemiology where it is
difficult to measure exposure at the individual level
Generally use pre-existing data, which have been collected for
other purposes and which can be available on large
populations
• there are many sources of environmental data, from environmental
surveillance systems which can be used (air pollution, drinking water
quality, temperatures…)
Can also be used for studyng infectious diseases, especially in
surveillance, in order to detect signals which can be alerts
Also useful to evaluate the impact of intervention programs or
policies
12. Data on many variables (e.g., behaviors, or medical hx) may
not be available at the population level, and the results of
ecologic analyses are subject to certain limitations.
• unlikely to give reliable and accurate estimates of risk but they can
be useful for the purposes of formulating hypotheses, that may
then be examined by other methods.
Ecological studies look for associations between the
occurrence of disease and exposure to known or suspected
causes.
However, the unit of analysis is not an individual but a
population / a group of persons
Disease rates and exposures are measured in each population
and their relation is examined
13. The principle of this study design is
To identify changes in the frequency of an health event in
space and/or time
To relate these changes with environmental factors
14. The 2 major objectives of ecological
studies are:
4) to generate or to test etiologic hypotheses, in order to
explain disease occurrence;
5) to evaluate the impact of intervention programs or
policies (if we have the knowledge of disease etiology).
15. Use of agregated data and not the individual data
The statistical unit corresponds to a group of persons
• Levels of exposure are not measured individually ; but use mean
exposures in the population
• Health outcome - use the frequency of the disease in the
statistical unit
The principle is to look if there is an association between
• variation of exposure levels and
• variation of health indicators
16. For example:
• to look for an association between mean chlorination
levels in tap water in the district and the incidence
rate of bladder cancer in different geographical areas
(geographical study), or;
• to look for an association between daily levels of air
pollution and the daily numbers of hospital admissions
for respiratory diseases (time series study)