18 terms

Ecological Study


Terms in this set (...)

Ecological vs. cross-sectional study design
Assess exposure and outcome at the same time for a single time point or period. Ecologic studies focus on the group as the unit of analysis, while cross-sectional studies focus on the individual.
Ecological study basic design 1
1. Uses summary measures of outcome (disease or mortality rates) and summary measures of exposures
2. Unit of analysis is the group vs. the individual
Ex: country, state, county, city, census tract, hospital, school
3. Descriptive (exploratory): no a priori hypothesis, only outcome is measured
Ecological study basic design 2
4. Analytic: a prior hypothesis; multiple group, time-trend, and mixed ecological
5. Ecological studies tend to be more hypothesis generating than testing because we can generate hypotheses for the individual level based on group level
Types of Ecological studies
1. Multiple group studies or ecological comparison studies
2. Time trend studies
3. Mixed ecological studies
Multiple group studies or Ecological comparison studies
1. Involve assessment of the correlation between exposure rates and disease rates among different groups or populations over the same time period
2. The ecological units are places (countries, states, institutions, etc.)
Time trend studies 1
1. Involve correlation of changes in exposure with changes in disease over time within the same aggregate group
2. The ecological units are time periods within one group
Time trend studies 2
3. Assess associations between average change in exposure and average change in disease frequency to:
a. screen etiological hypotheses
b. generate new etiological hypotheses
c. evaluate interventions or policies applied to groups
Mixed ecological studies
1. The ecological unit is defined by both place and time
2. Examine associations between exposures and outcomes for several populations over time
3. Combine features of multiple group and time trend studies
Types of variables for ecological study 1
1. Aggregate measures: summarize the characteristics of individuals within a group as the mean value of a certain parameter or the proportion of the population or group of interest with a certain characteristic
Types of variables for ecological study 2
2. Environmental measures: represent physical characteristics of the geographic location for the group of interest
3. Global measures: represent characteristics of the group that are not reducible to characteristics of the individual (e.g., type of political or health care system in a region)
Strengths of ecological study 1
1. Quick and relatively inexpensive
2. Relatively simple to conduct
3. Data available from surveillance programs and disease registries (secondary)
Strengths of ecological study 2
4. May lead to more accurate conclusions than individual level data when: a) within-population variability of the exposure is low but between-population variability is high; b) the implications for prevention or intervention are at the population level
Limitations of ecological study 1
1. Ecological fallacy: aggregation bias, making causal inferences about an individual phenomenon or process on the basis of group observations
- Associations among groups may not hold true at the individual level
2. Summary measures of exposure or outcome may hide variability in the data and lead to imprecise measures of association
Limitations of ecological study 2
3. Information bias: within group misclassification and lack of adequate data leads to proxy measures and imprecise measurement of exposure or disease
4. Difficult to identify and control confounders at the group level
5. Temporal ambiguity
Bias in ecologic study
1. Classification bias: generally non-differential, insufficient quality of exposure and outcome classification.
2. Surveillance bias: can be differential, ascertainment of disease or exposures or both may differ from one place to another.
3. Ecologic fallacy: inability to generalize data gathered at the group level to results at individual level.
Ecologic fallacy
A logical fallacy in the interpretation of statistical data where inferences about the nature of individuals are deduced from inference for the group to which those individuals belong
Source of ecologic fallacies
1. Information bias within groups
2. Between group confounding
3. Effect modification by group
4. Others (i.e. immigration)
How to control confounding in ecologic study?
1. Multivariate model, treat as covariate in the analysis
2. Standardization for possible confounders in the exposure and/or outcome.