L6 Ecological studies

Question 1

Define an ecological study.

Answer 1

Observational study where data are analysed at the population or group level, rather than individual level.

Question 2

What are the advantages of ecological studies. (3)

Answer 2

• Inexpensive
• Easy to undertake – use routinely collected data
• The population context of individual characteristics has been demonstrated to be stronger determinant of disease at population level than individual level risk factors

Question 3

What are the advantages of ecological studies. (4)

Answer 3

• Prone to bias & confounding
• Care must be taken:

1. When extrapolating to individuals within the measurement areas
2. Or to higher population level

• Other designs generally considered more reliable – with respect to inference & causation

Question 4

Read

Why do an ecological study?

• To monitor population health
• Informing & develop public health strategies
• To make large scale comparisons - e.g. between countries
• To study the relationship between population-level exposure to risk factors & disease
• To study contextual effect of risk factors on population • If individual measurements are not available - e.g. confidentiality precludes release of data on individuals – so anonymized by aggregation of data to small area level
• If disease being studied is rare – so aggregation of data is needed for analysis to be done

Question 5

Read

Types of measurement in ecological studies

Health Outcomes:

• Aggregates of individual health data

1. E.g. Prevalence
2. Incidence
3. Rate of disease

Ecological risk / Exposure data:

• Aggregate measures
• Data are summaries of individual level data - e.g. mean, percentage, area-level deprivation indices

Environmental measures:

• Equivalent individual level data are conceivable - e.g. mean annual exposure to fluoridation

Global measures:

• No equivalent individual level data- E.g. number of GP practices = Population density

Question 6

Name the 3 types of ecological studies.

Answer 6

• Geographical
• Longitudinal
• Migration

Question 7

Read

Ecological fallacy

• Type of confounding specific to ecological studies
• Occurs when relationships found for groups are assumed also to be true for individuals
• There may be unknown confounding factors that display the same pattern of spatial heterogeneity
• E.g. an area with a majority of girls may have a higher mean level of dental decay than another area with a majority of boys
• It might be inferred that girls have a higher mean level of dental decay than boys
• HOWEVER – 2 types of potential confounding:

1. Confounding common to all observational studies
   
   1. Due to unknown characteristics of the study area
   2. E.g. first area might have greater deprivation than the second area
   3. It might be inferred that children from deprived areas have greater levels of dental decay
2. Confounding which is known as the ‘ecological fallacy’
   
   1. Even if adjustments are made for all confounders the data are aggregated – does not provide exact information on those individuals who have high levels of dental decay

• ​There are only a few boys in the first population
• However, these might be the only children with high levels of dental decay – perhaps very high levels which make the mean level greater than the second population
• If this is true then the inference from the ecological analysis would be incorrect & possibly reversed at the individual level

Question 8

Read

Types of Ecological study

Geographical

• Compares one geography with another
• Assesses health of population of each
• Exposures for geographies may also be measured & included in analysis
• Also, other possible confounders - e.g. demographic, socioeconomic data Longitudinal

Population

• Monitored to assess changes in disease over time
• Confounding factors may be included in the analysis

Migration

• Migrant population data collected & analysed
• Unit of interest – population type
• Not time or place