What are ecological studies? How do they differ from cross-sectional studies?
•These are observational studies as opposed to experimental studies
•They study the determinants of health in people living in their own environment
•Ecological studies are carried out at the level of the population
Cross sectional studies are carried out at the individual level
What is the ecological fallacy?
•We must not jump to conclusions about individual based on data from populations
What are some advantages of ecological studies?
•Data that are difficult to measure on individuals (such as alcohol intake) are easier to measure on groups
•Sometimes population health interventions target groups rather than individuals; in a case like this, an ecological study is appropriate
What are some limitations of ecological studies?
•Population level associations may not be the same as individual level associations (ecolocial fallacy)
•So ecological studies provide only weak evidence that a risk factor actually causes a disease
•Data are often only available for large population groups that are so heterogenous that the real sources of variability are masked
•Data may not be available in the detail needed to test a hypothesis (eg age-specific disease rates, or rates for non-fatal disorders such as arthritis)
•Variations in methods of data collection over time and between countries/regions
What is standardisation?
•Standardisation is a process of adjusting disease and death rates to take into account differences in demographics
•Direct standardisation is used to adjust death rates to eliminate the effect of differences in the age distribution
•It uses a standard population which has a defined percentage of people in each age group
What are descriptive cross-sectional studies used to monitor?
the health and risk factors of a population
What are analytic cross-sectional studied used to test?
they test ideas about the factors that are associated with health outcomes
What is a 'representative sample'?
–Ideally a random sample
–But this may not be possible (no sample frame)
–Other sampling strategies are possible, but the overall goal is representativeness
what is a non-response?
it is an important issue
–Decreases representativeness and introduces bias
–Can be partly compensated for by using ‘sampling weights’ to adjust prevalences
–But non-responders will differ from responders, so the greater the non-response rate, the greater the potential for biased results
what are the limitations of cross-sectional studies?
•Cross-sectional studies survey the people who are present in the population at a given moment
–This may under-represent people with diseases that are rapidly fatal, or last only a short while
•They measure the risk factor and the disease at the same time point
–Risk factors may be changed by the presence of disease (for example, people with heart disease changing their diet or quitting smoking)
•Period prevalence may be affected by recall bias
–People asked about “the last year” are likely to include events that happened before then, so prevalence is biased upwards
–People may forget, too, but this is less likely with significant health conditions
what are the strengths of cross-sectional studies?
•By using a standard approach to measurement, cross-sectional studies can keep a finger on the pulse of health
•Allowing changes to be detected over time, and comparisons made with other countries