Lecture 11 - Cross-setional and ecological studies Flashcards
Analytic epidemilogy
1. What does it do?
Quantify the degree to which an exposure increases/decreases the likelihood of the outcome
Descriptive Epi is what term in the definition of Epi? What about analytic?
What is the definition of Epi?
“The study of the determinants and distribution of health and diseases in populations”
Determinants for Descriptive and distribution for Analytic
Decriptive Epi is…..
- Has what studies
- First part of what?
- PPP
- Studies like this are good for?
- What are examples of descriptive epi?
- Has observational studies (so observing what’s already happening in the population)
- First part of public health model, “defining and measuring the problem - who, where, when, what”
- Person, place, time:
- person = used to describe population of interest - age, gender, proportion of ethnicity etc
- place = geographical locations matter
- time = seasons etc matter - For generating hypotheses/ideas
- Cross-sectional studies and ecological studies but they’re also vaguely analytic epi
Analytic epi……
- Has to do with what?
- Studies what?
- Has what kind of study?
- It deals with what part of the health model?
- Deals with what one word?
- Has to do with associations between exposures and outcomes
- Studies causation (cause and effect between exposures and outcomes)
- Has observational or intervention studies e.g. randomised control trials
- It deals with the second part of the public health model: Describing causes and consequences
- Deals with “Why” (not what, where, when etc) - quantify the association between exposures and outcomes
Cross sectional studies
- Measures what? (more than a sentence)
- There is a what?
- What can “point in time” mean?
- Measures exposures and/or outcomes at one point in time (assessing them at same time i.e. snapshot in time)
- There is a defined population - not selected because have particular exposure/outcome - they include group which is representative of the pop (take a cross-section of the population)
- Point in time can be a date, specific event, specified period of time
What do CSS measure?
- formula
- affected by both
- duration can be affect by
Prevalence
- (no. ppl w/ dis)/(total no. ppl in pop)
- affected by both duration and incidence
- duration can be affected by recovery and survival
What are CSS used for?
- To describe
- To compare
- To generate
- To……
- To describe e.g. prevalence of health condition
2 To compare (amongst other sub groups or if repeat study then overtime) prevalence
- To generate hypothesis
- To plan e.g. health service delivery (what’s likely to be required in terms of health services)
What are the limitations of CSS? 4
- Temporal sequencing
- can’t determine whether exposure exposure or outcome came first because they were assessed at the same time - Measures prevalence, not incidence
- can’t tell about onset of disease - Not good for studying rare outcomes and exposures - very few people would have it so wouldn’t provide meaningful info
- Not good for assessing variable and transient exposures/outcomes
- So your findings may be difference when you do your CSS at diff points in time
Why do CSS?
- Can assess multiple exposures and outcomes at same time
- Depends on your research question
- want to measure prevalence + distribution of prevalence in pop
- stable exposures and outcomes
- hypothesis generating - Can be less expensive
- Rel quick
Ecological studies
Basic principle about individual/group level study
Compares exposures and outcomes at GROUP not individual levels (so a data point is a pop, not individual
What are they used for? 3
- To compare bw pops
- To asess pop lvl factors (e.g. air pollution in diff cities and lvl of lung disease)
- To consider hypothesis (can’t confirm it but can see if ecological data is consistent with hypothesis)
Limitations of Ecological? 3
- Ecological fallacy - you can’t ascribe characteristics of that group to individuals within that group
- Can’t control for confounding
- Can’t show causation
Why do Ecological/
- Depends on research Q
- pop lvl exposures - might wanna look at them
- consideration of hypothesis - Data is often routinely collected
- may be relatively easy to do
- may be relatively inexpensive
