unit 12

what are observational studies

• non experimental

• observe describe and analyse individuals or populations without intervention

what are the 2 major catagories of epidemiological studies

• experimental

• non-experimental

experimental + descriptive

none

experimental + analytical

clinical trials

field trials community trials

non-experimental + descriptive

ecological

cross-sectional

case reports

case series

non-experimental + analytical

Cohort

case-controlled

ecologic study is what (catagotiries)

• non-experimental

• descriptive

Ecological study

what is the unit of analysis in an ecologic study?

the group or community level - not the individual

Ecological study

what does an ecologic study examine

intergroup variability

(compares rates of disease between different populations or areas)

Ecological study

what is the main purpose of an ecologic study

to generate hypotheses about potential associations

Ecological study

if a finding in an ecologic study shows a correlation what is the necessary next step

the findings must b confirmed in more rigorous studies on individual

Ecological study

an example of how ecologic studies are commonly used

measuring disease rates for different geographic areas and relating them to area characteristics like median income, air pollution.

Ecological study

what is a classic example of an ecologic study at the country level?

examining the relationship between country’s gross domestic product (GDP) and its populations life expectancy

Ecological study

what is ecological fallacy

the error of assuming that association observed at the group/population level is true at an individual level

Ecological study

ecological fallacy

the data is often correct however the correlation is what is wrong

Explain Emile Durkheim's suicide study as an example of ecological fallacy.

A: Group level: Protestant areas had higher suicide rates than Catholic areas (attributed to social cohesion).

Individual level: Catholics living in Protestant areas had a higher suicide rate than Catholics in Catholic areas. The individual risk was linked to social isolation within the broader community, not just religion.

read

done

Ecologic studies

what is aggregation bias (caused by ecological fallacy)

data may show an opposite effect to what is happening at individual level

cross sectional

• non-experimental

• descriptive

what is a cross-sectional study

snapshot of a population at one point in time

what can a cross sectional study measure

attitudes, behaviours, health conditions (past or present) or risk factors (past or present)

cross sectional study

how are participants selected

from a target population using a simple random sample

cross sectional study

during sampling you can also

oversample particular subgroups/strata (e.g favour a specific age group)

cross sectional study

what is a DESCRIPTIVE cross-sectional study

measures one parameter

(e.g measuring the prevalence of diabetes in adults over 40)

cross sectional study

what is an ANALYTICAL cross sectional study

measuring both outcome and exposure to look for an association

(e.g measuring the prevalence of obesity and the prevalence of diabetes in the same population to see if they are linked)

case-control

non-experimental

analytical

case-control

what is known about the outcome at the time of the case-control

it has already occurred

case-control

what is known about the exposure at the time of the case-control

data on exposure to potential risk factors is collected retrospectively ( usually form questionaries)

case-control

what is the direction of the investigation

from outcome back to exposure

case-control

how does a case control study begin

identifying individuals with the disease (cases) and without the disease (controls ) and looking back to compare past exposure

case-control

what is the primary measure of association

the odds ratio

case-control

why is relative risk not calculated

because the investigator chooses the number of cases and controls (rather than following a whole population from exposure to outcome)

calculate odds ratio

(90/10) / (30/70) = 21

calculate relative risk

(90/120) / (10/80) = 6

(risk exposed / risk unexposed)

case-control

issues

what biases

• selection bias

• information bias

• reverse causality

case-control

how do we reduce selection bias

selecting controls from the same population as cases

(making them as similar as possible expect for the outcome)

case-control

how to reduce information bias

asking both cases and controls about a number of exposures in the same manner

(bc sick people more likely to remember or report an exposure)

case-control

how to reduce reverse causality

need other study designs to confirm findings

case-control

what is matching in cross sectional studies

cases are individually matched to control key characteristics

(e.g age, sex)

and a matched pair analysis takes place

case-control

what is a nested cross sectional study

cases and controls are selected from within an existing cohort study.

to examine other possible risk factors

Cohort

non experimental

analytical

cohort study

how does it begin

identifying groups (cohorts) of people based on their exposure status (exposed vs not exposed)

cohort study

what is known about the outcome at the start of the study?

the outcome has not occurred yet

cohort study

what happens after the exposed and unexposed cohorts have been identified

they are followed up

cohort study

what is the direction of the investigation

moves forward in time

from exposure (present) to outcome (future)

cohort study

why is the follow up essential to happen ASAP

so that any differences can be explored

cohort study

what is a bias

measuring bias

cohort study

how to reduce measuring bias

collecting data on exposure and outcomes