L17 - Causation and study design

Question 1

label study designs A-D

Answer 1

A - cross sectional study

B - case control study

C - cohort study

D - RCT

Question 2

CROSS SECTIONAL STUDY

i) name two things that are simultaneously measured in a defined population
ii) how are samples selected?
iii) how is prevalence reported?

Answer 2

i) defined population is surveyed to simultaneously measure disease/condition status and exposure
ii) samples are selected using inclusion/exclusion criteria
iii) prevalence is reported for the population as a whole and often for subgroups

Question 3

COHORT STUDY

i) what is it used to calculate and identify?
ii) what two groups are participants divided into? what should the two groups be?
iii) are variables not of interest matched or not matched?
iv) what is compared at the end of the study?
v) can it be retrospective, prospective or both?

Answer 3

i) calculate incidence (new onset of cases) and identify exposures (risk factors for a particular outcome)

ii) participants divided into exposed vs non exposed
- groups should be similar as possible

iii) variables not of interest are matched
iv) at the end of the study - relative numbers of new disease occurences are compared across groups
v) can be both retrospective and prospective

Question 4

CASE CONTROL STUDY

i) what is it used to identify? what is it not used for?
ii) which two groups of participants are selected?
iii) are they retrospective or prospective?

Answer 4

i) used to identify relevant exposures and not used to calculate incidence
ii) select one group with the condition (case) and one group without the condition (control)
iii) always retrospective - past exposures in both groups

Question 5

RANDOMISED CONTROLLED TRIAL

i) what are they used to test? (2)
ii) how participants allocated to recieve the intervention of interest or control?
iii) what should be ensured before allocation?
iv) when are outcomes compared?

Answer 5

i) safety and efficacy/effectiveness of the interventions
ii) random allocation
iii) ensure the two groups are essentially identical before allocation
iv) outcomes compared at the end of the trial

Question 6

EVIDENCE FOR CAUSAL ASSOCIATION

i) which study has the highest evidence for causal assoc?
ii) what type of studies are cross sectional and case studies? do these have a high level of evidence for causal assoc?
iii) what are the two components of an analytical study design - give three examples of this type of study

Answer 6

i) RCT
ii) cross sec and case study are descriptive - low level of evidence for causal relationshop

iii) analytic > hypothesis and a test
- case control, cross sectional and cohort study

Question 7

CONFOUNDING

i) what is a confounder?
ii) what can it influence? what can it make it look like?

Answer 7

i) a variable that influences both the dependent and independent variable causing spurious association
ii) can influence exposure/disease and can make it look like the exposure has caused the disease

Question 8

BRADFORD HILL CRITERIA

explain each step

• strength of assoc
• consistency
• specificity
• temporality
• biological gradient
• plausibility
• coherence
• exper evidence
• analogy

Answer 8

• strength of assoc - a high association is likely to be causal

consistency - consistent findings by different people

specificity - specific cause leads to specific effect

temporality - cause must preceed the disease

biol gradient - strength of assoc increases as exposure increases

plausibility - explained by biol mechanisms

coherence - doesnt contradict well known facts

experimental evience - showing absence of A = absence of B

analogy - similarties between the observation and other observations eg smoking and exposure to second hand smoke can have the same effects

Question 9

AIDS - STUDY DESIGN

i) what type of study is good when a new disease arises?
ii) which study is good to establish a case definition
iii) which large study was launched in 1984? what type of study was this?

Answer 9

i) case studies
ii) cross sectional

iii) MACS - multicentre aids cohort study
- cohort study

Question 10

SUMMARY

i) name two retro/prospective studies
ii) name three analytical-observational
iii) name one analytical experimental
iv) which study gives best evidence for causal association between exposure and outcome?
v) what are kochs 4 postulates?

Answer 10

i) cross sectional/cohort
ii) cross sectional, cohort, case control
iii) RCT
iv) RCT

v) presence of MO in diseased organism
- MO can be isolated and cultured
- intro of MO to healthy organism causes disease
- MO can be re isolated from inoculated disease organism and identified as the original agent