CAT: COHORT

Question 1

What is clinical equipoise?

Answer 1

there is genuine uncertainty in the expert medical community over whether a treatment will be beneficial for a group

Question 2

What is a cohort study?

Answer 2

A type of longitudinal observational study in which a group of people with a particular exposure and a group of people without this exposure are followed overtime. The outcomes of the people in the exposed group are compared to the outcomes of the people in the unexposed group to see if the exposure is associated with particular outcomes

Question 3

What are the strengths of a cohort study?

Answer 3

Best information for causation as temporal sequence if explicit in the study design
Able to examine a range of outcomes
Good for rare exposure
Incidence can be directly calculated
No recall bias like case-control
Can do studies that may be considered unethical to do as an RCT

Question 4

What are the limitations of a cohort study?

Answer 4

Long follow up - expensive and time-consuming
risk of loss-to-follow-up which could introduce bias
Susceptible to confounding and without appropriate control it will influence observed associations
Bad for rare outcomes
Bad for long latency periods
Potential for selection bias

Question 5

What is the difference between a primary and a secondary outcome?

Answer 5

Primary outcome = judged by the researchers to be the outcome on which the overall effectiveness of the intervention will be judged
Secondary outcomes = considered to be of interest and potentially important, but not the main basis for judging effectiveness

Question 6

Why is it considered good practice to distinguish between primary and secondary outcomes?

Answer 6

Clarity in study design
When outcomes are pre-specified and categorised it reduces the risk of post-hoc analysis bias
Ensures study resources and analyses are prioritised for the most importance questions
Guides future research
This forms the basis for establishing power and sample size for the study

Question 7

What is bias?

Answer 7

The systematic introduction of error into a study that can distort the results in a non-random way

Question 8

What is selection bias?

Answer 8

A systematic error in assigning individuals to groups leading to differences in the group’s qualities and this may influence the outcomes
Also means the sample will not be representative of the population

Question 9

What is information bias?

Answer 9

Aka measurement bias
A statistical error that arises from inaccuracies in the measurement or collection of data i..e information about participants is incorrect

Question 10

What is procedure bias?

Answer 10

Aka performance bias
When there are systematic differences in how different groups in a study are treated or managed, affecting the study outcome

Question 11

What is confounding?

Answer 11

The distortion of the association between an exposure and an outcome by a 3rd variable which is related to both exposure and outcome

Question 12

What is confounding by indication?

Answer 12

a form of confounding in which the clinical indication for receiving the study treatment is a risk factor for the outcome of interest

e.g. pts with high bp will be given an anti-hypertensive andthe outcome is a reduction in cardiovascular event. If the study finds that pts on these drugs have a higher cardiovascular event rate this may be due to the fact they have a higher risk of this in the first place not because the medication causes the outcome

Question 13

How can a study address confounding?

Answer 13

Randomisation
Restriction
Matching
Multiple variable regression
Stratification e.g. propensity score
Adjusted results e.g. adjusted odd ratios
Sensitivity analysis

Question 14

Ways you can assess the interval validity of a cohort study?

Answer 14

Appropriate study design to address the question?
Clear definition of cohorts?
Any selection bias in selection of participants?
How did it measure exposure? Was it consistent and accurate?
Good level of completion and follow up to identify meaningful complications?
Standardisation of data collection?
Blinding of outcome assessors?
Did it identify and address confounding?
Was the sample size sufficient for the power desired?

Question 15

Ways you can assess the external validity of a cohort study?

Answer 15

Is it UK NHS? Setting
Are patients comparable to pt group? - demographics, geographic? Inclusion criteria etc
Consider whether exposure pattern is likely to be similar to local population
Are treatments different within the study and the local population
Cost-effectiveness

Question 16

How can the way in which a group of individuals is recruited to a cohort study may influence the validity of the study?

Answer 16

Internal validity:
Selection bias from non-random selection
Attrition bias from follow up challenges
If non-comparable groups teres the potential for confounding variables to affect the validity

External validity:
If narrow inclusion/exclusion criteria or demographic differences, the resulting sample might not be representative of the broader population, limiting generalizability

Question 17

Explain why inaccuracy in exposure measurements could lead to biased results?

Answer 17

Non-differential error i.e. random misclassification - could bias towards H0 and hide true relationships (T2 error)
Differential error could bias in either direction creating spurious associations (T1 error) or obscuring true ones (T2 error)

Misclassfication of exposure can obscure true associations between exposure and outcome making it difficult to determine causality = reduced internal validity
Measurement errors cause the study results to inaccurately reflect the true relationship can reduce external validity

Question 18

Explain why inaccuracy in outcome measurement can lead to biased results?

Answer 18

Non-differential - If outcome measurement reporting is uniformly poor then it will tend towards H0 = type 2 error which underestimates true associations
Differential - If outcome measurement reporting is more accurate in 1 exposure group then you may get spurious positive relationship = type 1 error (can also bias the other way to cause type 2 error)

Question 19

What methods in a cohort study could enhance the accuracy of an outcome measurement?

Answer 19

Standardisation of measuring an outcome measurement
Outcome established blind to exposure category
Consistent protocols
Statistical techniques for measurement error e.g. regression or sensitivity analysis
Comprehensive follow up to reduce attrition bias

Question 20

What would the consequence be for a study if all important confounders were not identified?

Answer 20

The association between exposure and outcome could be biased
Can create spurious associations or distort the strength of the association = compromises the internal validity as causal relationship cannot be estimated
Limits applicability = reduces external validity

Question 21

Explain why completeness of follow up can influence the validity of a study?

Answer 21

Incomplete follow up allows for selection bias, attrition bias and may reduce the power of a study even if it is non-differential between exposure groups
Differential follow up will lead to bias

Question 22

Explain why length of follow up can influence the validity of a study?

Answer 22

Insufficient length of follow up may mean that relevant outcomes are not observed and associations are not detected

Question 23

What is restriction in terms of minimising confounding?

Answer 23

When you limit the study to subjects in 1 category of the confounded to ensure all participants have the same level of the confounder e.g. if smoking is the confounder then you could limit the population to non-smokers

Question 24

What is matching in terms of minimising confounding?

Answer 24

Matching the subjects with respect to the confounding variables
E.g. each smoker who is enrolled is matched to a non-smoker of similar age

Question 25

What is attrition bias?

Answer 25

bias that arises from systematic differences in the way participants are lost from a study.

Question 26

“What are the implications of this cohort study for practice”

Answer 26

1 observation study rarely provides sufficiently robust evidence to recommend changes to clinical practice
For some questions, observational studies provide the only evidence
Recommendations from the study will be stronger if supported by other evidence

ChatGPT ideas:
Is it relevant to clinical practice?
Could it influence or change current clinical guidelines?
Is it costs effective?
Does it identify areas where more research is needed?

Question 27

What analysis methods are used in cohort studies?

Answer 27

Relative risk
Hazard ratios
Rate ratio