Research Studies - Study Design, Bias, Confounders

Question 1

What is bias

Answer 1

One outcome is systematically favoured

Question 2

Selection bias
-definition
-forms of selection bias

Answer 2

Error in assigning individuals to groups leading to differences which may influence the outcome

Sampling - subjects not representative of population
Volunteer - people more likely to participate
Non responder - people are less likely to respond

Question 3

Recall bias
-what is it
-which kind of study is more likely to have this

Answer 3

Difference in the accuracy of the recollections retrieved by study participants

Case controls
-people affected may search their memories more thoroughly for a history of exposure

Question 4

Publication bias
-what is this
-which kind of study is more likely to have this

Answer 4

Failure to publish results from valid studies, especially as they show negative or uninteresting results

Meta-analyses

Question 5

Work-up bias/verification bias
-what is this

Answer 5

When comparing new diagnostic tests with gold standard tests, clinicians may be less likely to order the gold standard test unless the new test is positive, as the gold standard is often more invasive

Question 6

Expectation bias
-what is this
-when is this a problem

Answer 6

Observers measure or report data in a way that favours the expected study outcome

More likely in non-blinded trials

Question 7

Hawthorn effect
-what is this

Answer 7

Group of people change their behaviour because they know they’re being watched

Question 8

Late-look bias
-what is this

Answer 8

Gathering info at an inappropriate time
-studying a fatal disease many years after many participants may have died

Question 9

Procedure bias
-what is this

Answer 9

Subjects in different groups receive different treatment

Question 10

Lead-time bias
-what is this

Answer 10

When 2 tests for a disease are compared, 1 test diagnoses the disease earlier but there is no effect on the outcome of the disease

Question 11

Confounders
-what is it
-common examples

Answer 11

Distortion between an exposure and health outcome by an additional factor (cofounder)

Non random distribution of risk factors
-age
-sex
-social class

Question 12

How to detect publication bias

Answer 12

Tendency for studies with postive results to be published

Funnel plot - graph used to see if there is an evenly weighted publication results
:) symmetrical inverted funnel
:( asymmetrical funnel

Question 13

RCT key features

Answer 13

Random allocation to intervention or control

Practical or ethical problems may limit use

Question 14

Cohort study key features

Answer 14

Observational
Prospective
Looking for the relative risk

Participants selected according to exposure to see how they may develop an outcome

Question 15

Case control study key features

Answer 15

Observational
Retrospective
Looking for odds ratio
Quick
Good for rare conditions
Prone to confounders

Participants with cases are matched with controls
Data collection on past exposures for a possible causal agent

Question 16

Cross-sectional study key features

Answer 16

Prevalence studies
Weak evidence of cause and effect

Question 17

Study design - levels of evidence

Answer 17

Meta analysis - GOLD STANDARD

Primary goal - increase power and improve estimates of effect size, especially when individual studies may be too small