INTRODUCTION TO COMPETING PRIORITIES 20/9/16

Question 1

What 3 competing priorities do clinical studies require you to balance? (3)

Answer 1

1. design and conduct a study that provides the correct answer
   - by addressing potential biases
2. maximise efficient use of resources
3. comply with ethical, legal, institutional and professional regulations

Question 2

How do we go conduct studies which allow you to balance the three competing priorities? (7)

Answer 2

1. identify a problem and review previous studies
2. formulate a question to address a gap in the evidence
3. select a study design appropriate to this question
4. sample/select participants/ data sources
5. collect data/measure variables
6. analyse and interpret the analysis findings
7. disseminate the study’s results

Question 3

What are potential sources of bias in terms of sampling/selection bias?

Answer 3

• external validity = non-representative samples

- confounding = selection influences exposure and outcome

Question 4

What are potential sources of bias in terms of measurement bias? (3)

Answer 4

1. information bias
   - extent of information varies amongst participants
2. observer bias
   - influenced by prior knowledge or belief
3. recall/response/prestige bias
   - influenced by prior knowledge or belief

Question 5

What are the potential sources of bias in terms of analytical bias? (3)

Answer 5

1. loss to follow up
   - specific participants excluded
2. omitted variable bias
   - imprecise adjustment for confounding
3. attributional bias
   - interpretation of causality

Question 6

What are the potential sources of dissemination bias?

Answer 6

1. publication bias

- eventful results more likely to be published

Question 7

What is a priori study design and what does it do?

Answer 7

• avoids bias by design

Question 8

How does a cross-sectional study avoid bias?

Answer 8

provides evidence of association with in a sample

Question 9

How does a case-control study avoid bias?

Answer 9

provides evidence of association between samples

Question 10

How does a cohort study avoid bias?

Answer 10

provides evidence of directionality of associations

Question 11

How does a trial study avoid bias?

Answer 11

provides evidence of causality

Question 12

How does a meta-analysis study avoid bias?

Answer 12

provides evidence of reproducibility/generalisability

Question 13

Which bias is mostly tackled?

Answer 13

attributional bias

Question 14

What is the hierarchy of analytical designs for evidence of effect? (5)

Answer 14

Pyramid top

1. meta-analysis
2. trial
3. cohort
4. case control
5. cross sectional

Question 15

What are the two key approaches to studies? (2)

Answer 15

1. inductive reasoning

• observation
• pattern
• tentative hypothesis
• theory

2. deductive reasoning

• theory
• hypotheiss
• testing
• observation
• reject/not reject

Question 16

What are descriptive/inductive studies? (2)

Answer 16

• case study/series

- cross-sectional

Question 17

What are analytical/ deductive studies?

Answer 17

1. observational
   - selective sampling
2. experimental
   - selective exposure