L4 - Clinical Trials

Question 1

Define CT

(note: CT aka intervention trials)

Answer 1

A resh study where HUMANS are prospectively assigned to ≥1 health-related interventions to evaluate health outcomes

Question 2

Describe CT

Answer 2

• Investigator assign exposure with one tx and ctrl grp.

- Can be random (RCT) or non-random (nRCT)

Question 3

Purpose of randomisation

Answer 3

1. Decrease selection bias (random chance for subject to be assigned to each group)
2. Decrease Cfing by ensuring similar baseline characteristics (hence any difference is only caused by intervention)

Question 4

Purpose of blinding

Answer 4

1. Decrease information (observation) bias

Question 5

Distinguish between allocation concealment and blinding

Answer 5

• Allocation concealment: BEFORE expt, prevent investigators from unconsciously choosing assignment for participants
• Blinding: DURING expt and analysis, ensure different parties unaware of intervention

Question 6

Types of blinding

Answer 6

1. Open-label (None)
2. Single Blinding: Participants blind
3. Double Blind: Participants, investigators
4. Triple blind: Participants, investigators, data anlysers

Question 7

Distinguish between Parallel and Crossover CT design

Answer 7

1. Parallel: Sample receive separate tx and outcome evaluated separately
2. Crossover: Sample receive separate tx, evaluated, then washout and receive the OTHER tx and evaluated again.
   - Order of tx is randomised

Question 8

What is the benefit of crossover CT?

Answer 8

1. Less between-subject variability (since same subject receiving the two interventions)
2. Decrease the total sample size required (since you just need “one” group)

Question 9

Describe Factorial design for CT. State one benefit of factorial design

Answer 9

• Sample randomised to ≥2 interventions, both separately and in combination against the control

Benefit: Allows evaluation of interactions between two interventions as well

(E.g. Drug A, Drug B, Drug A + B, no tx)

Question 10

What are the problems of loss to follow-up (f/u) in a trial?

Answer 10

1. Affect internal validity (you calc power already but sample size keep dropping)
2. Introduces bias, especially differential rates of loss
   (E.g. Tx group < Placebo grp, tx grp drop due to SE, while placebo no se and no one drop)
3. Decreases power of study (due to lower sample size, may even go below threshold)

Question 11

How to minimise LTFU?

Answer 11

1. Reminders
2. F/u calls
3. Exclude those who are likely LTFUs before study commence
4. Recruit 10-20% more sample than you require (does not minimise LTFU but reduce damage done by it)

Question 12

Describe the primary analysis method of RCT.

Answer 12

Intention-to-treat (ITT) analysis

• Analyse subjects as assigned
• Helps preserves comparability obtained by randomisation (preserve baseline)

Question 13

Describe the two secondary analyses methods

Note secondary analyses aka non-randomised comparisnons

Answer 13

1. Per protocol (PP) analysis: - - Only analyse fully compliant subjects (and drop non-compliant)
2. As-treated analysis
   - Analyse according to tx received
   - E.g. participant drop out of tx grp, we can analyse that subject as placebo instead

Question 14

What is an example of standard reporting practice?

Answer 14

CONSORT Statement:

- Consolidated Standards of Reporting Trials

Question 15

List and describe other considerations of CTs

Answer 15

1. Clinical Equipoise: If got current useful drug, cannot give placebo
2. Data safety and monitoring Board
3. HSA requires CT registration

Question 16

Describe the strengths of CTs

Hint: stems from design of expt

Answer 16

1. Randomised = less selection bias + Less CF

2. Blinding = less info (observation) bias

Question 17

Describe the disadvantages of CTs

Answer 17

1. Ethical issues: may require OS instead
2. Costly and less feasible
3. Issue of compliance and LTFU
4. Limited generalisability
   - E.g. long-term safety cannot profiled
   - I/E criteria causes results to only be applicable for defined ppn