Intro to RCTs

Question 1

What are observational studies?

Answer 1

• no intervention
• observe and record behaviour, Sx, attitudes

Association between risk factors and disease/outcomes
Cohort, cross sectional studies

Question 2

What are experimental studies?

Answer 2

research intervenes deliberately to influence events
records effect of intervention
- effectiveness of new drugs, Rx (comparative)
- RCTs, cross-over trials

Question 3

What 2 types of outcomes should be considered in a RCT?

Answer 3

• positive: good
• negative: bad

ideally, following I want to increase the abundance of positive ones and reduce negative ones

Question 4

What kind of study design is RCT?

Answer 4

parallel design: patients randomised to a group, which are studies in parallel

then compare outcomes between Intervention and Control groups

Question 5

What controls are used in a RCT?

Answer 5

• Standard Rx: active or positive control
• Placebo (negative control)
• No “additional” Rx (unlikely)

these are collectively known as the CONCURRENT CONTROLS: they reflect current natural epidemiology

Question 6

What is the placebo in the RCT?

Answer 6

referred to as “treatment”
no active therapeutic ingredients
similar appearance and taste as active Rx
- administered in same way

Question 7

What music new Rx or therapeutic regime RCT include?

Answer 7

a control group: to know how the condition would proceed normally (without new intervention)

allows comparison of how much recovery is down to new drug vs. natural improvement of condition

Question 8

What is the aim of sampling for any study?

Answer 8

to minimise SELECTION BIAS

e.g. systematic difference between population and sample [characteristics]

Question 9

What is the purpose of using random allocation to group the participants?

Answer 9

removes ALLOCATION BIAS and minimised confounding

Question 10

What is the difference between ALLOCATION vs SELECTION bias?

Answer 10

ALLOCATION: bias using characteristics that designate participants to either control or intervention group (certain characteristics for control, certain for intervention)

SELECTION: bias made when choosing the initial sampling population. Results in systematic differences between sample and population and can make the study less representative

Question 11

How does use of a placebo aid RCT design?

Answer 11

ENSURES BLINDING

• minimises ASSESOR BIAS
• minimises RESPONSE BIAS

(all needed to be able to infer causality)

Question 12

What is the most ideal situation with sampling a population? Why is this not generally done?

Answer 12

investigate the ENTIRE population

all adults meeting inclusion criteria, potentially infinite

-> v. expensive and time consuming

Question 13

What are the methods of sampling a population?

Answer 13

• convenience sampling (people taken as they are referred or present)

Question 14

What is a method of making the sample more representative of the population?

Answer 14

using patients from multiple centres or countries or regions

Question 15

What the importance of making the sample representative of the general population?

Answer 15

affects ability to GENERALISE study results to rest of population
If this is not possible to do, then there is SELECTION BIAS present

Question 16

What is a bias?

Answer 16

= systematic difference
(not down to chance, but also may not be deliberate)
e.g. study contains more women or >65yo individuals than general population
this may be influenced by other non-deliberate factors: people who are more likely to volunteer after selection

Question 17

What is the benefit of random allocation?

Answer 17

probability of an individual being allotted to Rx group is equal

eliminates allocation bias

minimises confounding (there should be no differences in baseline characteristic between groups to explain differences in outcomes

Question 18

How does random allocation facilitate blinding?

Answer 18

• requires placebo

- double blind

Question 19

What is the benefit of performing a RCT with double-blinding?

Answer 19

= group allocation blinded to researches and participants

reduces ASCERTAINMENT BIAS: - Response bias (patients)
- Assessment bias (researchers)

Question 20

What are the advantages of using a placebo?

Answer 20

• RESPONSE BIAS
  participants may alter responses if they know which group they are in
• may report responses to please researchers
• difference between patient response and “truth”
• ASSESSMENT BIAS
  researchers may bias outcomes: subjective interpretation to favour their theory
• LOSS TO FOLLOW UP
  participants less likely to drop out since they are unaware if they’re in the placebo group or not

Question 21

What is the ‘TREATMENT RESPONSE’?

Answer 21

the response in the outcome for the intervention group

components: natural epidemiology of condition, placebo effect PLUS treatment effect

Question 22

What is the ‘PLACEBO RESPONSE’?

Answer 22

response in outcome for the placebo group

components: natural epidemiology of condition PLUS placebo effect

Question 23

What is the empirical difference between the ‘Treatment’ and ‘Placebo’ responses?

Answer 23

the inclusion of the ‘treatment effect’ in the treatment response

this is the pharmacology (of the intervention)

Question 24

What is the Hawthorne effect?

Answer 24

It is a non-specific treatment effect;

Change in behaviour as motivational response to interest, care, or attention received through observation and assessment.

Question 25

What is the PLACEBO EFFECT?

Answer 25

NOT the placebo response FYI

psychological effect, seen especially in subjective outcomes

Components:

• response to investigation and therapeutic ritual
• subsequent response to observation and assessment
• response to doctor-patient interactoin

Question 26

What effect is the Hawthorne effect a part of?

Answer 26

part of the placebo effect

Question 27

How is the treatment effect calculated?

Answer 27

= pharmacological effect (of intervention)

= Treatment response - Placebo response

Question 28

What component in present in all RCT groups response?

Answer 28

• natural epidemiology of condition
• placebo effect

difficult to quantify these effects and separate one from the other

Question 29

For valid consent, what should be included?

Answer 29

[information should be understood]

communication of

• risk
• random allocation
• use of placebo
• pros/cons of Rx

Question 30

Why is informed and valid consent important for RCTs?

Answer 30

participants need to make an “INFORMED DECISION” as to whether or not they want to be involved

this can only be done if they are provided with and fully understand the info on all its aspects

Question 31

What is a cross-over trial?

Answer 31

derivative of the RCT design
patients receive both Rx (one could be a placebo) but ORDER of Rx allocation is random

washout period exists between 1st Rx and 2nd Rx

Question 32

What is the role of a WASH OUT period in the cross-over trial design?

Answer 32

allow any residual effect of previous Rx to be removed before 2nd Rx

Question 33

What is the benefit of receiving the Rx in a random order in a cross-over trial?

Answer 33

minimises an ORDER EFFECT

calculate: control Rx - Inv Rx
across participants and then average

AVG: removes the variation from within an individual

Question 34

When can Cross over trials NOT be used?

Answer 34

for conditions that are cured by Rx

must be chronic or Rx only relieves Rx short term

e.g. pain relief efficacy in cancer
pain relieved NOT cured

Question 35

Can blinding occur in a RCT where the placebo is no Rx?

Answer 35

no because blinding is dependent on randomisation of Rx (but for this to work there must be Rx for both groups (placebo must be a treatment, just not active)