lecture 1: RCT overview

Question 1

Fundamental question in RCT

Answer 1

How do we know if a treatment works or not (or is even harmful)?

Question 2

What do you need/ acknowledge to answer whether treatment works?

Answer 2

1. comparison (Spontaneous improvements?)
2. regression to the mean: start treatment at the worst moment
3. might be a selective group of patients

Question 3

Essence of a RCT

Answer 3

compare 2 groups and distribute people at random

Question 4

Advantages of RCT

Answer 4

controls for confounding, factors that we know and measure, but also the ones that we don’t know about or don’t measure

Question 5

confounding

Answer 5

there is an association between A and B but this can be explained by a third variable (e.g. psychological treatment and symptoms –> e.g. time, therapeutic alliance…)

Question 6

What types of questions does a RCT answer?

Answer 6

• effectiveness of something (NOT e.g. mechanism of change)

- -> cornerstone of Evidence Based Medicine

Question 7

Types of RCTs

Answer 7

1. Explanatory (or efficacy) vs. pragmatic (or effectiveness)
2. Superiority vs. non-inferiority
3. Individual vs. cluster

Question 8

Explanatory trial

Answer 8

=efficacy trial

• high internal validity
• conditions are as controlled as possible (laboratory)
• group as homogeneous as possible, adhere to treatment, the very best therapist
• usually if we don’t know much about the treatment yet (to see if it has any effect at all)

Question 9

Pragmatic trial

Answer 9

=effectiveness trial

• high external validity
• conditions are as similar to real world as possible
• heterogeneous group, e.g. with comorbidities, drop-out, average therapists

Question 10

Which one should I choose? Explanatory vs. pragmatic

Answer 10

depends on the rq!
but in psychotherapy it is not possible to create laboratory conditions, but also not completely routine practice (Because not every single person wants to take part..), you lean more towards one or the other though; you do a mixture

Question 11

superiority trial

Answer 11

treatment X is better than Y

Question 12

non-inferiority trial

Answer 12

treatment X is not worse than treatment Y (if there are other benefits, e.g. costs, more acceptable, shorter…)

Question 13

equivalence trial

Answer 13

treatment x is not better but also not worse than treatment Y

Question 14

individual vs. cluster

Answer 14

does not depend on rq, but on the level of randomization (usually you randomize individuals)

Question 15

Cluster

Answer 15

randomization on a higher order level, such as therapist or research center (i.e. cluster)
reasons:
- providers wants this (center only offer 1 treatment), pragmatic argument, not very good
- not possible to randomize individually (e.g. classroom interventions)
- rarely: high risk of contamination (e.g. only one therapist works at a clinic and only offers one therapy)

Question 16

Problems with cluster RCT

Answer 16

• patients within one cluster might be different from those in another cluster (wealthy area vs. rich area –> confounding)
• 2 arms might be disbalanced (patients in treatment center randomized to control: benefit of participating in RCT? Extra burden: filling out questionnaires)

Question 17

CONSORT statement

Answer 17

• consolidated standards of reporting trials (report!!)

- standard for high quality research

Question 18

SPIRIT guidelines

Answer 18

-designed to help with protocol papers

Question 19

good research question

Answer 19

PICO:
P: patient or population of interest (concise)
I: intervention (specific)
C: comparison (specific)
O: outcome (specific)

Question 20

primary outcome

Answer 20

outcome on which you decide whether it is effective or not