lecture 18 - critical appraisal I Flashcards
describe phase 1 and 2 clinical trials
Phase 1 (first in humans)
typically 20 – 100 patients, several months to complete
health volunteers / paid participants
ADME and acute toxicity
Phase 2 (first in patients)
small scale treatment groups (RCTs)
typically 100 - 200 patients, up to a couple of years to complete
basic clinical effectiveness, ADME in patients
describe phase 3 clinical trials
Large-scale randomised controlled trials (RCTs)
Typically many 1,000 patients
Several years to complete
Basis of clinical usefulness
Detailed clinical effectiveness
Acute and chronic side-effects
Evidence for marketing licence
describe retrospective versus prospective studies
retrospective is select two or more groups with different outcomes and look into their past for different behaviours. always look without an intervention
prospective is set up 2 or more groups with different behaviours. monitor into the future to look for specific outcome,s groups can be chose for the behaviour (non-interventional) and behaviour can be imposed on a specific group.
what are the 6 types of clinical trial design ?
case studies - retrospective study individual patients in phase 1 or 2
case control studies retrospective study of groups of patients, phase 1 – 4, no intervention
Cohort study
prospective study of groups of patients, phase 3 – 4, no intervention
Randomised clinical trials
prospective study of groups assigned treatments, phase 3 – 4
Practice guideline/systematic review
review of published literature, phase 1 - 4
Meta-analysis
numerical combination of data from multiple phase 3 / 4 studies
what scalar continuous data ?
Indication of effect /side effect in each patient
based on a physical measurement
Can be physiological measures
weight, BMI, blood pressure, blood glucose
Can be rating-scale scores
pain index, depression index
Group data presented as mean (with confidence interval
what is binary (dichotomous) data
Binary (yes/no) response in each participant
a “responder” is someone who meets a predefine measure of outcome
Is a therapeutic/side-effect effect seen?
of at least a predefined magnitude (who determines this?)
Gives a simple measure of intervention outcome
the Responder Rate
simplifies comparison many different clinical trials
what does rr of 0.7 mean
An RR = 0.7 means that 70% of participants met the condition for being classed as a “responder”
what is confidence interval ?
Confidence interval (usually 95%)
“there is a 95% probability that the true population value, that I do not know precisely, lies within this stated range”
determined by sample size (larger sample, smaller confidence interval)
Standard error
less commonly quoted in clinical trials
more complex to understand
“a measure of the statistical accuracy of an estimate, equal to the standard deviation of the theoretical distribution of a large population of such estimates”
why does a statistical test fail?
There is no difference
there is no elephant-in-the-room
no amount of data will reveal its presence
The test was not rigorous enough
test based on insufficient data
causes the confidence interval to be too large
there is an elephant-in-the-room
you can’t see it, because you haven’t looked hard enough
what is the difference between statistical significance and clinical significance
Statistical significance
a purely mathematical concept
mathematical proof of a difference in a single outcome between two interventions
can be either “effect” or “side-effect”
Clinical significance
a medical, sociological, economic, moral, ethical concept
must be some statistically significant justification
depends on effect size, side-effects, cost and many other factors
