Clinical Trial Flashcards
Randomisation
The chances of bias are reduced if subjects are randomly allocated to treatments
This should provide groups that are equal in terms of potential confounding factors in the baseline characteristics
Stratified randomised subgroups
Groups based on characteristic
Blocked randomising subgroup
Grouped by time
Clustered randomising subgroup
Grouped by geographical
Open label clinical trial
The doctor and patient know which drug or vaccine is being administered
Opens up the potential for significant bias -> but sometimes it is difficult to conceal allocation from doctor or even patient
Single blinded clinical trial
The patient doesn’t;t know which treatment they are getting
Double blonde clinical trial
Neither the doctor nor the patient knows which treatment the patient is getting -> this is what the aim should be
Controls
Necessary to assess the impact of any treatment on health -> control group should be exposed to a comparator treatment
Can be an inactive placebo or an active control
Historical controls
Are unreliable
Outcome measurements
Objectively measured -> not open to interpretation
Clinically relevant -> they should be outcomes that are direct ally relevant to patients
Surrogate end-points
Used if end points are hard to assess -> investigators make assumption that that surrogate end point can be used to extrapolate to likely impact on later clinical outcomes
Equipoise
Genuine uncertainty over which of the two treatments is better
Critical appraisal
Process of systemically reviewing clinical research
Recruitment
Allocation
Maintenance
Measurements
Weaknesses in clinical trials
Poor randomisation Baseline inequalities in groups Maintenance of study groups unequal Inadequate blinding Subjective outcome assessment High drop out rates Subgroup analysis Underpowered trial Author bias in the interpretation of results
