Flashcards
Experimental Studies
Cards in this deck (37)
- 1Randomised clinical trialsTreatment assigned by randomization rather than physician patience choice.
This is the standard for excellence - 2Characteristics of intervention (3)Generalizability
Strength
Complexity - 3Randomised control studies are similar toCohort studies, except for the randomization
- 4Comparison groups usePlacebo if no treatment exists
Usual care or intervention
Another intervention - 5Issues in comparison groupsUnpredictability
Placebo effect
Regression to the mean (if you do an experiment many times, results get closer to the mean) - 6Importance of randomisationCOntrolling for bias
- 7Types of random assignmentSimple
Random block
Randomization of pairs of subjects - 8Blinding/ MaskingInvestigators and/or participants continue their assignment without knowing which intervention a patient is receiving
- 9Specify outcomes before or after studyBefore the study
- 10Assessment of outcomesCOMPARE outcomes using
RR
AR
OR - 11Summarizing treatmentBy determining change in outcome resulting from the intervention
RRR
ARR
Number needed to treat (NNT) - 12Relative Risk ReductionDescribe the magnitude of the effect
Represents the percentage reduction in risk of studied outcome achieved by use of intervention
Control event rate - Experiment event rate/ Control event rate - 13Absolute Risk ReductionRisk difference in outcome between patients who have undergone one therapy and those who have not
Tells how many are spared from adverse outcome as a result of having received the experimental rather than control
Control event rate - Experimental event rate
Puts low risk or high risk into consideration - 14Number Needed to Treat (NNT)Number of patients needed to treat to prevent one patient from getting the outcome
1/(control event rate - experimental event rate)
Changes inversely with baseline risk - 15Analysis of results (2 approaches)Intention to treat
Explanatory - 16Intention to treatAnalysis according to group assigned to the patient by randomization regardless of whether they received the treatment or not i.e. regardless of compliance.
It is reflective of real world as not all patients are compliant, and it tests effectiveness - 17ExplanatoryAnalyse according to if treatment was received or not i.e. Did it make any difference.
- 18Judging the treatment's resultEfficacy
Effectiveness - 19EfficacyJudging if the treatment works in ideal settings (clinical trial setting)
Compliant patients so can use both intention to treat and explanatory trial - 20EffectivenessDoes the treatment work in ordinary real life circumstances in patient care?
Analysed by intention to treat as it puts into consideration non-compliant patients - 21Attitude that can be employedPragmatic Attitude
Explanatory Attitude - 22Explanatory AttitudeParticipants are highly selected and adherent
Intervention is strictly reinforced
Can the intervention work
Indirect relevance to clinical settings - 23Pragmatic attitudeNot strict
Does the intervention work in clinical practice
Direct relevance to those making decisions about treatment options - 24Issues after randomizationAdherence, crossovers, losses
Co-interventions
Compliance
Blinding
Are the outcomes appropriate
Are the primary outcomes presented? - 25RCT strengthControl over study situation
Random assignment
Control of exposure (timing etc) - 26WeaknessExternal validity (application to real world)
Adherence to protocols
Ethics of withholding treatment - 27Power and sample size relationshipPower is dependent on sample size
If low power and no effect, it may be that sample size was not large enough to detect a small effect. So you can keep power constant and adjust sample size
So avoid power problem by doing sample size calculation ahead of time - 28Quantify when large sample size is neededWhen there's only about 25% improvement in treatment of new drug compared to traditional. Large sample size is needed to show this effect
- 29Statistical vs. Clinical significanceBoth influenced by sample size
Statistical significance uses figures, clinical significance is judgement call.
If samples size is small, there may be no statistical significance, but clinically, the result may be important.
If large sample size, there may be statistical significance but result not clinically significant. - 30Experimental event ratea/(a+b)
- 31Control groupc/(c+d)
- 32ARR vs RRRThe smaller the event rate in control group, the larger the difference between ARR and RRR
RRR is good for estimating risks for patient undergoing a treatment
ARR is good for estimating whether benefits outweigh harm
RRR is more impressive than ARR - 33Tip 1RRR - Drug may be good regardless of high risk or low risk
ARR - To assess whether benefits outweigh harms - 34Tip 2Estimate risk of outcome without treatment to tailor clinical trial data about treatment result
- 35Tip 3100/ARR for percentage and 1/ARR for proportion
Calculate NNT from risk difference
Help estimate risk reduction to look for in therapy
For personal absolute reduction of risk for patients, you can get NNT from patients similar to yours...because clinicians sometimes misestimate baseline risks - 36RR is constant or varies across populationConstant
- 37—
Information
- Description
- Flashcards on Experimental Studies, created by sotee4u on 03/08/2015.
- Created by
- sotee4u
- Views
- 0
- Rating