Evidence Based Medicine Flashcards
What is a confounder?
A confounder is a factor associated with the exposure and is independently a risk factor for the disease
What is “concealed allocation and why should randomisation have “concealed allocation”?
This means that the randomisation sequence is prepared by someone who is not entering patients. Now often done by 3rd part computerised randomisation.
It is done to ensure observers cannot bias - so do not know what group person will be in until after entering trial (or after finish if bias).
When is bias difficult?
Surgical procedures
Psychotherapy vs antidepressant
Alternative medicine
Lifestyle interventions
Prevention programs.
What is an explanatory trial “as treated” analysis
It analyses only those who completed follow-up and complied with treatments.
Compared the physiological effects of the treatments.
But, looses the effects of randomisation.
Non-compilers likely to be systematically different form compilers - selection bias and confounding
What is a pragmatic trial / “intention to treat” analysis?
Analyses according to the original allocation to treatment groups (regardless of whether they completed follow-up or adhered to treatment)
Compared the likely effects of using the treatments in routine clinical practise.
Also, it preserves the effects of randomisation - minimal selection bias and confounding.
What is the difference between “as-treated” and “intention to treat”
“As treated” - analyses tend to give larger sizes of effect (more impressive results)
“Intention to treat” - analyses tend to give smaller and more realistic sizes of effects.
Clinical trials should normally be analysed on an “intention to treat” basis
What is a confidence interval?
A range in which 95% of the ranges will contain the mean. Means there is a 95% chance that the true value will lie within those values.
For a confidence internal to be statistically significant it must not include 1.
The bigger the sample size, the smaller the confidence Interval.
What is the difference between deductive and empirical scientific approaches?
Deductive:
Identify the basic process
Deduce the best procedures
Apply to the clinical situation
Empirical:
Identify the basic process
Postulate alternative approaches
Test ideas experimentally.
Empirical is better.
What issues should you look at when assessing an RCT?
Randomisation - concealed allocation to randomisation sequence?
Was there blinding of outcome assessment?
Did they conduct an ‘intention to treat’ analysis?
Also: Inclusion / exclusion criteria Baseline comparability Defined primary outcome Good follow-up and of adequate length Simple size calculation
What is the difference between relative risk and absolute difference?
Relative risk of 1 = no difference
Done as % of the other
Absolute difference of 0 = no difference
Is literally one value - the other value
What is the Cochrane collaboration?
Cochrane is a British charity formed to organise medical research findings so as to facilitate evidence-based choices about health interventions faced by health professionals, patients, and policy makers.
What are the problem with using RCTs?
Not everything needs to be tested with an RCT (e.g. replacing blood volume after a massive haemorrhage)
Not everything can be tested with a RCT (e.g. smoking)
Not everything has been answered yet
What is PICOS?
A way of breaking down your search when searching for articles
P - population or patient group I - Intervention or invest Investigation considered C - comparator O - Outcomes considered S - Study design (may not need this)