Topic 11: Critically Appraising Papers Flashcards
What are four things to ask when critically appraising a methodology paper
Advantages, limitations of the method. How does the proposed method relate to my specific scenario and how does it relate to the wider literature.
What is a frequentist approach
Treats probability as equivalent to frequency and draw conclusions by emphasising the frequency of proportions of findings in the data
What is a Bayesian Approach
Uses Bayes rule to update the probability of hypotheses as more evidence/information/data becomes available/
What does CONSORT stand for
CONsolidated Standards Of Reporting Trials
What does the CONSORT statement emphasise
Readers can only understand the design, analysis, interpretation and assess reliability of the results when there is complete transparency from authors.
When is a CONSORT checklist uploaded
With the submission of manuscript for publication of a trial.
What critique might you make for a single centre trial
Not necessarily generalisable to other centres
What might you question when critiquing eligibility criteria
Does this fully represent the intended target population.
If method of concealment is not given in the trial report, what is there a risk of
Selection bias
Should the method of generating the randomisation list be reported
Yes
Why is it not appropriate to conduct formal statistical tests to assess differences in baseline characteristics
Because given that the method of randomisation/allocation concealment was sufficient and appropriate, any imbalance would have occured by chance.
What might a critique be for a trial that uses only intention to treat analysis population
A sensitivity analysis should have been completed for using different kinds of analysis population.
What are some points to consider when critiquing a trial (question, randomisation and blinding)
Does the study address a clear question, could the randomisation procedure have been predicted, was blinding every broken
What are some points to consider when critquing a trial (method, analysis and results)
Are the analysis methods appropriate for the type of data and to answer the question. Are the methods transparent. Is standard of care or control treatment defined Are the results reliable and valid. What are the analysis methods and populations. Do the authors discuss the applicability of the results to general practice. Are there limitations of the study in terms of wider application.
What are some points to consider when critiquing a trial (the patients)
Are the eligibility criteria clearly stated and are they generalisable. What was loss to follow up like. Are the groups similar at baseline. Is the primary analysis population appropriate.
