How to critique a paper- Basics

Simply go through the paper trying to answer each of these questions – that will be a very thorough critical analysis. From our experience the area most presenters will experience problem is the use of statistics. It is useful to acquaint yourself with this weak area of many doctors.

  • Was the title of the study appropriate? Does it reflect the objective of the paper
  • Was the abstract accurate yet concise? Has the authors stated the conclusions properly including unfavourable findings if any?
  • Authors’ experience and qualifications
  • Place of study- multicentric studies are thought to be better quality.
  • Length of study- longer the length, better the study.
  • Any financial grant from industry which might bias the investigator? Are other conflicts of interest clearly stated?
  • Was the aim of the paper clearly stated?
  • Were the outcome measures or endpoints (primary and secondary) clearly defined and measurable?
  • Was the study well designed? All the steps, devices, questionnaires, investigations etc should be defined in very clear terms so that if any other group wants to reproduce the study there is no confusion or chance of guesswork.
  • Was there any bias and methods to control bias?
  • Was the research well carried out? Randomised trials can be assessed using the CONSORT statement.
  • Was ethical approval obtained? Adverse effects monitored and reported?
  • How were patients selected? Is there clearly defined inclusion and exclusion criteria?
  • Was consent obtained?
  • Were the baseline characteristics of both groups matched?
  • How was randomisation performed?
  • Were appropriate controls used- either placebo or gold-standard?
  • Were the study groups and the investigators blinded as to allocation?
  • Were patients excluded, lost to follow up, did they die, refuse participation, or have notes missing?
  • Were appropriate statistics used?
  • Was the study powered for the answer it was seeking to find? That means appropriate sample size.
  • Were the study limitations discussed?
  • Did authors meet the objective they set out for?
  • Are the patients similar to those seen in your unit? Can you extrapolate these results to your patients? A study done with 25yr old healthy volunteers might not be applicable to an elderly population with co-morbidities.
  • Clinical as opposed to statistical significance?
  • Did the study persuade you to change your mind?
  • In bibliography- is there a lot of reference to tertiary literature rather than primary?

Last modified: September 2, 2010

Post a Comment