The fundamentals of critically appraising an article

Abstract

CASP has specific checklists to use for critically appraising randomised controlled trials, systematic reviews, qualitative studies, cohort studies, diagnostic studies and case control studies. 3 ¾ The research question should be concise and focused. This should be clearly highlighted in the title and abstract. In particular, the relevant population, intervention/exposure/risk factor, comparisons/controls and the outcome of interest should be stated well. ¾ This question will enable you to understand whether the correct and most appropriate study design has been chosen to answer the question intended: 4  Randomised controlled trials: patients are randomly allocated to experimentally receive or not receive an intervention. This is the best study design to assess the effectiveness of a treatment.  Cohort study: follows up a particular cohort of people of interest who may have an exposure to see whether they obtain the designated outcome. This group's outcome would be compared with a control group without the exposure. This study allows observance of temporal associations.  Cross-sectional study: collects data at one point in time to give understanding about a particular feature of the intended population. Population of interest ¾ The inclusion and exclusion criteria should be clearly stated. Any specific details should be included. 4 ¾ Any measures taken to reduce bias will be found in the methods section. ¾ Any conclusion may only be valid for the population studied. But have any groups of the studied population been excluded? This would be an example of selection bias. Methods ¾ Details about the intervention or outcome data and how they were collected, and collection accuracy should be carefully analysed. Information about how the endpoint was measured and categorised should have been included. ¾ Confounding factors can introduce bias in any study, especially observational studies. These are factors that are independently associated with both the exposure and outcome variables. These should have been considered and any relevant adjustments considered to minimise their effect. Depending on the study question, confounding data may have been collected, so it is worth bearing this in mind throughout your appraisal. 4 ¾ Details of data not collected or presented should be provided with an appropriate explanation.