Skip to Main Content

Systematic Reviews

Five steps to conducting a systematic review

 

The following shows the Five Steps to Conducting a Systematic Review  reproduced from the paper by Khan et. al. 

Khan, K. S., Kunz, R., Kleijnen, J., & Antes, G. (2003). Five steps to conducting a systematic review. Journal of the Royal Society of Medicine, 96(3), 118-121. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC539417/ 

 

(1) Framing questions for a review

Questions should always be

  • Clear
  • Unambiguous
  • Structured

Usually, a protocol for a systematic review (plan or set of steps to be followed in a study) will only be modified if it becomes clear that there are alternative ways of defining your population, intervention, outcomes or study designs.

 

(2) Identifying relevant work

Involves extensive searching of a wide range of online or print resources (e.g., MEDLINE, Embase, The Cochrane Library,etc. ) without language restrictions.

How to identify what studies to keep and what to reject?

  • Define eleigibility criteria for study selection from your review questions
  • Record all reasons for inclusion and exclusion

[Useful article: Selecting Studies for Systematic Review: Inclusion and Exclusion Criteria]

 

(3) Assessing the quality of the studies 

Note that study quality assessment is relevant to every step of a review.assessment. For example, Question formulation (Step 1) and study selection criteria (Step 2) should describe the minimum acceptable level of design. 

Studies included in the review should be subjected to quality assessment using the critical appraisal guides and design-based quality checklists (refer to the Critical Appraisal tab). Quality assessment is used to:

  • Explore heterogeneity and informing decisions regarding suitability of meta-analysis (Step 4).
  • Help in assessing the strength of inferences and making recommendations for future research  (Step 5)

(4) Summarizing the evidence

  • Tabulation of study characteristics, quality and effects
  • Statistical methods for exploring differences between studies and combining their effects (meta-analysis)

 

(5) Interpreting the findings

  • Check for any possible risk of bias

  • Check heterogeneity to help you determine if your overall summary can be trusted. If you have doubts, then check the effects observed in high-quality studies to help generate inferences.

  • Grade recommendations by reference to the strengths and weaknesses of the evidence