Skip to main content

Systematic Review for NURS 7323

This guide covers the basics of conducting a review as assigned in NURS 7323.

The Process

Formulate Question

The systematic review pulls together studies that answer the same question; thus, it is very important to carefully consider your research question.

Put your question in the PICO format to make it searchable. Click here for a refresher.


Reviews of qualitative studies may require a different type of question. Instead of PICO, try PEO:

PEO
P Population/Problem(s) Affected users, patients, community and their symptoms, age, gender, problem
E Exposure This must be specific, like "violence in the home" or "transcendental meditation education"
O Outcome Looks at the patient's experience, for example, did their perceived quality of life improve

Or SPiDer

S Sample
Pi Phenomena of Interest
D Design
E Evaluation
R Research Type

 

Develop Protocol

The protocol sets the systematic review apart from other literature reviews, and is part of what makes it a valid research methodology. It is very important. It will include:

  • Your question
  • Where you plan to search and your search strategy
  • Inclusion criteria
    • Type of studies
    • Participants
    • Interventions/exposures
    • Outcome measures
    • Etc.
  • Data extraction plan
  • Assessment plan
  • Data synthesis plan
  • Anticipated begin and complete dates
    • Use a calendar to set a timeline
    • Include all aspects of your search in the timeline

Conduct Search

The part we've all been waiting for! It's time to search!

  • Search the databases
    • Follow your search protocol
    • Record your results in the PRISMA flow chart
  • Search the grey lit
    • Government documents may be applicable, but may not appear in the databases
    • Advocacy groups, Industry reports, and more
    • May not be applicable to your question
  • Perform hand searching
    • Not all journals are published electronically
    • Not all trial reports are accessible through the databases
    • Requires going through a journal or conference proceeding page by page
  • Contact experts
    • If researchers have unpublished studies that may be valuable to you
    • Helps combat publication bias

Select Studies

Screening your studies by title and abstract:

  1. Use Rayyan for blinded review
  2. Read through the titles and abstracts and mark include, exclude, or undecided
  3. If you decide to exclude a study, make note of the reason
  4. Once your review by title and abstract is complete, compare with the results from your blinded 2nd reviewer
  5. Send any conflicting results to a third blinded reviewer
  6. Studies that make it through this process will need to be read in full and appraised

Appraise Studies

Using a predetermined assessment tool (usually an Excel doc), evaluate the:

  1. Methodological quality of your particular study type
    • If you include multiple study types, you will need multiple assessment tools
  2. Risk of bias in individual studies

Extract Data

You will work with Drs. Puga and Richardson to create tables or forms for recording data. You may hear this referred to as a "Synthesis Table."

It's an important step because it allows "apples to apples" comparison and will more easily inform your next step, which is Data Synthesis.

Often, this step is completed in tandem with the full-text screen. As you read full-text articles to determine suitability for inclusion, you will also code them.

If you are interested in creating a meta-analysis, please consider the following:

Blinded reviewers (ideally two) will extract the data.

Analyze Results

There are 2 types of synthesis:

  • Narrative
  • Metanalysis

Narrative seeks to answer the following:

  • Does the intervention work? Why? and for Whom?
  • How can the results of all studies be described so that they will be identifiable and comparable across studies?
  • What is the relationship of the data within and between studies?
  • How robust is the synthesis?

Narrative analysis includes a "Discussion" section with the following subheadings:

  • Overall completeness and applicability of evidence
  • Quality of the evidence
  • Potential biases in the review process


Metanalysis

  • Statistically combines data from all included studies to give an overall result
  • May need to use software to ensure that data is combined appropriately
  • Requires separate metanalysis for each included study type
  • Will render as a Forest Plot
  • Also needs some discussion

Interpret Results

Divided into two sections:

  • Implications for practice
  • Implications for research

Consider the following:

  • Quality of evidence
  • Balance of benefit and harm
  • Values and preferences
  • Resource utilization

From the Cochrane Handbook