Week 10. Systematic Reviews Flashcards
h-Index
author-level metric that measures productive and citation impact of the publications of a scientist/scholar
“a scholar with an index of h has published h papers each of which has been cited in other papers at least h times”
Journal Impact Factor (IF)
measure reflecting the yearly average number of citations to recent articles published in that journal.
- used as a proxy for the relative importance of a journal within its field
- higher IF deemed to be more important that lower ones
Types of SR
- literature/narrative
- scoping review
- state-of-the-art review
- systematic review
Narrative review
- summarize different 1* studies to draw conclusions with interpretations informed by the reviewers’ own experience, existing theories and/or models
- allows experts to synthesize and interpret information and form opinions/hypothesis
- most susceptible to bias
Example of narrative review
- like a mini-CPG
- practice points
- up-to-date overview from clinical experts
Scoping review
- Preliminary assessment of potential size and scope of available research literature
- Aims to identify the nature and extent of research evidence (usually including ongoing research)
- Casts a “broad net”
example of scoping review
- Informs future research
- identifies gaps in knowledge
- inventories tools, methods, or approaches
Systematic Review
literature review that involves “a comprehensive search for relevant studies on a specific topic, and those identified are then appraised and synthesized according to a predetermined and explicit method.”
Example of systematic review
- Rigorous approach
- Least susceptible to bias
- Unfortunately, often recommend more research is needed
Relationship of review
Narrative review < quick scoping review < rapid evidence assessment < full systematic review
Types of SR
- Intervention SR
- Diagnostic accuracy SR
- Prognostic SR
- Risk/Etiology SR
- Measurement tool properties SR.
- Qualitative SR
where does SR fit in the 5S Hierarchy of Evidence
Lowest level of secondary: Synthesis
Secondary (filtered resources) 1. Systems 2. Summaries 3. Synopses 4. Syntehsis Primary (unfiltered resources) 5. Studies (RCT, Cohort, Case control, Case series, case report)
*6. expert opinions: classical textbooks
Why SR?
- Explicit methods reduce bias which means that conclusions are likely to be reliable
- When a meta-analysis is done, it helps increase the precision of the answer provided by many smaller studies (i.e. shrinks CIs)
- Saves the time of healthcare providers
What are SR used for?
- Recommendations and guidelines
- Policy decisions
- Research agendas
- Individual patient care
- Patient decisions
What is a meta-analysis
- Meta-Analysis – a statistical method to synthesize or pool data from multiple studies.
- Systematic reviews ≠ meta-analysis
- Systematic reviews may use meta-analysis to quantitatively combine studies
- Meta-analyses are not always systematic (can also be done as easily on an arbitrary grouping of studies)
Components of SR
- Focused clinical question (PICO)
- Comprehensive search for evidence
- Explicit inclusion and exclusion criteria for studies
- critical appraisal of the included studies
- systematic synthesis (if quantitative then also referred to as a meta-analysis)
Steps to conducting a SR
- Assemble a team
- Ask the right question
- Conduct a systematic and comprehensive search
- Screening and selecting studies
- Data extraction
- Quality assessment (QA) of included studies
7a. Narrative Data synthesis
7b. Meta-Analysis
Steps to conducting a SR: (1) Assemble a team
- Assemble a team
- Content experts, SR method experts, statistician, searcher/librarian
Steps to conducting a SR: (2) Formulate Question
- Ask the right question
- PICO, determines the scope of project and facilitates subsequent steps of the SR
Steps to conducting a SR: (3) Conduct a systematic and comprehensive review
- find all relevant studies that answer question
- consult/work with librarian/expert searcher
Steps to conducting a SR: (4) screening and selecting studies
- Search results are compiled into a database (e.g. Refworks) and duplicates are removed.
- 2 people review ALL abstracts and select articles to minimize mistakes and bias.
- Need a clear process for resolving disagreements
Steps to conducting a SR: (5) Data extraction
- Scan each included study for the data you would need to compare and synthesize studies (populations, settings, methods, treatment effects, adverse events)
- Create spreadsheets/tables
- Ideally 2 independent reviewers extract data from each study to reduce bias/errors
Steps to conducting a SR: (6) Quality Assessment of included studies
- Not all evidence is high quality
- Quality = absence of bias
- Validated QA tools available
- Different QA tools for different study designs. (e.g. Cochrane risk of bias tool favoured method for QA of RCTs)
- 2 independent reviewers do QA for all studies
Steps to conducting a SR: (7a) Narrative Data synthesis
- Narrative synthesis is done with non-quantitative data or because the studies are too heterogeneous
Organize results as appropriate (e.g. study design, study quality, differences in population (mild, severe)
Provide conclusions for research and clinical practice
Steps to conducting a SR: (7b) Meta-analysis
- A statistical methodology to pool effect sizes across studies
- Essentially creates a new larger, more precise study
- Studies must be clinically homogenous in terms of interventions and outcomes
- Create graphs called Forest Plots or Blobbogram
Critical Appraisal of SR
AMSTAR = A MeaSurement Tool to Assess systematic Reviews’
Credibility and Applicability: 4 questions before applying results of SR to patient
- Was the SR done in a rigorous way?
- Are the included studies high quality = low bias?
- What are the results?
- Are results applicable to my patient?
