Exam 3: Systematic Reviews and Meta-Analysis Flashcards
Difference between systemic review and narrative review
- Systemic review: identify, appraise, and synthesize all empirical evidence that meets pre-specified eligibility criteria to answer a given research question; explicit methods used to minimize bias
- narrative: qualitative summary of evidence on a given topic; informal and subjective methods to collect and interpret info (written by “expert” in field)
types of review
- effectiveness of interventions
- accuracy of diagnostic tools
- identification of prognostic factors
- methodological factors
who is Archibald Cochrane and what is he known for and why
-wrote book “effectiveness and efficiency: random reflections on health services” led to the development of the Cochrane collaboration
Cochrane collaboration
- international not-for-profit organization dedicated to promotion of clinical trials evidence and development and dissemination of SRs of healthcare interventions
- primary product: Cochrane database of systematic reviews
Cochrane review
-Def: SRs of research in healthcare and health policy that are published in the Cochrane database of SRs
types of SRs published in Cochrane reviews
- intervention reviews
- diagnostic test accuracy reviews
- methodology reviews
the purpose of a systematic review
- to identify and combine studies using explicit methods to reduce bias
- do not typically define appropriate actions or incorporate values
what is the process of conducting an SR?
state the study objective–>develop the protocol–>develop a search strategy–>conduct the search–>retrieve relevant papers–>screen and select papers that meet established criteria–>evaluate methodological quality of selected studies–>analyze and synthesize findings–>determine if stats data are sufficient for further analysis (if no, the report results of SR); if yes, analyze the effect size estimates–>report results of meta-analysis
are PICO questions used in systematic reviews
yes, used to formulate the research question
how is the research question for an SR defined?
- a well formulated question that guides the review process
- assists with determining: eligibility criteria, searching for studies, collecting data from included studies and presenting findings
is the literature review for a systemic review exhaustive?
- yes??; need to pick ones specifically for your review
- select key words
- identify resources that include relevant information (Cochrane, previously published SRs, databases available, reference checking)
what databases are searched when conducting a SR?
- search at least 3
- Medline, Embase, Cinahl, Web of Science, EBM reviews
7 types of bias’ that can influence the outcome/quality of a SR
- selection bias
- confounding bias
- allocation bias
- performance bias
- attrition bias
- detection bias
- reporting bias
- selection bias
may distort tx effects
-random allocation and concealment are essential
- confounding bias
systematic differences in factors related to intervention and outcome
- allocation bias
Systematic differences due to the way that tx groups are assemebled
- performance bias
differences in the provision of care to experimental and control groups
-blind those who receive and give care
- attrition bias
differential loss of subjects across comparison groups
-intention to treat analysis, people drop out
- detection bias
outcome differs across comparison groups
-quality assessment tools: quality assessment may provide explanations for differences in study results
- reporting bias
systematic differences between reported and unreported findings
what are the rating scales used for an SR
- Jaded scale, Pedro scale, QUADAS scale
- no gold standard, most scoring systems have not been validated
- PT-related scales should take into account: pt adherence, standardization of tx protocol, precision of performance of intervention, validity, reliability, and responsiveness to the tests and measures
how many people review the systematic review prior to its publication?
- minimum of 2 primary reviewers
- disagreement= describe if resolved by consensus or by resolution from a 3rd party
how do you read a Forest Plot?/how is it constructed?
-has 5 odds ratios (squares, proportional to weights used in meta-analysis), with the summary measures (center line of diamond) and associated confidence intervals (lateral tips of the diamond), and a solid vertical line of no effect
purpose of Forest Plot
-illustrates results of individual studies and a cumulative summary; visually understand the inconsistency in findings and large variance in several studies for this review
what is meta-analysis?
statistically combining results from 2 or more separate studies to estimate an “average” or “common” effect
-improves the precision of the estimated treatment effect
what are the two states in a meta-analysis?
Stage 1. calculation of confidence intervals for each study
- extraction of data from each individual study and the calculation of a result for each study with an estimate of the chance variation we would expect with studies like that
- Evaluate heterogeneity (chi-square test)
Stage 2: calculation of pooled average result across all studies
- give greater weight to the results from studies
- effect size, effect size index, weighting effect size
what is the relationship between a meta-analysis and a SR
-SR refers to the entire process of selecting, evaluating, and synthesizing all available evidence, while meta-analysis refers to the statistical approach to combining the data derived from an SR
What is PRISMA
Preferred Reporting Items for SRs and Meta-Analysis: a checklist for reporting them
- making sure it has all of the components
- abstract, intro, methods, results, discussion, funding