Systematic review Flashcards
Systematic review
- scientific investigations in themselves, with pre-planned methods and an assembly of original studies as their “subjects.”
- They synthesize the results of multiple primary investigations by using strategies that limit bias and random error.
- These strategies include a comprehensive search of all potentially relevant articles and the use of explicit, reproducible criteria in the selection of articles for review.
- Primary research designs and study characteristics are appraised, data are synthesized, and results are interpreted.
**BEST TYPE OF EVIDENCE!
Meta-analysis
a subset of systematci reviews: -A statistical synthesis of the data from separate, but similar (i.e. comparable) studies, leading to a quantitative summary of the pooled results -pooling of the data across studies
systematic vs narrative reviews
narrative: subject to bias
Foreground information
- info that is good for us to know, what interventions work well with each patient population/ what’s best used to treat a patient
Elements of a system review
- PICO question
- consider inclusion and exclusion criteria to assess relevance to your pt
- literature search
- relevant dtatbases identified
- unpublished studies are considered
- non-english studies are considered
- review to titles –> abstracts
- look at thousands of titles, smaller number of abstracts, even smaller number of full articles
- Articles are selected for review
- investigators retrieve and review potential articles
- apply inclusion and exclusion criteria to articles to assemble the final set
- typicalyl involves multiple reviewers
- Data abstraction
- use PICO criteria
- methodologic quality appraisal
- consensus on validity of studies reviewed
- heterogenous or homogenous results??
- heterogenous: findings across studies are inconsistent
- homogenous: findings across studies is consistent
- Data analysis:
- data is pooled if appropriate
- forest plot is used
forest plot
- have symbols that represent the point estimate (each of the top four lines represents studies)
- the line represents the confidence interval (null value is one)
the line either crosses or does not cross the line, ones that cross the line are not significant.
- the size of the symbol is related to the sample size,
- small sample size=large confidence intervals.
Publication bias
- tendency to publish studies with significatn positive findings with large effects
- assessed by a funnel plot
funnel plot
- x-axis shows the effect sixe or odds ratio, the y-axis as you go up the sample size gets larger.
- Shows evidence of publication bias. Assymetrical suggests publication bias, symetrical suggests no publication bias
- may be useful to assess the validity of meta-analyses
- will look like an inverted triangle if there is no publication bias
heterogeneity
= Inconsistency of findings from multiple: i.e. some positive, some negative, some null.
Expressed by a coefficient of heterogeneity (e.g. “Q”) and its p-value.
- If “significant” (small p-value), differences in findings are not likely to result from chance. If this is the case you cannot pool the data.
- Instead variation in the findings is due to true differences in the studies (i.e. heterogeneity)
Appraise a systematic review: Is the systematic review valid?
- Does the methods section describe:
- method for identifying and including all relevant trials?
- assessment of individual study validity?
- Were the results consistent across studies?
- Look for heterogeneity statistic
- Was publication bias assessed?
- Look for assessment based on “funnel plot”
Appraise a systematic review: is it important?
Is the data pooled?
If so, does the analysis indicate significant effect(s) of the treatment of interest?
- Effect size
- Number needed to treat
appraise a systematic review: clinical utility
Do the findings of the study apply to my patient ?
What is the population to whom the findings apply?
Do the findings apply to a broad or narrow clinical population?
