Narrative and Systemic Review
T or F: Appraisal methods are used to evaluate included articles
False - they are not used
The % variability across studies that is due to heterogeneity and not chance
I2 test
Narrative review is considered to be ____ literature
tertiary
T or F: Systematic reviews define their search strategy (mesh, key words, limits, etc.)
true
T or F: Systematic reviews have a rigorous critical appraisal of studies that are included
true
Statistical analysis is done to complement ___ ____
visual analysis
What tests can be done to see if the studies combined were similar or consistent?
visual assessment of forest plot statistical testing - cochran Q (x2), I2 difference between studies or heterogeneity
What is the point of no difference for continuous data?
0
What is the point of no difference for discrete data?
1
How should the studies in a meta-analysis be selected? What kind of bias is this?
Based on methods - not results! Selection
How does one assesss a study for risk of bias?
Blind manner and completed by more than one person independently, to then be compared
What is the leading resource for systematic reviews?
Cochrane database of systematic reviews (CDSR)
What are the best tests (2) for the random effects model?
DerSimonian and Lard test
T or F: Meta-analysis overcomes bias
F
T or F: Meta-analysis increases sample size resulting in a decrease in power and precision
F --- increases power
What test is best for descriptive statistics?
I^2
How do qualitative and quantitative systematic reviews differ?
In the analysis of the data - statistics
How is publication bias assessed?
Inverted funnel plot
What is the most-common test for the fixed effects model?
Mantel-Haenszel test
These can help answer a clinical question that is currently unanswered by small RCTs
Meta-analysis
These take less time and they are less costly than a large RCT
Meta-analysis
what is another name for quantitative review?
Meta-analysis
Outcome is predicted according to 1 or more explanatory variables
Meta-regression
Outcome variable is effect estimate and the explanatory variables are study characteristics that might influence the effect size
Meta-regression
What are the 2 types of review articles?
Narrative Systematic
What does a P value > 0.05 suggest when assessing a publication bias?
No publication bias
T or F: meta-analysis can establish statistical significance with studies that have conflicting results
T
PRISMA stands for:
Preferred reporting items for systematic reviews and meta-analyses
Studies with a positive result will more likely be published than those with a negative result
Publication bias
What do asymmetrical inverted funnel plots suggest?
Publication bias
What are the 2 types of systematic reviews?
Qualitative Quantitative
Data is pooled Qualitative or quantitative
Quantitative (Meta-analysis)
T or F: SR can increase power to detect rare events
T - reduce need for large expensive trials
(SR/Narrative) cannot reduce systematic error
SR
____ can look for differences in effect among subgroups
SR
What do gaps in inverted funnel plots mean?
Small studies favoring the control are missing
Compares estimates of treatment effects for different subgroups or study types
Subgroup analysis
Meta-analysis is a: narrative review or a systematic review?
Systematic
Focuses on a specific topic and attempt to answer a specific question
Systematic reviews (SR)
Has a predefined research methodology
Systematic reviews (SR)
T of F: Poor studies lead to poor SR
T
T of F: Simple average would give each study equal weight which is incorrect
T
T or F: The focus of a meta-analysis is more narrow than that of a narrative review
T
T or F: Weighing gives more emphasis to studies that provide more precise estimate of effect size
T
Why should a search be detailed and include the dates and limits?
To allow one to duplicate it
What is the purpose of having 2 or more investigators participate in decisions in the systematic review process ?
To reduce error and guard against observation/interpretation bias
T or F: Narrative reviews include bias
True - the author usually supports their own hypothesis and excludes those that do not
Is publication bias a key concern for SR? why?
YES Because depends on thorough search of both published and unpublished studies to represent the population of completed studies
Does an author need to report how they assessed risk of bias?
Yes
What is the traditional test for heterogeneity?
cochran's Q (chi-squared)
Mean difference and standardized mean difference are used for ____ data
continuous
A search should be 2 things:
detailed exhaustive
Odds ratio and risk ratio is used for ____ data
discrete
critical appraisal relates to the ____ criteria. These define __.
eligibility pico
PICO is a form of: (a) internal validity (b) external validity
external validity
T or F: Pooling data is acceptable when studies are different
false - this would not make sense
T or F: Narrative review use mostly a scientific approach to how the review is conducted
false --- less scientific
T or f: In a narrative review, the search strategy along with the studies included and excluded must ALWAYS be mentioned and described
false --- not always
"What is the best estimate of the treatment effect" fixed / random effects model
fixed
considers only within-study variability fixed / random effects model
fixed
Assumes one true effect size
fixed effects model
This can be a source of unpublished literature (ex: conference abstracts, thesis)
grey lit
Random effects model incorporates _____
heterogeneity
What does a low p value <0.10 indicate? is pooling appropriate?
heterogeneity and source of heterogeneity should be explored - pooling is not appropriate
If studies are not similar enough to be pooled, they are: homozygous vs heterozygous
heterozygous
For a fixed effects model, ___ is assumed
homoegenity
What does a high p value > 0.10 indicate? is pooling appropriate?
homogeneity and increases confidence that the studies can be pooled
If studies are similar enough to be pooled, they are: homozygous vs heterozygous
homozygous
searching for unpublished information is important to avoid____ __
publication bias
(Narrative/Qualitative and or Meta) - variable and subjective; has new data produced
meta
(Narrative/Qualitative and or Meta) have statistically evaluated results
meta-analysis
This is a general summary on a given topic that does not need to answer a specific question
narrative review
Fixed effects model gives a (narrower/wider) CI than random model
narrower
Having 2 or more investigators participate in decisions in the systematic review process
observation bias
Enables the author to search for the studies for inclusion and enable the reader to apply the info
pico/eligibility
statistical summary of the data
pooled estimate
27-item checklist and flow diagram
prisma
A checklist containing a minimal list of items relevant when report systematic reviews
prisma
a standard way for authors to prepare the manuscript of trial findings for complete and transparent reporting
prisma
(Narrative/Qualitative and or Meta) have a predefined strategy that is explicit and comprehensive
qual/meta-analysis
(Narrative/Qualitative and or Meta) includes unpublished literature
qual/meta-analysis
A narrative synthesis of the data is a _____ SR
qualitative
Data is not pooled Qualitative or quantitative
qualitative
Systematic reviews can be either ___ or ___
qualitative or quantitative
A synthesis of the data with statistical analysis is a _____ SR
quantitative (meta-analysis)
"what is the average treatment effect" fixed / random effects model
random
assumes true effect size could vary from study to study
random effects model
considers both between study and within study variability fixed / random effects model
random effects model
Summaries and analyses of available information
review articles
What are the 3 types of biases in systematic reviews?
selection observation publication
Studies for systematic reviews are chosen based on specific criteria and not results to avoid ____ ____
selection bias
Determines how sensitive the results are to changes in the way meta-analysis was conducted
sensitivity analysis
_________ ___ _ is used when studies have different endpoints to measure the same clinical outcomes
standardized mean difference