Evidence Based Medicine Flashcards
What is a confounding factor?
Confounder is associated with exposure but not a consequence of exposure. Also associated with outcome independent of exposure
Types of selection bias
Recruitment of unrepresentative sample population
- Sampling bias
- Response bias
Attrition bias
Number of dropouts differ significantly in different arms
Reduced with intention to treat analysis
What is a cohort study?
- observational study
- group of ppl followed over time to establish incidence of a condition
- identify risk factors for certain condition
- prospective or retrospective
What are the advantages of a cohort study?
- demonstrate an appropriate temporal relationship between exposure and outcome
- estimate incidence rates in exposed and non-exposed groups
- can assess multiple outcomes
- can study uncommon exposures
- insight into latent and incubation periods for communicable/non-communicable diseases
What are the disadvantages of cohort study?
- require large sample size
- long follow-up periods
- frequent re-evaluation of exposure
- portion of cohort lost to FU
- outcome misdiagnosed/misallocated if diagnostic criteria changed over time
- outcome assessment vulnerable to diagnostic suspicion bias
What are the disadvantages of cohort study?
- require large sample size
- long follow-up periods
- frequent re-evaluation of exposure
- portion of cohort lost to FU
- outcome misdiagnosed/misallocated if diagnostic criteria changed over time
- outcome assessment vulnerable to diagnostic suspicion bias
What is a CONSORT statement?
Consolidated Standard of Reporting Trials
set of recommendations for papers reporting the results of randomised clinical trials.
What guidance does the statement provide?
- 25 item check list
- template participant flow diagram
- guidance for each section of paper (abstract, intro, methods, results, discussion, other information
What is a PRISMA statement?
Preferred Reporting Items for Systematic Review and Meta-Analysis
Set of evidence-based guidelines intended to improve quality of reports for systemic reviews
What guidance does PRISMA provide?
- 27 item check list
- Template flow diagram diagram to describe the identification and inclusion/exclusion of trials/studies in the review.
What is impact factor?
Measure of a particular journal’s relative importance in its field
How do you calculate impact factor?
Number of citations in a given year / number of articles published
Any other bibliometrics similar to impact factor?
- Immediacy index
- Cited half-life
What is h-index?
a simple metric that provides a guide to an individual researcher’s impact on the field
It is the rank of the paper at which position in ranking equals citaiton count
Outline 1 a, b, c of oxford system for studies of intervention
1a: systematic reviews or RCT
1b: high quality RCTs with narrow CI
1c: all or nothing RCTs
Outline 2 a, b, c of oxford system for studies of intervention
2a: systematic review of cohort studies
2b: individual cohort study or low quality RCT
2c: Ecology studies
Outline 3 a, b of oxford system for studies of intervention
3a: systematic reviews of case control studies
3b: individual case-control studies
Outline level 4 and 5of oxford system for studies of intervention
4: case series, poor quality cohort or case control study
5: expert opinion
Alternative systems to classify level of evidence
NICE
SIGN (Scottish intercollegiate guideline network)
What is a meta-analysis?
Use of statistical methods to combine results of individual studies
What is systematic review?
Literature review conducted according to strict protocol to identify studies relevant to question under consideration
What elements comprise a good systematic review/ meta-analysis?
- well-constructed clinical question in PICO terms
- objective inclusion and exclusion criteria
- Key outcomes defined
- Structured search strategy
- Efforts to obtain missing data to avoid publication bias
- Data should be extracted by two or more observers independently using a prepared extraction pro forma
What are commonly used pooled outcome measures in meta-analyses?
- Weighted mean difference: continuous variables
- Pooled odds ratio: categorical variables in interventional studies
- Risk ratio/ relative risk: categorical variables in epidemiological studies
What are commonly used pooled outcome measures in meta-analyses?
- Weighted mean difference: continuous variables
- Pooled odds ratio: categorical variables in interventional studies
- Risk ratio/ relative risk: categorical variables in epidemiological studies
What are the advantages of meta-analysis compared to traditional literature reviews?
- evaluation for bias
- greater statistical power to detect effect over single study
- can control between study variation
- more reliable synthesis of literature, reduce info overload
- less likely influenced by local factors
Disadvantages to Meta-analysis
- vulnerable to publication bias
- pooling of several small studies may not predict results of large study
- unreliable results if source studies poorly designed
What is publication bias?
Studies with positive results more likely to be published than those with negative results
How is publication bias detected?
Using a funnel plot
Describe a funnel plot
X-axis: sample size
Y-axis: effect size (standard error of effect estimate)
How to identify publication bias from funnel plot?
asymmetric funnel
Possible causes of asymmetrical funnel plot
- Publication bias
- Heterogeneity
- Poor methodology leading to inflated effects in smaller studies
What is a forest plot
graphical representation of data from a meta-analysis
Describe the format of a forest plot
Box: effect estimates form single studies
Diamond: pooled result
Horizontal line: confidence interval
Vertical line: line of no effect
What is stratification?
a method of ensuring an equal distribution of key confounding factors between arms of a randomised trial
