Week 7 (lecture 6) Flashcards
(14 cards)
Differentiate systematic review and meta analysis
Systematic Review
Qualitative
Meta-analysis
Quantitative: aim is to produce a single estimate of treatment effect across included studies
But There is overlap
Systematic review: Define population, intervention, and comparator
Population: includes both the nature of the condition being studied, and the underlying characteristics of the population such as age, sex, race, previous treatments, early versus advanced disease, and other risk factors, such as other common conditions or current medications.
Intervention: refers to the drug, medical procedure, medical device, or other healthcare intervention that is being assessed. Considerations about the intervention include dose, duration, frequency, and intensity.
Comparator: outlines whether the intervention is being compared with placebo, usual care with another treatment, or is being added to an existing treatment. Considerations about the comparator also include dose, duration, frequency, and intensity.
Systematic review: Define outcome, timing, and setting
Outcome: is the endpoint measurement, which ideally should reflect a true health outcome that a patient can touch or feel, rather than an intermediate, surrogate endpoint. Considerations in defining outcomes include validation of the measure, sensitivity to change, relevance to the disease, and how the outcome data is collected.
Timing: reflects when the outcome will be measured, and how long measurement will continue. Outcome measurement must consider how responsive the population may be to intervention over time.
Setting: reflects aspects of how healthcare is being delivered in the population, including access and delivery models, such as urban academic medical centers as opposed to rural primary care.
What are steps 1-7 of a systematic review?
Step 1: PICO Question
Step 2: Inclusion / exclusion criteria & outcomes
Step 3: Literature Search
Step 4 & 6: Review Abstracts
Step 5 & 7: Review full-text articles
Describe step 2 of systematic review
1) Include
Peer-reviewed journals
Search terms: “dance-fever” -/+ “honky-tonkitis”
English language
2) Exclude
Greater than 10 years since publication
Observational trials???
3) Outcomes
Morbidity and mortality
Quality-of-life
Cost-benefit analysis
Describe step 3 of systematic review
1) Databases: Searches should be explicit in terms of timeframe, language, study type, intervention, condition(s), and characteristics of the population, such as age
2) Searches should be replicated across multiple databases whenever possible:
Medline
PubMed
Embase
Cumulative Index to Nursing and Allied Health Literature (CINAHL)
Scopus
Cochrane Library
Describe steps 4 & 6 of systematic review
1) Title and abstract reviewed by two independent reviewers
2) Any study identified as possibly relevant by at least one of these reviewers then moves on for full text review
3) An alternative approach to this process is to have a single person review all titles and abstracts and have a second reviewer only review those that are recommended for exclusion
-This approach minimizes review burden since the second person only has to review the list of possible excludes and verify coding for why the study was excluded
4) Remove duplicate studies
5) Hand-searching – assessing the journals included in your research for additional studies
-Not required
6) Grey literature – journal articles not published, white papers, or theses
-Not required
Describe steps 5 & 7 of systematic review
1) Assessing risk of bias and rating study quality
Cochrane
Selection bias
Performance bias
Attrition bias
Detection bias
Reporting bias
2) US Preventive Services Task Force
Subjective: consider each article as convincing, adequate, or inadequate; taking into consideration elements of internal and external validity
3) Create PRISMA diagram
4) Quality assessment via National Heart, Lung, and Blood Institute
What are 3 measurements you need to do in steps 5 & 7?
Grading of Recommendations Assessment, Development and Evaluation (GRADE)
A Measurement Tool to Assess Systematic Reviews (AMSTAR)
Risk Of Bias In Systematic Reviews (ROBIS)
Quantitative analysis of heterogeneous studies is called what?
Meta analysis
Describe meta analysis
1) Assign weight to each studies
-Smaller variance and large sample sizes = greater impact on final evaluation
2) Heterogeneity chi-squared
p-value <0.05 generally is interpreted as presence of statistically significant heterogeneity
3) I2 statistic (%)
Higher percentages are associated with greater heterogeneity
4) Forrest plots
The less the confidence intervals of individual studies overlap, the higher the heterogeneity
5) If statistically significant heterogeneity is found, then combining the studies into a single analysis may not be valid.
Describe the parts of meta analysis
Forrest plots are commonly used for continuous data
* A chi-squared statistic with a p-value <0.05 generally is interpreted as presence of statistically significant heterogeneity
*Drug A produces a statistically significant greater decrease in symptoms since the confidence intervals do not contain 0.0000
*Converting to dichotomous data
*Odds ratio (OR) or relative risk (RR)
Meta analysis:
1) Define publication bias
2) Describe funnel plots
1) Publication bias - likelihood that studies with positive findings are more likely to be published and these published findings are more likely to be identified and included in a systematic review and meta-analysis
2) Funnel plots
X-axis weighted mean difference
Y-axis standard error of the weighted mean difference
If all studies are within the pseudo-95% confidence interval pyramid, then there is a very small chance that publication bias is occurring
Egger and Begg test with p-values are options too
