B3.064 Prework 1 Systematic Review & Meta-Analyses

Question 1

why are individual studies at risk of inaccurately estimating the exposure-outcome relationship

Answer 1

bias
sampling variability
may not generalize across heterogenous populations and settings

Question 2

hierarchy of scientific evidence

Answer 2

1. meta analyses
2. systematic review
3. RCT
4. cohort
5. case-control
6. cross sectional/ecologic
7. case series/report

Question 3

systematic review

Answer 3

literature reviews focused on synthesizing all high-quality research evidence relevant to a research question

Question 4

meta-analyses

Answer 4

statistical synthesis of results from two or more primary studied that addressed the same hypothesis in the same way

Question 5

what do meta-analyses seek to accomplish?

Answer 5

1. systematically examine the strengths and weaknesses of the accumulated evidence
2. explore heterogeneity between studied
3. identify potential biases
4. provide overall estimate of the benefit or harm of treatment

Question 6

what is PICO

Answer 6

patient
intervention
comparison
outcomes
a model for searching literature

Question 7

algorithm for determining the current state of evidence (5 steps)

Answer 7

1. define a question
2. search the literature
3. assess the studies
4. combine the results
5. put the findings in context

Question 8

what do forest plots display

Answer 8

direction and size of effect

Question 9

how are forest plots assembled

Answer 9

estimated treatment effect of each study is represented by the box – note the direction and size of effect
confidence interval, provides a measure of the precision of the estimate

Question 10

size of each data marker on a forest plot

Answer 10

weighting factor assigned to the study

Question 11

combined result on forest plot

Answer 11

long diamond

width = 95% CI

Question 12

sources of heterogeneity

Answer 12

study population differences
differences in interventions
differences in outcomes
differences in methodology

Question 13

I2 statistic

Answer 13

represents the proportion of variation in treatment effects across studies that is due to heterogeneity rather than the usual sampling variability

Question 14

levels of heterogeneity

Answer 14

0-40% might not be important
30-60% may represent moderate heterogeneity
50-90% may represent substantial heterogeneity
75-100% considerable heterogeneity

Question 15

funnel plots

Answer 15

sampling variability decreases as the sample size increases

small studies scatter widely at the bottom with the spread narrowing among larger studies

Question 16

what should a funnel plot look like in the absence of bias

Answer 16

symmetrical inverted funnel

Question 17

how does bias manifest on funnel plots?

Answer 17

smaller studies showing no statistically significant effects remain unpublished

Question 18

search bias

Answer 18

small differences in search strategies can produce large differences in the set of studies found and can impact the overall conclusion

Question 19

selection bias

Answer 19

small differences in inclusion/exclusion criteria can produce large differences in the set of studies included in the analysis
can impact overall conclusion