7- reviews of evidence

Question 1

define systematic review

Answer 1

an overview of primary studies that used explicit and reproducible method

Question 2

how to make sure the research question is clearly defined

Answer 2

• eligibility criteria
  – types of study
  – types of participants
  – types of interventions
  – types of outcome measures

Question 3

how to find studies to systematic review

Answer 3

• medline/cochrane
• all keywords
• diff languages
• grey literature

Question 4

define meta analysis

Answer 4

a quantitative synthesis of the results of two or more primary studies that addressed the same hypothesis in the same way

Question 5

purpose of meta analysis

Answer 5

• synthesis of lots of study results
• collate study results
• reduce problems due to variations in sampling

Question 6

how to calculate pooled estimate odds ratio in meta analysis

Answer 6

• OR and 95% CI are calculated for all studies
• combined to givee pooled estimated odds ratio
• studies weighted according to their size and uncerainty of odds ratio

Question 7

what is a, b, c, d

what does size of square correlate to

what does width of diamond mean

Answer 7

a- individual odds ratio

b- individual 95% CI

c- pooled estimate

d- dotted = pooled 95% CI, not dotted = null hypothesis

size of square- weighting of study

width of diamond- pooled 95% CI

Question 8

problems with meta analysis

Answer 8

• heterogenity between studies
• variable quality of studies
• publication bias in study selection

Question 9

what is heterogenity

Answer 9

observed effects in studies are more different than we would expect by chance

Question 10

what causes heterogenity

Answer 10

methodological heterogenity- differences in methods

clinical heterogenity- differences in patients, interventions, outcomes

Question 11

how does forest plot show heterogenity

Answer 11

the individual 95% CI (solid horizontal lines) overlap

Question 12

what are the two approaches to calculating the pooled odds ratio considering study heterogenity

Answer 12

fixed effects model- studies are investigating the same true effect size, ie. assumes they are homogenous and any differences are due to random error

random effects model- incorperates heterogenity, assuming studies are investigating similar but not identicle true effect size. any differences due to heterogenity and random error

Question 13

compare forest plot of random effect vs fixed effects

Answer 13

• widening of CI
• more equal square sizes and therefore weighting

(on random effects model)

Question 14

what helps explsin the heterogenity

Answer 14

sub group analysis

Question 15

what can cause variable study quality

Answer 15

poor study design

poor design protocol

poor protocol implementation

bias prone studies eg. case control and non randomised

Question 16

what approaches can be taken to varible study quality

Answer 16

• define basic quality standard and only include studies that meet this
• score each study for its quality and incorporate this score into its weighting

Question 17

how does publication bias occur

Answer 17

studies with statistically significant results more likely to be published tham those with non statisfically significant results

Question 18

how to identify publication bias

Answer 18

• funnel plot- a measure of study size against odds ratio. if no bias then it will be balanced. if there are few small studies with unfavourable results then bias is unlikely
• statical test

Question 19

funnel plot no bias vs evidence of bias

Answer 19

-