systematic reviews and meta analysis course guide

Question 1

what is a systematic review

Answer 1

aims to answer defined question by collecting and summarising empirical evidence that fits eleigibility criteria

Question 2

what is a meta analysis

Answer 2

statsistical techniques used in a systematic review to integrate the results of the studies

Question 3

why do a systematic review

Answer 3

high volume of data
impossible for clinicians to analyse it all
individual studies - not enough power to make decisions
multiple studies = inconsistent results
review gives genrealisable conclusions

Question 4

advantages to a systematic review

Answer 4

transparent - explicit methods
meta analysis increases power and enhance precision accounting for sample size and uncertainties
demonstrate lack of adequate evidence

Question 5

stage 1 of the systematic review

Answer 5

plan
define the qn
frame qn around the participants, exposure, outcomes and study designs of interest

Question 6

stage 2 of the systematic review

Answer 6

identify research - clearly defined search criteria and a thorough search of literature
selection of studies - inclusion and exclusion criteria a priori
study quality assessment - assessed against recognised/user defined criteria eg identify biases

Question 7

stage 3

Answer 7

reporting and dissemination - details extracted and effect estimate
details tabulated
estimate an overall effect by combining the data

Question 8

strength of meta analysis

Answer 8

more subjects - more reliable and precise estimate

heterogeneity can be identified and explored

Question 9

weakness of meta analysis

Answer 9

if too heterogenous inappropriate to pool

Question 10

stats and visualisation of meta analysis

Answer 10

forest plot

graphical representation of each result and combined meta analysis result

Question 11

escribe publication bias

Answer 11

grewater likelihood of papers with stat significance to be published
failure to include all data (including null results) may mean effect is over/under represented

Question 12

how is publication bias explored

Answer 12

funnel plots

see if link between study size and effect of estimate

Question 13

heterogeneity exists because of

Answer 13

clinical, methodological differences and unknown study characteristics

Question 14

how can studies differ

Answer 14

population, exposure, outcomes and designs used

Question 15

how can you explore heterogeneity

Answer 15

Galbraith (radial) plots

Question 16

limitations in systematic reviews

Answer 16

if too few studies matching criteria the review might not add much
if methods are inadequate - findings may be compromised
publication bias

Question 17

guidelines and protocols for systematic review

Answer 17

Cochrane Database of Systematic reviews set standards. started to fascilitate the decisions people had to make for health interventions according to EBM
Preferred Reporting Items for Systematic Reviews and Meta Analyses (PRIMSA) evidence based minimum set of items for reporting, focus on evaluating randomised controlled trials - bases for reporting systematic reviews - can be used for other things eg evaluation of interventions

Question 18

prevalence

Answer 18

Number of cases of a disease within a defined population measured at a specific point in time. Prevalent cases include both new (incident) and existing cases.

Question 19

how can HIV be tested for

Answer 19

self report

saliva and blood testing - ELISA

Question 20

problem with measuring point prevalence

Answer 20

consider data available on census day - participation bias

Question 21

problem with HIV tests

Answer 21

look at sensitivity and specificity

Question 22

problem with self reporting

Answer 22

people might not know

or are ashamed

Question 23

overall incidence =

Answer 23

new cases in cohort/total number in cohort

Question 24

relative risk

Answer 24

incidence in exposed group/incidence in non-exposed group

Question 25

what does relative risk of 8.14 mean

Answer 25

risk in exposed group >8x higher than non-exposed group

Question 26

explanation for findings

Answer 26

chance
bias 
representative/generalisable?
confounding 
cause

Question 27

should we circumcise men for HIV protection

Answer 27

yes - reduce rates and prevent spreading until vaccine is produced
no - not invulnerable, should only be used as part of wider strategy

Question 28

summarise male circumcision for HIV prevention

Answer 28

reduce risk by 60%
3 randomised controlled trial shown circumcision with proper facilities is safe
only where there are heterosexual epidemics, high HIV and low male circumcision
should be combined with treatment for sexually transmitted infections; the promotion of safer sex practices; the provision of male and female condoms and promotion of their correct and consistent use.