lecture 8

Question 1

meta-analysis

Answer 1

type of systematic review in which data from studies that meet strict inclusion criteria is combined

2 steps

Question 2

2 steps of meta-analysis

Answer 2

extracting the size of the Rx effect from each study

pooling the data to obtain an average single estimate based on all studies - this increase sample size

Question 3

critical apprasial question

Answer 3

are the result valids
what ar the results
will the result hep me in caring for aptient

Question 4

are the results valid

Answer 4

was it clear which trials were to be reivewed
were the most relevant studies reviewed
was the quality of the reviewed studies taken into account

Question 5

was it clear which trials were to be reviewed

Answer 5

clear search strat
clearly defined scope
clear critria for inclusion and exclusion

Question 6

criteria for inclusion and exclusion

Answer 6

type of study
type of particpant
type of intervention
outcome measures

Question 7

were most relevant studies reviewed

Answer 7

conclude there is less evidence than there is
systematically different conclusion to those included in review
publication bias
language bias
grey lit

Question 8

why assess study quality

Answer 8

allow to determine biases

biased primary more likely to provide misleading info

poor quality trials are building blocks of review may follow high quality methods but quality of evidence is poor

Question 9

when to assess study quality

Answer 9

plan what criteria you will use

formal assessment of quality after trials

carried out by 2 independent reviewers

Question 10

reporting trial quality

Answer 10

some form of quality assessment widely recommended

allocaton of concealment
blinding particpant
better to describe than score
quality can be used to influence final meta-analysis results

Question 11

what are the results

Answer 11

continuous- size effect- difference in mean
precision- 95 percent ci or mean difference

dichotomous outcomes- size of efect- absolute risk reduction

precision- 95 percent confidence interval ARR
number needed to treat

Question 12

clinical/ statistical significance

Answer 12

Forest plot is used
does confidence interval cross 0 is p value greater than 0.05

Question 13

averaging studies

Answer 13

simple average gives each study equal weight

this seems intuitvely wrong