EIDM

Question 1

What are three qs to ask when looking at a study?

Answer 1

1) Does the study answer my practice question?
2) Are the results of the study valid? (minimizes bias, results not due to other factrors, results can be applied to larger population, outcomes analyzed)
3) Can I apply these results to my nursing practice?

Question 2

What are intervention studies?

Answer 2

• address a treatment/therapy or prevention
• RCT is the strongest design
• determine how the interventions affects the outcome
• RCT not always ethical, observational studies can be used

Question 3

What observational studies exist?

Answer 3

• cohort analytic study (one group, before and after)
• case control study

Question 4

How can participants be selected for an intervention study?

Answer 4

• convenience sampling (volunteer); potential bias, making results less valid
• random sampling (ideal)
• better if allocation can be concealed

Question 5

What is allocation sequence?

Answer 5

• random process of allocating participants into experimental/control groups

Question 6

What is allocation concealment?

Answer 6

• researchers not aware of how allocation occurred, unpredictable (remote randomization, opaque envelopes, medications look similar)
• participants blinded to what group they are in
• prevents bias

Question 7

How do we know whether randomization of participants occurred?

Answer 7

• compare experimental and control groups - similar in all known determinants of the outcome
• usually uses p-value to compare the two groups
• assume unknown determinants corrected with randomization

Question 8

Where could potential biases occur without blinding?

Answer 8

• participants if they know what group they are in (Hawthorne effect)
• people delivering the intervention could make intervention look more effective
• people evaluating the outcome

Question 9

What is intention-to-treat analysis?

Answer 9

• outcome analysis should include the participants who dropped out before completion of the study within the group they were originally allocated
• statistical analysis needs to include all participants, even those not completing the study

Question 10

What should we ask about follow-up?

Answer 10

• was it complete and was it long-enough?
• were all participants accounted for?

Question 11

What are the two types of variables?

Answer 11

1) dichotomous (yes/no)
2) continuous

Question 12

What are the types of continuous data?

Answer 12

• ordinal: sets of ordered categories (pain scale)
• interval: ordered set where series of intervals are the same size (temperature)
• ratio: interval with absolute 0 (weight)
• can be reported as dichotomous data but not vice versa

Question 13

How is treatment effect measured for continuous data?

Answer 13

• uses “mean” values to show whether data changed after exposure to intervention
• may include the range of values and/or standard deviation (variable around the mean)
• line of no difference is 0

Question 14

How is treatment effect measured for dichotomous data?

Answer 14

• line of no difference is 1
• outcomes reported as relative risk (RR) or odds ratio (OR)
• RR or OR = 1 means there is no difference between the control and intervention group results
• RR or OR > 1 means more of the outcome in the intervention group (less in the control group)
• RR or OR < 1 there is less of the outcome in the intervention group (more in the control group)

Question 15

What is OR/RR?

Answer 15

The odds or risk that an outcome will occur in the intervention group compared to the control group

Question 16

If the outcome is bad, risk is reported as…

Answer 16

• relative risk (RR) = relative benefit of a treatment
• absolute risk diference/reduction (ARR) = how much of a reduced risk if the result of the treatment

Question 17

If the outcome is good, risk is reported as…

Answer 17

• benefit increase

Question 18

How do we determine treatment effect?

Answer 18

• determine if the outcome is a“good” (desirable) or “bad” outcome
• treatment effect is the outcome difference between the control and treatment group
• more/less of a good outcome?
• more/less of a bad outcome?

Question 19

How do we determine whether results are valid?

Answer 19

• 95% CI does not cross the line of no difference
• p-value <0.05

Question 20

What is the 95% confidence interval?

Answer 20

• 95 times out of a 100 the treatment effect represents an estimate of the true difference that is not due to chance
• estimates the statistical significance of the intervention effect
• wider the CI range, the less precise the estimate of the treatment effect

Question 21

What is number needed to treat (NNT)?

Answer 21

• related to the effect of the intervention
• only used with dichotomous outcomes
• tells us whether the interventions is worth any potential harms or costs
• number of participants that must receive the intervention in order to promote one additional good outcome or prevent one additional bad outcome

Question 22

What is an acronym for the critical appraisal process for RCTs?

Answer 22

Design appropriate
Allocation
Demographic (variables)
Follow up
Intention to treat
Bias, Blinding
Statistical significance

Question 23

What is a systematic review?

Answer 23

• summary of multiple research evidence that answer the same specific and focused question
• quantitative studies
• attempt to overcome bias via rigorous searching, appraisal, data synthesis
• 2 or more people select research evidence to include based on inclusion criteria

Question 24

What is a meta-analysis?

Answer 24

A systematic review which includes a quantitative (statistical) combination of the results of similar studies

Question 25

What is a meta-synthesis?

Answer 25

A systematic review of primary qualitative studies on a particular topic

Question 26

How is a literature search conducted for a SR?

Answer 26

- studies that used most appropriate design for research question - uses exhaustive strategies: search terms, several databases, reference list of relative studies, hand-searching, consulting experts, multiple languages, inclusion dates

Question 27

How is the validity of the included studies assessed?

Answer 27

- poorer quality of studies tend to overestimate the effectiveness of the intervention - uses a specific quality check list (rating tool) - identify any risks for potential bias - 2 or more “raters” used - disagreement resolution method

Question 28

How is heterogeneity between studies tested?

Answer 28

- statistical test - compares any differences between the different study results - are the results of the studies similar? • the p-value is ≥ 0.05, any difference across the studies included in the review was due to chance alone - means they are homogenous - I2 test: 0% indicating the least amount of variability across the studies

Question 29

How can the SR be completed for heterogenous studies?

Answer 29

- if study results are heterogeneous – the results can be combined statistically by using fixed effects model or random effects model - can do a meta-analysis - or, can synthesize data narratively

Question 30

How is treatment effect shown in an SR?

Answer 30

- forrest plot or blob-o-gram - each study has a box and a line through it - size of box = sample size - line through the box = 95% CI - diamond = overall treatment effect, edges of the diamond = CI - statistically significant = does not cross the line of no difference - dichotomous outcome = 1 - continuous outcome = 0