Ratio measures assess strength of effect - how effective is thetreatment?
Difference measures take into account frequency of the outcome –can assess whether it is worthwhile (allocation of time and $$)
Both ratio and difference measures are needed
All these measures are estimates and are subject to sampling error– need confidence intervals to determine their precision
All the measures are limited by the study(ies) that generated them– they may vary by patient characteristics, adherence totreatment, duration of follow-up, etc)
Measures consider only beneficial and not adverse effects oftreatment.
Aspirin in prevention of MI among malesmokers(data from Physicians’ Health Study)
5-year incidence of MI:
aspirin group = 1.2%
placebo group = 2.2%
Risk ratio = 1.8
Relative risk reduction = 45%
Absolute risk reduction = 1.0% in 5 years
NNT = 100 for 5 years (to prevent 1 MI)
Antihypertensive treatment in 75-year old women with BP of 170/80(data from SHEP study)
•5-year incidence of stroke:
treatment group = 5.2%
placebo group = 8.2%
–Risk ratio = 1.6
–Relative risk reduction = 37%
–Absolute risk reduction = 3.0% in 5 years
–NNT = 33 / 5 years (to prevent 1 stroke)
Measures of effect in RCTs:continuous outcomes
•Example: RCT of antidepressant vsplacebo:
•Measures on depression scale at baselineand at follow-up
•Possible measures:
–Difference in mean scores at follow-up
–Difference in change scores from baseline tofollow-up
Measures of effect in RCT:adjustment for covariates
•Is it necessary?
•Compare characteristics of study groups at baseline(statistical testing not appropriate but may be requested!)
•Regression models:
–time to event: Cox proportional hazards
–categorical outcome at point in time: multiple logisticregression
–continuous outcome (at point in time or change score):multiple linear regression
Measures of effect inobservational studies
•Cohort studies:
–can use same measures as in RCTs but control ofconfounding is essential
•Case-control studies:
–odds ratio may be used to estimate relative risk undercertain assumptions
–relative risk reduction can be computed as:
1 - 1/OR
–risk difference and NNT cannot normally be computedfrom case-control studies
Example: a quasi-randomized trial of a 2-stage ED intervention for seniors
Every 2-3 days during 1st week,then weekly until discharge
If discharged before 8 weeks:8-week post discharge homeassessment
PRIMARY OUTCOMEPRIMARY OUTCOME
MEASURE (continued)MEASURE (continued)
Time to improvement in hospital
Improvement = MMSE scorepersistently at least 2 points higherthan initial score
Kaplan-Meier survival curves ofpercent with improved MMSE scoreKaplan-Meier survival curves ofpercent with improved MMSE score
Kaplan-Meier survival curves of percentwith improved MMSE scorestratified by dementiaKaplan-Meier survival curves of percentwith improved MMSE scorestratified by dementia
Measure of effect
•Hazard ratio (HR) for shorter time toimprovement = 1.10 (95% CI: 0.74, 1.63)