Intervention Flashcards
Intervention
the course of tx prescribed by the practitioner that ha the goal of improving the pts tatus with regards to the target disorder, illness, and predicament
Critical appraisal-3 questions
are results of the individual study valid, important, and applicable to our patient
Study valid?
randomization
follow up
analyzed in assigned group
blinding
equal tx groups
groups similar at start of trial
Results important?
significance?
magnitude?
precision of treatment effect?
traditional results- establish alpha level, administer treatments, compare means, calculate p value, determine statistical significance
if p > alpha -“not statistically significant”. Chances are what we are measuring came from the same targe population and infers the two txs have a similar effect.
if p < alpha - “statistically significant”. Chances are what we are measuring came from different target populations and infers that the two txs have a diff effect.
Statistically significant
not a measure of magnitude
does not address good vs harm
statistically significant finding not likely to have occurred if null hypothesis is true. Does not necessarily mean that the results are important, and is not an indication that the difference is large
Results important?
for continuous measures- effect size (ES)
for categorical measures- magnitude of tx effect (NNT and NNH) and precision of tx effect (conf interval around NNT and NNH)
control event rate (CER)- count # of events in control group
experimental event rate (EER)- count # of events in experimental group
absolute event rte reduction (ARR= EER-CER)
NNT= 1/ARR
Effect size (ES)
specific type of signal to noise ratio
numerator- diff in means
denominator- common SD
ES= (M1 - M2)/ Pooled SD
NNT
the # of patients that need to be treated in order for one to benefit
ARR= EER- CER
ARR= .816- .581
NNT= 1/ ARR= 1/.235
NNT= 4.25
NNH
large #s are good because they mean that adverse events are rare.
Small value for NNH are bad, because they mean adverse events are common
Forest plots
graphical representation of SR results
indicates treatment effect size
if meta-analysis (cumulative effect size)
need to identify- outcome, comparison, point estimates, confidence intervals
diamond shows everything together
