Stats Flashcards
t test and ANOVA
quantitative (ANOVA is means)
chi square (2 groups of qualitative)
categorical
kaplan meyer
small groups; estimates survival
relative risk
incdience of exposed/incidence in unexposed
power of test
probably of making correct consluion
1-probabily of TYPE II ERROR
0.8+ is adequate
power = 0.9
90% chance a positive value detects real difference
10% chance a negative value does not detect difference when there actually is one
accuracy equation
true positive + true negative / (ALL)e
deepdnency on prevalence
predictive value depends on prevalence
sens/spec does NOT depend on prevalence
NSQIP
outcome data for surgical QUALITY in US
JCAHO
side marking, timeout etc.
REAIM
reach
effectiveness
adoption
implementation
mainteannce
for getting evidence based practice out there
QI types
lean process and six sigman p
lean process
continue to reevaluate and get rid of stuff that doesnt work; LEAN METHODOLOGY
outcomes will contineu to improve.
six sigma
define measure analyze imrpove control
eiminate process errors
ERROR WASTE REDUCTION
median in kaplan meier survival curve
50% survival; what is the X value (time)
NNT
1/ARR
what bias in cross sectional study
no causality…
selection bias
survival bias
case control bias
only study on outcome (some causal effect) with
selection bias
recall bias
cohort study bias
some potential to establish causal effects
selection bias
RCT is not good for…
rare events
receiver operating curves
ROC used to eval what happens to sensitivity and specificity when threshold for positive vs negative test is changed
largest area under ROC
optimized/highest possible combination of sensitiity and specificity for a test
