44 Statistics and Patient Safety Flashcards
type 1 error
rejects the null hypothesis incorrently, falsely assumed there was a difference when no difference actually exists
type 2 error
accets null hypothesis incorrectly, 2/2 small sample size, tx are interpreted as equal when there is actually a difference
null hypothesis
hypothesis that no difference exists between groups
p value
convention is <0.05 rejects the null hypothesis … mean 95% likelihood that the difference between the populations is true … <5% likelihood that the difference is not true and occurred by chance alone
variance
spread of data around a mean
parameter
population
mode
most common value
mean
average
meadian
middle value, 50th percentile
trials and studies: list types
RCT, double-blind controlled trial, cohort study, case-control study (retro), meta-analysis
randomized controlled trial
prospective study with random assignment to treatment and non-tx groups, avoid treatment biases
double-blind controlled trial
prospective study in which patient and doctor are blind to treatment, avoids observational biases
cohort study
prospective study - compares disease rate between exposued and unexposed groups (random assignment)
case-control study
retrospective study in which those who have the disease are compared with a similar population who do not have the disease, the frequency of the suspected risk factor is then compared between the 2 groups
meta-analysis
combines data from different studies
list quantitative vs qualitative variables
quant = student’s t test, paired t test, ANOVA … qual = nonparametric statistics, chi-squared, kaplan-meyer
student’s t test
2 independent groups, variable is quantitative - compares means
paired t test
variable is quantitative, beore and after studies (i.e. weight before and after drug vs placebo)
ANOVA
compares quant variables (means) for more than 2 groups
non-parametric stats
compare categorical (qualitative) variable (i.e. race, sex, medical problems and diseases, meds)
chi-squared test
compares 2 groups with categorical (qualitative) variables - i.e. number of obese patients with and without DM versus nonobese pts with and without DM
Kaplan-Meyer test
small groups, estimates survival
relative risk
incidence in exposed / incidence in unexposed
power of test
probability of making the correct conclusion = 1 - prob of type 2 error (accepts null incorrectly) … likelihood that the conclusion of the test is true … larger sample size increases power of test
