Ethics/Statistics Flashcards
can you perform practice procedures on a newly deceased patient
Yes, but you must gain permission from patient before dying or a patient’s family member
what is information bias
bias introduced by imprecise measurement methods of exposure or outcome
difference between two sample t-test and two sample z-test
t-test compares means between two samples
z-test compares means between two populations (not frequently used)
what is ANOVA
Analysis of Variance: used to compare three or more means
what is the chi square test used for
to compare categorical data or proportions
what are the equations for relative rate and odds ratio
RR= [(a/a+b)/(c/c+d)]
OR=ad/bc
how do you calculate attributable risk percentage
ARP= (relative risk - 1) / (relative risk)
what is selective survival bias
seen in case control studies; selecting patients from the entire population as opposed newly diagnosed patients will give you a higher proportion of patients with less severe disease since most of the severe cases would have died already
what happens to the PPV and NPV of a test if a population has a high prevalence of the tested disease?
what about if a population has low prevalence of the tested disease?
high prevalence population: PPV increases, NPV decreasese (since with a neg test you would still suspect diseas)
low prevalence population: PPV decreases, NPV increases
what’s a quick and easy way to distinguish effect modification from confounding?
stratify by the proposed confounder; if there is no association in either of the stratified groups then the variable you stratified by was a confounder
if you stratify by the variable and there is a strong association found btwn exposure and outcome in one group but not the other than there was likely an effect modification
explain a factorial design
2 or more interventions are studied, with 2 or more independent variables studied within each intervention
explain cluster analysis
involves grouping different data into similar categories or randomization at the level of groups rather than individuals
what is Berkson bias
a form of selection bias in which studies conducted using hospital patients as sample are unrepresentative of entire population
what is attrition bias
a form of selection bias in which a significant amount of subjects are lost to follow up and may differ from those who remain
what kind of bias should you think of when you see a new screening test for a poor-prognosis disease (e.g. cancer)
lead-time bias
if a minor requires urgent or emergent care and the parents have joint custody, but one refuses the care what should you do
proceed with care; for joint custody cases only one parent’s consent is needed
what is the most important tool for decreasing medical errors due to failures of communication between physicians in sign-out?
implementing a sign-out checklist
a child needs chemo for ALL, but the parents refuse the treatment, what do you do?
obtain a court order; in the U.S. parents are not allowed to refuse life-saving treatment for their children (including for religious reasons)
a patient with a highly contagious disease does not want to stay in the hospital to be treated; what do you do
hospitalize against his/her will (court order not needed as this is a public health threat)
what is prevalence odds ratio and in what kind of study is it used?
used in cross-sectional studies to compare prevalences between two populations
how do you determine the median of a number set with an odd number of items?
even number?
odd: the median is the number that splits the number set exactly in half (located exactly in the middle)
even: the median is the average of the two middle numbers
what it type II error?
type II error=beta; think of the mnemonic “blocking”; type II error occurs when the study misses a true effect and incorrectly accepts the null hypothesis
what is type I error?
type I error=alpha; think of mnemonic “assumes”; type I error occurs when the study picks up an effect/ association that isn’t really there in actuality
the null hypothesis is incorrectly rejected
what is power in terms of either alpha or beta?
power= 1 - beta
what does the confidence interval tell you about statistical power
the narrower the confidence interval, the greater the statistical power
define statistical power
power is the probability of finding a difference that truly does exist
define primary, secondary and tertiary prevention
primary: preventing a disease (decreasing incidence)
secondary: early detection and slowing progression
tertiary: mitigating morbidity and mortality/ reducing complications