Biostats/Other Flashcards
When to use a chi-square test for measuring statistical associations? What about z-tests, t-tests and ANOVA?
Categorical data = chi-square
2 means = t and z tests
3+ means = ANOVA
What does a hazard ratio represent?
Event rate in treatment group compared to comparison group
Cluster analysis
Study randomization occurs at the group level, not the individual level
False positive ratio
1 - specificity
False negative ratio
1 - sensitivity
Study population alters behavior because they know they are being observed
Hawthorne effect
What does power represent?
The ability of a study to identify a statistically significant difference between 2 groups
Number needed to treat
Absolute risk reduction = incidence in comparison group - incidence in treatment group.
NNT = 1/ARR
Susceptibility bias
A form of selection bias where a treatment regimen depends on the severity of their disease and is not randomized
Method of characterizing exposure-disease association in case-control trials
Odds ratio = ad/bc
Prevalence
Incidence x time
Attributable risk percent
(Risk in exposed - risk in unexposed)/risk in exposed
Also = (RR-1)/RR
How can you control for confounding?
Matching, restriction and randomization during the study design stage.
Stratification and statistical modeling during the analysis stage.
Requirement for a patient to enter hospice care
Prognosis ≤ 6 months
When can you accept nonmonetary gifts from interested third parties?
Direct benefit to patient care and of small value