Biostatistics Flashcards
What is the mean?
Average
What is the Median?
Middle value
What is the Mode?
Most frequent value
What does sensitivity tell you?
People that have the disease w/ pos test
What does specificity tell you?
People that don’t have the disease w/ neg test
What does positive predictive value tell you?
Probability of having a disease w/ a pos test
What does negative predictive value tell you?
Probability of not having a disease w/ neg test
What does incidence tell you?
New cases
(Rate per unit time)
What does prevalence tell you?
Total cases
(At one time)
What does the Odds Ratio tell you?
Diseased are x times more likely to see risk factor
What does a Confidence Interval = 95% tell you?
95% sure it lies within the interval
What does Relative Risk tell you?
Risk of getting disease w/ known exposure
What is NNT?
Number Needed to Treat to change 1 life
What does a p-value <0.5 tell you?
Random chance that you will be wrong 1 time out of 20
What does a Null hypothesis tell you?
Nothing’s happening
What does Power tell you?
Probability of detecting a true intervention
What is a type I error?
p-value error
False negative
“Too optimistic”
What is a type II error?
Power error
False positive
“Too pessimistic”
What is Accuracy?
Validity
“Truth”
What is Precision?
Reliability
“Keep making the same mistake”
What are the phases of clinical trials?
Phase I: Toxicity “hurt pt?”
Phase II: Efficacy “help pt?”
Phase III: Comparison “any better?”
Phase IV: post-marketing surveillance “can they screw it up?”
What is a Cohort study?
Prospective study that provides incidence (new cases)
Uses Relative Risk (RR)
What is a Case Control study?
Retrospective study that provides prevalence (total cases)
Uses Odds Ratio (OR)
What is a cross-sectional study?
Provides prevalence
“Snapshot”
What is a Case Report?
Describes an unusual patient
What is a Case Series report?
Describes several unusual patients
What are Consensus panels?
Panel of experts provides a recommendation
What is Clinical wisdom?
“I think…” paper
What is a Meta-analysis?
Tries to combine data from many trials
