Card Set Information
Epidemiology and Biostatistics
What is prevelance and how is it calculated?
P = (number of exisiting cases of disease at a given point in time)/(total population)
P = incidence x duration
What is incidence?
Number of new cases that occur during a period of time.
What is cumulative incidence?
Proportion of people whop became diseased during a perior of time (probability/risk of disease during the period of time).
What is sensitivity and how do you calculate it?
Proportion of those with disease who tested positive
Sensitivity = TP/(TP + FN)
What is specificity and how do you calculate it?
Proption of people without disease who test negative
Specificity = TN/ (TN + FP)
What is the PPV and how do you calculate it?
Proportion of people who test positive that are really diseased
PPV = TP / (FP / TP)
What is the NPV and how do you calculate it?
Proportion who test negative and really aren't diseased
NPV = TN / (TN + FN)
What factors change the PPV & NPV?
changing disease incidence /prevalence
What's a type 1 (alpha) error? What is it's measurement?
Stating that there is an effect or difference when none exists;
is probablility of making type 1 error
What's a type 2 (beta) error?
Stating that there is not an effect or difference when one exists;
What is the power of a study?
Probablilty of getting the right answer to the study.
What factors change the power of a study?
Number of points in the population
Differnece in complaince between groups
Size of expected effect
What's a confidence interval?
Range of values in which a specified probability of the means are expected to fall.
When is a confidence interval useless?
Interval includes 0 or 1 = H
Intervals overlap = not significantly different
What is an experimental study with random allocation?
Randomised controlled trial
What's an experimental study without random allocation?
Non-randomized controlled trial
what's an ovservational study with no comparison group?
What's an ovservational study with a comparison group?
Analytical Study (cohort, case control, cross sectional)
What's a cohort study?
Taking an exposure group and looking forward for an outcome
What's a Case-control study?
Analytica study looking from an outcome back to evaluate for exposure
what's a cross-sectional study?
Looking at the exposure and outcome status at the same moment in time
What's the correlation coeficient and what does it mean?
(r) bewtween -1 and +1
The closer to +1 the stronger the correlation between the two variables
What is the coefficent of determination?
this is the value that's reported
What is a case report and what are it's purposes?
Describe experience of individula/population
First step in identification of new things
1. describe and evaluate trends in outcomes
2. provide data for panning
3. suggest new research
What's a correlation/ecological study and what are strengths/limitations?
Study of group characteristics; measure of association; non-comparative to another group
: quick, cheap, easy
: inability to link exposure and disease; lack of control over confounding factors
What is ecologic fallacy?
Inapporpriate inference from ecological data (relationship of groups doesn't mean relationship of individuals)
What are the pro/cons of cross sectional study?
: quick, cheap, establish "burden of diease"
: confuses causation and correlation
What are the pro/con of case control studies?
: evaluation of rare diseases with long latency, quick and cheap, multiple factors for a single disease
: limited for rare exposures, no incidence rates, difficult to establish temporal relationship, prone to bias
What are the strengths of a cohort study?
Rare exposures, good information collected
Least bias, clear temporal relationship
direct measurement of incidnce in exposed vs non-exposed
What are limitations of cohort studies?
Ineficient for rare diseases
Expensive and time consuming
Retrospectives are more vulnerable to bias
What is an experimental study? What are key components of the study?
Randomization, blinidng, placebo controlled
Cross over - each subject serves as their own control
Factorial design - allows for evaluation of multiple hypothesises
What's primary disease prevention?
Remove the risk (ie vaccination)
What's secondary disease prevention?
Early detection and treatment (ie, pap smear)
What's tertiary disease prevention?
reduce complications of the disease (ie; insulin for diabetes)
How do you fill in the box for all the calculations?
What's an odds ratio and how is it calculated?
odds of being diseased if exposed
: odds of being diseased if unexposed
(A/B) / (C/D) = OR
Whats an absolute (attributable) risk and how's it calculated?
% increase of disease attrributabe to exposure
What's an absolut (attribuateble) risk reduction and how's it calculated?
% differnce in diseased between normal and treated groups
What's the relative risk and how's it calculated?
The probability of getting diseased in exposeure group when compared to normal
What's the Relative risk reduction and how's it calculated?
% attributable of diseased to a factor (ie
: 1/3 of pneumonia is caused by smoking)
How do you calculate Confidence Interval
Confidence Interval = (mean ±Z( σ /√N))
Z = 2; σ = SD