- Shuffle
Toggle OnToggle Off
- Alphabetize
Toggle OnToggle Off
- Front First
Toggle OnToggle Off
- Both Sides
Toggle OnToggle Off
Front
How to study your flashcards.
Right/Left arrow keys: Navigate between flashcards.right arrow keyleft arrow key
Up/Down arrow keys: Flip the card between the front and back.down keyup key
H key: Show hint (3rd side).h key
PLAY BUTTON
PLAY BUTTON
81 Cards in this Set
- Front
- Back
|
total # cases/total population at risk
|
prevalence
|
|
new cases/total population at risk
|
incidence
|
|
what must you subtract from total population at risk when calculating incidence
|
those currently or have previously had the disease
|
|
what will increase prevalence
|
chronic disease
|
|
when do prevalence and incidence equal each other
|
acute diseases
|
|
when is odds ratio used
|
in case-controlled studies - approximates relative risk if prevalence of disease is not too high
|
|
calculate relative risk using table
|
(a/a+b)/(c/c+d)
|
|
calculate odds ratio using table
|
(a x d)/(b x c)
*cross multiply in the table |
|
how much of the positive results can be attributed to that risk factor vs. the general population
|
attributable risk
|
|
calculate attributable risk using table
|
(a/a+b) - (c/c+d)
|
|
calculates the percentage of the disease that is due to exposure to the risk
|
(risk total - risk in unexposed) / risk total
(Pe)(RR-1)/[1+(Pe)(RR-1)] |
|
selection bias
|
nonrandom assignment into groups
|
|
sampling bias
|
the study sample population does not represent accurately the entire population
|
|
lead time bias
|
early diagnosis falsely appear to prolong survival
|
|
late look bias
|
people who have already died of a disease are not included in the sample and the cases studied are not as severe
|
|
the reduction/increase in risk associated with treatment as compared to a placebo
|
absolute risk reduction/increase
|
|
number needed to treat
|
1/ARR
|
|
number needed to harm
|
1/attributable risk
|
|
calculate sensitivity using table
|
a/a+c
|
|
calculate specificity using table
|
d/b+d
|
|
LR+
|
sensitivity/1-specificity
|
|
LR-
|
1-sensitivity/specificity
|
|
this test has few false negatives
|
highly sensitive test
|
|
A negative with this test effectively rules out the diagnosis of disease
|
highly sensitive test
*SnNout |
|
this test has few false positives
|
highly specific test
|
|
A positive with this test effectively rules out the diagnosis of disease
|
highly specific test
*SpPin |
|
If the prevalence of a disease in a population is low, what does this do to the PPV
|
makes it low no matter how high the specificity and sensitivity of the test is
|
|
test with a PPV close to 1
TP/TP+FP |
high specificity
*very few FP |
|
test with NPV close to 1
TN/TN+FN |
high sensitivity
*very few FN |
|
standard error of the mean
|
standard deviation/square root (number of items in sample)
|
|
positive skewed distribution
|
tail on right
mean > median > mode |
|
negative skewed distribution
|
tail of left
mean < median < mode |
|
no difference between hypothesized mean and the population mean
|
null hypothesis
|
|
p value < 0.05
|
indicates that there is less than a 5% change that the difference was attributable to random chance
|
|
compares means between two groups
|
t-test
|
|
compares means between 3+ groups
|
ANOVA
|
|
compares 2+ percentages/proportions
|
Chi-square
|
|
compares 2 continuous variables
|
linear correlation
|
|
doing a study and saying there is a difference when one really doesn't exist, happened by chance
|
type 1 error
"saw" difference that didn't exist |
|
Doing a study and saying that there is no difference when there really was
|
type 2 error
didn't "see" difference that does exist |
|
False negative error
|
type 2 error
|
|
study that looks at the disease state of the patient at a given time
|
cross-sectional
|
|
study that compares a group with a given exposure or risk factor to a group without to see what will happen
|
prospective
|
|
study that compares a group of people with disease to a group without to look for prior exposure or risk factors
|
case-control
|
|
the ability to detect a difference between groups if it is truly
there |
power
|
|
calculate power
|
1 - B
|
|
False positive error
|
type 1 error
|
|
what happens to the confidence interval when the sample size is larger
|
width of confidence interval decreases
|
|
reliable test
|
reproducible in that it gives similar results on repeat measurements
|
|
patient has positive test and wants to know probability he is actually positive
|
PPV = TP/TP+FP
|
|
patient sample is selected and separated into 2 groups, exposed vs. unexposed, and look at over period of time to see outcomes
|
prospective cohort
|
|
effect of main exposure is modified by the presence of another variable, not a bias
|
effect modification - natural phenomenon
|
|
calculate attributable risk percent or ARP% using RR
|
RR-1/RR
|
|
absolute risk increase/reduction
|
treatment event rate - control event rate
|
|
what is associated with a confidence interval that never crosses 1.0
|
p value < 0.05
|
|
relative risk
|
event rate in exposed/event rate in unexposed
|
|
how can there be increased prevalence without increase in incidence
|
increased duration of disease - improved quality of care
|
|
what does increased prevalence do to PPV
|
increases PPV
*this of diseases of foreign countries compared to USA |
|
ARR
|
CER - TER
*control event rate - treatment event rate |
|
ARI
|
TER - CER
*treatment event rate - control event rate |
|
probability of being free from disease when the test is negative
|
NPV
|
|
relative risk reduction
|
(control AR - treatment AR)/control AR
|
|
used to compare proportions in two groups such as 2x2 table
|
chi-test
|
|
used to compare exposure of people with disease to exposure of people without disease
|
case-control study
|
|
main measurement used in case-control study
|
exposure odds ratio
|
|
confidence interval calculation: 95%, 99%, 99.7%
|
Mean +/- Z(SE)
Z = 2, 2.5, 3 |
|
tendency of study population to affect an outcome due to knowledge of being studied
|
Hawthorne effect
|
|
selection bias by selecting hospitalized patients as control
|
Berkson's bias
|
|
effect that a researcher's beliefs in the efficacy of treatment that can potentially affect the outcome
|
pygmalion effect
|
|
loss to follow-up can result in what
|
selection bias
|
|
differentiate which tests are better at picking up people without disease vs. picking up people with disease
|
pick up disease - highly sensitive (few false negatives)
pick up non-disease - highly specific (few false positives) |
|
if a test is not accurate, it is what
|
biased
|
|
a reproducible/reliable test is defined as what
|
precise
|
|
differentiate types of variables
|
nominal - no measurement
ordinal - ranking scale of degree of severity dichotomous - normal/abnormal continuous - measured values on a scale (BP, weight, height) |
|
The closer the value is to 1, the stronger the association is between the two variables
|
linear correlation
|
|
When computing a confidence interval, when do you use t and when do you use z
|
use z when standard error is known
use t when standard error is estimated |
|
What is the effect of sample size on the width of a confidence interval
|
the larger the sample size the smaller the width of the confidence interval
|
|
Classifies the disease and exposure status of each individual in a group at a given time
|
cross-sectional study
|
|
results in this study measure prevalence, not incidence
|
cross-sectional study
|
|
This type of study describes the natural history of the disease process. Can calculate incidence to exposure of risk factor
|
prospective study
|
|
Which test is most appropriately used to evaluate the relationship between body weight and systolic blood
pressure |
linear correlation
|
