Reading...
Play button
Play button
Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
105 Cards in this Set
- Front
- Back
- 3rd side (hint)
|
The epidemiological triad?
|
host, environment, agent
|
|
|
|
epidemic
|
occurence of an illness in a population that is higher than normal
|
|
|
|
pandemic
|
occurence of an illness worldwide that is higher than normal (pan = everywhere)
|
|
|
|
endemic
|
usual occurence EN the neighborhood
|
|
|
|
epidemiology
|
basic science of public health, the study of disease and health in populations
|
|
|
|
Sensitivity
|
test positive / # with disease
|
Probability of catching the disease (P=P/D), PPD!
|
|
|
Specificity
|
ability of test to ID those without disease: negative test / those without disease
|
|
|
|
100% sensitive test: patient gets negative
|
patient MUST not have the disease (true negative)
|
high SNOUT, SeNsitivity rules OUT
|
|
|
Type of test used for screening diseases with low prevalence?
|
high sensitivity
|
|
|
|
Confirmatory tests after a positive screening test should be?
|
highly sensitivty
|
SPIN, specificity used to rule IN
|
|
|
ELISA
|
highly sensitivy:
high false positive rate. confirme with a specific test |
|
|
|
Western blot
|
highly specific:
high false negative rate |
|
|
|
all positives must be true positives with:
|
Highly sensitive test
|
|
|
|
PPV?
|
number of positive test results that are true positive.
|
TP/P
|
|
|
PPV of a test detecting a low-prevalence disease?
|
it will always be low, regardless of spencificity or sensitivity
|
lots of people getting tested, most don't have the dz, but some will have false positives!
|
|
|
NPV
|
probability that a person is disease free given a negative test result
|
TN/FN+TN
|
|
|
What is statistical significance?
|
when values observed are not due to chance alone.
|
|
|
|
prevalence > incidence
|
chronic disease
|
|
|
|
incidence x disease duration
|
prevalence
|
|
|
|
prevalcne = incidence
|
acute disease
|
|
|
|
Chi squared test df?
|
# of rows - 1
|
|
|
|
normal p-value that is considered statistically significant?
|
0.05
|
|
|
|
What is alpha?
|
this is the false positive error: the probability that you said "there is a difference" when there really isn't
|
|
|
|
what is beta?
|
this is false negative error: the probability that you said "there is no difference" when there really is
|
|
|
|
What is the power?
|
1-beta. TRUE NEG:
the probability of rejecting the null hypothesis correctly. |
|
|
|
What parameters does power depend upon?
|
total number of end points experienced by a population
2) differences in the compliance between two treatment grps (differences in mean values) 3) size of expected effect |
|
|
|
SEM
|
standard error of mean: SD/root(n)
|
|
|
|
1, 2, and 3 STDs?
|
68-95-99.7
|
|
|
|
Chi-squared tests for?
|
differences between percentages/proportions of categorical outcomes.
|
with disease or without disease is an example of categorical outcome
|
|
|
What is a P-value?
|
the probability that the null hypothesis is incorrectly rejected (false positive)
|
|
|
|
Biostatistics
|
the application of stats in order to answer biological questions
|
|
|
|
Descriptive Statistics
|
presenting organized QUANTITATIVE data in order to answer Q's
|
|
|
|
Inferential statistics
|
drawing conclusions about a population based upon observation of a sample
|
|
|
|
Standard deviation?
|
average distance of each obersvation from the mean
|
|
|
|
SE=
|
STD/root(n)
|
|
|
|
Mean mode and median are identical
|
normal distribution
|
|
|
|
mean>median>mode?
|
positive skew MEME MO!
|
|
|
|
mode>median>mean
|
negative skew
|
|
|
|
Normal range of values in normal distributrion?
|
+/- 2 stds to get 95% of values
|
|
|
|
Normal range of values in skewed distribution?
|
2.5th -97.5th percentile
|
|
|
|
definition of standard error
|
variability of the sample means
|
|
|
|
necessary and sufficient:
|
one and only one thing always causes the other.
|
|
|
|
sufficient but not necessary:
|
A always causes B, but C and D can calso cause B
|
|
|
|
necessary but not sufficient:
|
one thing always causes the other but in combo w others
A always leads to B if C and D are present |
|
|
|
Koch's Postulates?
|
organism is always found in the disease,
2) organisms not found with any other disease 3) organism can be cultured and innoculated into another host |
|
|
|
Bradforld hill critera of causation?
|
based upon strength, specificity, consistency, temporal and dose relationships
|
|
|
|
whatdoes nominal data mean?
|
categorical data with no order like male/female
|
|
|
|
ordinal data is?
|
data with order: ie class rank
|
|
|
|
requirement for T test?
|
normal distributions and equal variances
|
|
|
|
T test determines?
|
if the means of two groups are statistically different from one another
|
|
|
|
what does spearman rank order statistic tell you?
|
ordinal data: p
ie birth order and education |
|
|
|
Artificial increase in incidence?
|
new diagnostic procedure
|
|
|
|
Adjusted mortality rate by age?
|
take proportion of those in each age category, multiply by change of mortality.
|
|
|
|
What is PMR?
|
proportion of overall mortality due to a specific cause, not a rate.
|
|
|
|
differences in PMR due to?
|
can be related to morbidity in other diseases
|
|
|
|
PMR is useful for?
|
looking at major causes of death but not risk of dying from a disease
|
|
|
|
Underlying cause of death?
|
trigger that lead to chain of events for death
|
|
|
|
Who fills out underylying cause of death?
|
physician
|
|
|
|
who fills out death certificate?
|
funeral director except for underlying cause of death
|
|
|
|
death certificates coded by
|
underlying cause of death
|
|
|
|
case fatality rate?
|
number died of specific disease/number w disease
|
|
|
|
Kaplan-meier method?
|
tick marks on a time scale that represents when person was censored
|
|
|
|
Log-rank test?
|
comparing observed with expected and seeing if observed is statistically significant (prognostic tool)
|
|
|
|
CoxHazard model?
|
uses regression to determine prognosis
|
|
|
|
Crude mortality rate reported as?
|
deaths per 1,000
|
|
|
|
RR<1
|
exposure is harmful
|
|
|
|
relative risk
|
exposed risk to unexposed risk ratio
|
%45 risk of cancer w smoking / %10 risk without smoking
|
|
|
odds ratio
|
the ratio of the number of % occuring : % not occuring
|
ie. p(home run) = .1
p(no home run) = .9 odds ratio : .1/.9 |
|
|
ratio used for case-control studies?
|
C.C.oDDs
odds ratio |
|
|
|
with a disease with low prevalence, what does odds ratio approximate?
|
relative risk
|
|
|
|
Risk calculated for cohort studies?
|
relative risk
|
|
|
|
What is attributable risk?
|
the difference in risk between exposed and unexposed groups
|
|
|
|
the proportion of disease that are attrubtable to the exposure?
|
attributable risk
|
|
|
|
what is more easy to interpret? RR or OR?
|
RR is more accurate and easy to interpret.
|
|
|
|
1-RR
|
relative risk reduction
|
|
|
|
formula for relative risk reduction?
|
%risk in control - %(risk in exopsed) / (%risk in control)
|
|
|
|
Absolute risk reduction formula
|
% control - % in experiemnetal
|
|
|
|
What is absolute risk reduction?
|
difference in precent reduction between control and experimental risks.
|
|
|
|
number needed to treat?
|
1 / absolute risk reduction
|
|
|
|
number needed to harm?
|
1/ attributable risk
|
Harm Attributable Risk m
HARM |
|
|
Population attributable risk formla?
|
incidence in exposed - incidence in unexposed / incidence in exposed
|
|
|
|
What is number needed to treat?
|
the number of indivuals needed to be trated to prevent one adverse event
|
|
|
|
Screening time biases?
|
1) patient selection
2) lead time 3) legnth time 4) overdiagnosis |
|
|
|
How might patient selection be a form of screening time biase?
|
PEople who want to get tested are likelier to be healthier and adhere to treatment and therefor recover!
|
|
|
|
What is length time bias for screening biases?
|
It is idea that screenable tests naturally have longer preclinical phases and therefore have a better prognosis regardless
|
|
|
|
PPV?
|
probability that a person that tests positive has teh disease
|
|
|
|
NPV?
|
prob that a negative test really is negative
|
|
|
|
Ecological studies
|
unit of analysis is a group, not individual (ie incidence by country or ethnicity)
|
|
|
|
Aggregation bias?
|
bias that occur because of variables on an aggregative level may not represent the association that exists at the individual level
|
|
|
|
What are ecological studies used for?
|
observation across groups and time based upon wide variations in disease rates and in exposure
|
|
|
|
Case series used for?
|
establishing hypothesis and ideas
|
|
|
|
What is case series?
|
observation of patients across TIME
|
|
|
|
Cross sectional study?
|
data at ONE POINT IN TIME that assesses the frequesnce of disease ie incidence
|
|
|
|
What is a case-control study?
|
copares a group of people with diease without (and asks, what happened)
|
odds ratio used
|
|
|
Observational and retrospective study?
|
case-control studye
|
|
|
|
What is observational and prospective study?
|
cohort study to assess risk factors
|
relative risk
|
|
|
Cohort study
|
compares grp w given risk factor to a group without to assess risk factor, asks what will happen?
|
|
|
|
Disease prevention: primary
|
Prevent occurence
|
PDR
|
|
|
Disease prevention: secondary:
|
Detection
|
PDR
|
|
|
Disease prevention: tertiary:
|
Reduce disability
|
PDR
|
|
|
Clinical trial participants
|
already sick
|
|
|
|
Field trial participants
|
not yet sick but diasease is prevalent in population
|
usually more severe and common in population because waiting around for Dz is expensive
|
|
|
internal validity
|
the observed differences can ONLY be attributable to the hypothesis under study
|
|
|
|
external validity
|
generalizability of the test results: can study produce correct inferences in larger population?
|
|
|
|
gold standard for interventional studies?
|
randomized controlled trials.
|
|
