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;
64 Cards in this Set
- Front
- Back
|
Define rate from slides in class. Give epi example.
|
The changes in one quantity per unit change in another quantity. (example: incidence density)
|
|
|
Define ratio from slides. Give an example of one with a dimension and one without.
|
A fraction where the numerator is not included in the denominator. (example w/ dimension: # of hosp. beds/1,000 people; example w/o dimension: # of epi students/ non-epi students among MPH students
|
|
|
What is a common epidemiological ratio? Formula? (hint: only actual measure you can get from case-control data)
|
odds ratio
[ (# Dexp/ Not D w/exp) / (# D w/o exp / # not D w/o exp) ] |
|
|
Define prevalence ( aka prevalence proportion) and give an epidemiological example.
|
The proportion of a population with a disease determined at a specific point in time.
|
|
|
Define odds in an epidemiological manner.
|
The # of persons in a population with a disease divided by the the # of persons in a population without a disease.
|
|
|
Give an example of prevalence using swine flu in the berkeley school of public health.
|
There are 5 cases of swine flu amongst the 200 MPH students. Thererfore PP = 5/200 = 0.025
|
|
|
Define incidence. Give an example. Is it a rate, ratio or proportion?
|
The frequency of development of new cases of disease in a population. There was 5 cases of swine flu in 200 person-years. It is a rate.
|
|
|
Define cumulative incidence.
|
The proportion of subjects at risk for a disease at baseline who develop the disease during a certain period of time. (ex. 20 nervous break downs per one week of finals at Berkeley)
|
|
|
Does cumulative incidence (CI) have a dimension?
|
No bc it is a proportion and ranges from 0 to 1.
|
|
|
In 18 months, out of 10,000 shoppers visiting the dollar store, 25 of them come down with cleptomania due to the economic crisis. What is the CI for the 18 mo. period?
|
CI = 25/10,000 = .0025 or 0.25%
|
|
|
Define incidence density. Give an example.
|
The # of incident (aka new cases) divided the total person time. (ex.20 cases of diptheria in 8,000 person years)
|
|
|
What is the unit for prevalence?
|
cases per persons at risk
|
|
|
What is the unit for an odds ratio?
|
None --it is dimensionless
|
|
|
What is the two different units for incidence density?
|
# of cases per person time OR 1/ time (according to Klienbaum)
|
|
|
1,000 men with no history of CVD are observed for 24 months. 10 of them develop CVD on the first day of the 13th month. What is the cumulative incidence?
|
CI = 10 / 10,000 = 0.001 or 0.1% per 24 months or 0.001 per 2 years.
|
|
|
Define crude rate.
|
The # of experiences or events in a total population in certain period.
|
|
|
Give two common examples of crude rates.
|
1. crude birth rate
2. crude death rate |
|
|
Define adjusted rate.
|
A mathmatical transformation of a crude rate to account for differences in population structure (i.e. age, race, etc)
|
|
|
Define case fatality rate
|
# of cause-specific dths among cases/ # of incident cases of the same cause
|
|
|
Define attack rate
|
# of incident cases of disease/ total population at risk over a restricted period of time
|
|
|
What is attack rate commonly used for?
|
outbreak investigations
|
|
|
Is case fatality rate a rate, ratio, or proportion?
|
Although is says "rate" in its name, it is actually a true proportion
|
|
|
What is postnatal mortality rate?
|
index of the risk of death in infants aged 1- 11 months
|
|
|
Define postneonatal mortality rate.
|
# of dths in children aged 1-11 months / # of live births during that calendar year
|
|
|
What is infant mortality rate most commonly used for?
|
risk of death during the first year of life
|
|
|
Define infant mortality rate.
|
# of infant deaths in a calendar year/ # of live births during the same calendar year
|
|
|
Define crude birth rate. How is it expressed?
|
# of live births during a given time interval / # estimated total population at a given mid- interval
Expressed as per 1,000 |
|
|
Define curse fertility rate. How is it expressed?
|
# of live births during a given time interval/ estimated # of women aged 15 -44 years
Expressed per 1,000 |
|
|
Define maternal mortality rate.
|
# of deaths in a year from puerperal causes/ # of live births during the same year
**all live births, not all pregnancies |
|
|
How is maternal mortality rate expressed?
|
per 100,000
|
|
|
Define the crude rate of natural increase in terms of natality rates.
|
(# of live births - # of deaths during a given interval) / estimated total population at a given interval
|
|
|
How is crude rate of natural increase expressed?
|
per 1,000
|
|
|
Define low birth rate ratio.
|
# of live births under 2500 gm during a given interval / # of live births during a given interval
|
|
|
Give the formula for prevalence using incidence of a rare disease and does not have long duration.
|
P = Incidence * duration of disease
|
|
|
What is the formula for prevalence using incidence when the disease is not rare OR the disease has a very long duration?
|
P = (ID* T) / (ID * T) + 1
|
|
|
Give the formula to estimate risk using the simple cumulative method.
|
R = CI = I (t0 , t) / N'0
The # of incident cases between time zero and the end of the study divided by the total number of participants at the beginning of the study (disease free of course) |
|
|
What is sensoring in terms of risk estimates?
|
The inaccurate portrayal of risk estimates because participants were not followed the entire period or the lack of information conc???
|
|
|
How does one estimate the "effective number" of persons at risk using the actuarial method?
|
= (N'o - W/2)
The number of persons beginning the time interval minus half of the # of withdrawls. |
|
|
What measure of disease is used to test "etiologic hypotheses"?
|
Incidence density
|
|
|
What are the three ways to measure person time?
|
1. sum of how much each individual time each person was in study
2. "steady state" pop --people come in and out often enough to assume PT = N * dt 3. Don't know indiv. PT as there is no "steady state" so just use mid-interval disease free population as the avg. pop. size |
|
|
What are the three ways to calculate CI or risk?
|
1. simple cumulative
2. actuarial 3. density method |
|
|
What assumption is made when using the simple cumulative method?
|
that are study participants are followed for the entire study either till the end or until they get disease --aka no death to competing causes or withdrawls
|
|
|
What is an assumption made when using the actuarial method to calculate risk?
|
Withdrawls occur mid-interval of observation time
|
|
|
What a major assumption made when using the density method?
|
a closed cohort with no competing risk and the age interval ID remain constant over that ID(not sure what this means--will ask in OH)
|
|
|
What is the most exact method for calculating risk?
|
density method
|
|
|
What are the three assumptions that must be met in order to use CI instead of ID in the density method?
|
1. rare disease (< 10%) --from rare disease assumption
2. event is inevitable --no competing risks (rate is constant w/in each age interval) 3. each age-specific ID is constant over that time period |
|
|
Which of the three methods for calculating risk assumes not censoring or withdrawls?
|
simple cumulative
|
|
|
What measures of disease can be used for cross-sectional studies
|
1. point prevalence --only existing cases included
2. period prevalence -- longer time frame, can include both new and existing cases |
|
|
What MOD is best for follow-up studies/prospective studies?
|
Since only new cases are measured--cumulative incidence or incidence density
|
|
|
ratio, rate, or proporiton:
1. CI 2. ID 3. prevalence 4. odds 5. risk ration 6. rate ratio |
1. proportion
2. rate 3. proportion 4. ratio 5. ratio 6.ratio |
|
|
What method uses rates to calculate risk?
|
Density method
|
|
|
Risk or rate?
1. instaneous potential for change in disease status per unit change in time 2. cumulative incidence 3. probability that an indiv. will develop outcome during a specified time pd., conditional on not dying from something else 4. incidence density 5. 0- infinity 6. has no units 7. 0 -1 8. units -- cases per person time 9. used to predict invid. change in disease status & test etiologic hypotheses if risk pd is restricted 10. test etiologic hypothes if risk time is extended |
1. rate
2. risk 3. risk 4. rate 5. rate 6. risk 7. risk 8. rate 9. risk 10. rate |
|
|
Can you compare SMR's between studies?
|
Nope
|
|
|
Can you compare SRR's between studies?
|
Yep.
|
|
|
What are 4 ways to reducing confounding?
|
1. stratificatin
2. adjustment 3. propensity scores 4. inverse probability of treatment weighting (IPTW) |
|
|
What are some common adjustment techniques?
|
1. linear regression
2. logistic regression 3. poisson regression 4. proportional hazard models |
|
|
What is simpson's paradox?
|
thinks hats --if that helps
it is the reversal of cause/effect when the samples are combined |
|
|
List three methods in study design to reduce confounding.
|
1. Matching.
2. Randomization 3. Restriction |
|
|
List two method in study analysis to reduce confounding.
|
1. Multivariate analysis
2. Stratify |
|
|
List the three criteria for confounder
|
1. not on the causal pathway
2. a risk factor for the disease in the unexposed 3. be associated w/ exposure variable |
|
|
True/False
1. If it is a confounder then it has to be an effect modifier 2. If it is an effect modifier then it has to be a confounder |
1. False
2. False |
|
|
What it the null hypothesis for the test for homogenity?
|
The pooled estimator (ORmh) is not different from the stratified estimates.
|
|
|
How do you calculate the degrees of freedom for the test for homogenity?
|
N (# of strata) - 1
|
|
|
What does rejecting the null hypothesis indicate?
|
That there may be effect modification and that it is necessary to report the stratified estimates rather than the overall effect estimator
|
