Epi 1

Question 1

Prevalence

Answer 1

The proportion of a population who HAVE the disease at or during a specified time (new + old cases in a population)
P = (total # cases at time t) / (# individuals in pop at time t)
it’s a proportion
ranges from 0-1 or 0% - 100%
the probability an individual will have a disease at a given point in time

Question 2

Cumulative Incidence (risk)

Answer 2

The probability an individual will develop a disease over time; the number of NEW cases per unit of population
CI = (# new cases over specified time) / (population at risk over specified time)
properties:
proportion
ranges from 0-1 or 0-100%
estimated for population but can be applied to the individual

Question 3

Incidence Rate

Answer 3

The rate at which individuals develop a disease in a population
IR = (# NEW cases of a disease over a time) / (person-time over the specified time)
Usually written as per 100 person-years

Question 4

Person-Years

Answer 4

Used as denominator when calculating incidence rate; the sum of individual follow-up times when they are at risk for the disease for all people in the population
Person-time ends if the person develops the disease, or is no longer observed (lost to follow-up, death, study ends, etc)

Question 5

Average Rate in a Population

Answer 5

When incidence rates are not the same over time, we cannot sum the rates and divide by the number of years. We must take a weighted average of the incidence rates in each year using the person-years at risk as the weights.

Question 6

Mathematical Relationship between Cumulative Incidence and Incidence Rate

Answer 6

CI = IRTime
Only applies if IR is constant and IRTime is small
If IR changes with time, risk is calculated in smaller time intervals (survival analysis)

Question 7

Incidence vs. Prevalence (what are they used for?)

Answer 7

Both can be used to measure disease within a population and compare disease frequencies between populations
Incidence rates are used when studying etiologies of disease
Prevalence rates are used mainly when planning health care programs

Question 8

What could a change in an incidence rate reflect?

Answer 8

A change in prevalence of etiologic factors
A change in the amount of risk factor exposure
The effect of a preventive program (i.e. screening)

Question 9

What could a change in prevalence reflect?

Answer 9

A change in incidence rate
A change in duration or outcome of the disease
A change in the immigration or emigration of persons with the disease

Question 10

Mathematical Relationship between Incidence and Prevlanece

Answer 10

P = IR*D
Prevalence = Incidence Rate x Average Duration of Disease
Only applies to diseases whose incidence and average duration are stable over a time

Question 11

Mortality Rate (Death Rate)

Answer 11

A measure of the rate at which individuals in a population die during a specified time period
MR = (# of deaths over time period) / (Total person-time observed over time period)

Question 12

Cause-Specific Rates

Answer 12

Event rates for specific diseases

CSR = (# NEW events of a specified disease over time period) / (total person-time observed over the time period)

Question 13

Sex-Specific Rates

Answer 13

Event rates with gender groupings

SSR = (# new events in specified gender over a time) / (total person time of that gender over that time)

Question 14

Race-Specific Rates

Answer 14

Event rates within race groupings

RSR = (# new events in specified racial group over a time) / (total person years of that race over that time)

Question 15

Case Fatality Rate

Answer 15

The proportion of people, among those who develop a disease, who then die from that disease
It’s actually a proportion, not a rate!
Measures the severity of the disease over a time, can help assess benefits of a new treatment
CFR = (# deaths in a population due to a disease) / (total # of cases of that disease in the population)

Question 16

Proportional Mortality

Answer 16

The proportion of deaths due to a specific cause out of all deaths (helps with disease burden analysis)
PM = (# deaths from a disease in a pop over a time) / (total # deaths in the pop over that time)

Question 17

Crude Rates

Answer 17

The proportion of individuals in a population who experienced the event within the time period. Crude rates are often useless in comparing rates between populations if they differ in age or gender distributions (generally age is the biggest confounder) - i.e. an older pop has a higher crude death rate than a younger one (older people have a higher probability of dying than young people)
Differences in crude rates can be due to differences in age distributions or the event rates within specific age groups. (usually the latter is of most interest)

Question 18

Age Specific Rates

Answer 18

Event rates within age groups (aka ‘age strata’)

ASR = (# new events in an age group over a time) / (total person-time in that age group over that time)

Question 19

Age-Adjusted Rates

Answer 19

Overall event rates that are adjusted for age; sometimes called age-standardized rates
To compute the AAR: choose a standard population, multiply the age-specific rates for the study pop by the numbers of persons in the corresponding age strata in the standard population to get the expected numbers of age-specific events; sum the expected numbers of age-specific events for each population; divide the sum of the expected events by the total standard population.
The AAR is not a true rate - it is a hypothetical rate. It is the rate expected if the population of interest had the same age distribution as the standard pop.

Question 20

Measures of Association

Answer 20

Are used to compare the frequency of disease occurrence in 2 groups.
Summary parameters that estimate the association between exposure and risk of disease - difference measures or ratio measures

Question 21

Risk (or Rate) Difference (RD)

Answer 21

Risk (or rate) of disease among exposed persons that can be attributed to the exposure. A difference of the risk or rate among the exposed and the unexposed (attributable risk, AR)
RiskD = (Cumulative Incidence of Exposed) - (Cumulative Incidence of Non-Exposed)
RateD = (Incidence Rate of Exposed) - (Incidence Rate of Non-Exposed)

Question 22

Relative Risk (RR)

Answer 22

An estimate of the magnitude of an association between exposure and disease - the risk (or rate) of developing the disease in an exposed group relative to those who are non-exposed
Risk Ratio, Rate Ratio, and Odds Ratio are measures of relative risk

Important properties:

Range of values for RR is from 0 to infinity.
RR = 1 indicates no effect/association (H0)
RR > 1 indicates positive association (the higher the value, the stronger the + assn - exposure may be a risk factor)
RR < 1 indicates negative association (closer to 0, the stronger the association - exposure may be a protective factor)

A 95% confidence interval for RR is the interval within which the true RR of the population from which the sample was drawn lies with 95% certainty (it’s computed from the RR and standard error of RR)

When testing whether an estimate of the RR is different from 1: if a p-value is < 0.05 then the value 1 is excluded from the 95% confidence interval, and vice versa (when this occurs, you can conclude the observed RR is statistically significant from 1.0)

RRs measure the strength of association between exposure and disease.

Question 23

Risk Ratio

Answer 23

The cumulative incidence (risk) of disease among those exposed divided by the cumulative incidence (risk) of disease among those unexposed.
RR = CI (exposed) / CI (unexposed)

Question 24

Rate Ratio

Answer 24

The ratio of 2 incidence rates for two groups or populations (the exposed and the unexposed)
RR = IR (exposed) / IR (non-exposed)

Question 25

Odds Ratio

Answer 25

Commonly used in case-control studies, where participants are selected on the basis of disease status so it’s not possible to calculate the rate of developing the disease given the presence or absence of an exposure
The incidence rate ratio is estimated by calculating the ratio of the odds of exposure among those with disease (cases) to that among those without disease (controls)
OR = (odds of being exposed among diseased) / (odds of being exposed among non-diseased)

In a 2x2 table, OR = (a/c) / (b/d) = ad/bc
disease+ outcome-
+ exposure a b
- exposure c d

Question 26

Attributable Risk (AR and AR%)

Answer 26

The proportion of disease among the exposed that is due to the exposure - the proportion of disease among the exposed that could be prevented by eliminating the exposure (aka attributable proportion, attributable fraction, etiologic fraction)

AR = IR(exposed) - IR(non-exposed)

AR% = [IR(exposed - IR(non-exposed)] / IR(exposed) x 100%
AR% = [(Relative Risk - 1) / Relative Risk] x 100%

AR and AR% measure the expected potential benefit among exposed individuals if they could eliminate their exposure.

Question 27

Preventable Risk (PR and PR%)

Answer 27

Essentially the inverse of the AR
If the exposure is protective, the rates in AR equations are reversed so the higher (non-exposed) rate occurs first in the numerator and is also used in the denominator. This is called the preventable risk and preventable risk percent (aka preventable proportion/fraction).
In a clinical trial in which the exposed group gets an experimental treatment and the non-exposed group gets a placebo/control treatment, these values are called the absolute risk reduction and the relative risk reduction (ARR and RRR), respectively (can be assessed using cumulative incidence or incidence rates)

Question 28

Number Needed to Treat (NNT)

Answer 28

The number of individuals that would need to be treated (or exposed to a protective factor) to prevent a single event can be estimated by taking the reciprocal of the Absolute Risk Reduction
NNT = 1/ARR

The reciprocal of an Attributable Risk gives the number of individuals exposed to a risk factor that would need to become unexposed to prevent a single event.

Question 29

Population Attributable Risk (PAR and PAR%)

Answer 29

The amount or proportion of disease in the population that can be attributed to the exposure (can be assessed using cumulative incidence or incidence rates). Computed as the difference between the rate (or risk) in the entire population and the rate (or risk) among the unexposed. Usually expressed as a percent of the rate in the population. PAR is used to measure the potential benefit to be expected in the entire population if the exposure of interest were eliminated.

PAR = IR(population) - IR(non-exposed)

PAR% = [IR(pop) - IR(non-exposed)] / IR(pop) x 100%