Lec 2 TB Flashcards
2x2 or fourfold table
Margins
Cells
2x2 or fourfold table
2 categories of exposure (Y/N)
2 categories of disease (Y/N)
Margins: outermost row and column (eg a + b)
Cells: Inner area (eg a,b,c,d)
2x2 table when comparing incidence rates: 2 modifications
2x2 table when comparing incidence rates
Modifications
#1: remove # of ppl w/o disease (don’t need in incidence calculations)
#2: substitute counts w/ p-t in margin
2 ways to compare measures
1: absolute comparisons: subtract from another, difference
Compare measures of disease freq in 2 ways
#2: relative comparisons: divide by one another
Rate or risk difference formula
Define RD
Re
Ru
RD interpretation
If RD is w/ assumption of causality, RD interpretation
RD = Re – Ru
RD: rate or risk difference (e.g. incidence rate diff, cumulative incidence diff, prevalence diff)
Re: rate or risk (incidence rate, cumulative incidence, prevalence) in exposed gp
Ru: rate or risk (incidence rate, cum incidence, or prevalence) in unexposed gp
Interpretation: excess rate or risk of disease (RD) among exposed pop
RD w/ assumption of causality
If the exposure is the case of disease, RD is used to calculate the # of disease cases that will be elim if the exposure is elim (reduced to the lv of reference gp)
RD graph: where is excess risk?
Graph
NOTE: excess risk or rate x # of exposed ppl = # of excess cases
population
Pop risk or rate diff interpretation
2 ways to calculate
Pop includes exposed and unexposed ppl
Pop risk difference or pop rate diff: impact of exposure on the total pop
Calculation: PRD = Rt – Ru
• PRD: pop rate diff (or pop risk diff)
• Rt: rate or risk (incidence rate, cum incidence, prev) in total pop
• Ru: rate or risk (incidence rate, cum incidence, prev) in unexposed gp
Calculation 2: PRD = RD x Pe
• Another way to get pop rate diff
• RD: rate diff; (i.e. incidence rate diff, cum incidence diff, or prev diff)
• Pe: prop of pop that is exposed
For the total pop (Rt) or prop of exposed (Pe), we can get it from the study pop or general pop
Pop rate or risk diff interpretation
Pop rate diff describes the impact among total pop
• Refers to excess # of cases in total pop that is associated w/ exposure
• OR, if the exposure causes the disease, PRD refers to the # of cases of disease that will be elim in total pop if the exposure were elim (or reduced to lv of ref gp)
PRD help PH officials determine the most key exposure and prioritize prevention efforts
2 other measures of comparison that show PH impact of exposure
Interpretation Attributable prop among the exposed
Formula
Define APe
Re
Ru
1: Attributable prop among the exposed (Ape): prop of disease among the exposed that will be elim if the exposure were elim
2 other measures of comparison that show PH impact of exposure
o #1: attributable prop among the exposed (aka etiologic fraction)
o #2: Attributable prop among the tot pop
o Assumption: exposure causes disease
o Calculation: APe = [(Re - Ru)/Re] x 100
Ape: attributable prop among the exposed
Re: incidence rate, cum incidence, or prev in exposed gp
Ru: incidence rate, cum incidence, or prev in unexposed pop
Interpretation Attributable prop among total pop (APt)
Formula
Define Rt
Ru
What is the value of comparisons (rate diff, pop rate diff, APe, APt) if there is no relationship b/w exposure and disease?
2: Attributable prop among total pop (Apt)
o Describe prop of disease among total pop that will be elim if the exposure were elim
Calculation: [(Rt - Ru)/Ru] x 100
Rt: incidence rate, cum incidence or prev in total pop
Ru: incidence rate, cum incidence, or prev in unexposed pop
Apt helps determine priorities of PH action
o If there is no relationship b/w exposure and disease (rate or risk of disease among exposed = that of unexposed), the value for all absolute comparisons (i.e. rate diff, pop rate diff, APe, APt) = 0
Define prevention fraction
Formula
Define Ru
Re
- Prevention fraction (PF): used when an exposure is thought to protect against the disease
o Formula: PF = [(Ru - Re)/Ru] x 100
o Ru: incidence rate, cum inc, prev of unexposed
o Re: inc rate, cum inc, prev of exposed
3 types of relative measures of comparison
Formula for RR
Define RR
Re
Ru
Interpretation for RR
What does RR tell us
RR = 1
RR < 1
RR > 1
Excess relative rate meaning
Formula
Relative measure of comparison: It is based on 2 measures of disease freq
- Precise terms: inc rate ratio, cum inc ratio, prev ratio
- Calculation
o RR = Re/Ru
RR: rate or risk ratio
Re: incidence rate, cum inc, prev in exposed gp
Ru: incidence rate, cum inc, prev in unexposed go
Interpretation: strength of relationship b/w exposure and disease: the # of time higher or lower the rate or risk is among exposed compared to that of unexposed
- E.g.
Relative rate of death among 6 cities:
(16.24/1000 PY)/ (10.71/1000 PY) => 1.51
IOW: high polluted area has 1.5x higher death rate compared to low polluted area
o Graph
o - RR = 1.0: no relationship b/w exposure and disease
- RR < 1: the exposure prevents disease
o E.g. relative risk = 0.33: exposed gp has half the risk compared to unexposed gp; IOW: 67% decreased risk of disease - RR > 1: +ve relationship b/w exposure and disease (exposure increases rate of disease)
AKA “excess relative rate”
Calculation: (RR – 1) x 100
o E.g. Using the relative rate of death among 6 cities (1.52)
o The excess relative rate: (1.51 – 1) x 100 = 51%
o IOW: ppl living in highly polluted areas has 51% increased rate of death
Preferable type of measure
What happens if we look at the cum incidence for a very long time?
Attributable proportion among exposed formula
Define APe
RR
Attributable proportion among total pop formula
Define APt
RR
Pe
- Epidemiologists prefer relative method of comparison b/c it is based on the baseline value
- Caution in interpretations: If we look at the cum incidence for a very long time, RR approaches 1.0 as the disease freq increases
Attributable proportion among exposed
APe [(RR - 1)/RR] x 100
Ape: attributable prop among exposed
RR: relative measure of comparison
Attributable prop among total prop
APt = [Pe(RR - 1)/Pe(RR - 1) + 1] x 100
Apt: attributable prop among total pop
RR: relative measure of comparison
Pe: exposed prop in either the study or general pop
crude rates
What is it difficult to compare crude rates
Solution
2 methods to find age-standardized rates
Direct method formula
Direct standardization
- Crude rates: rates based on raw data
- It is hard to compare crude rates as the compared gps differ on a characteristic that affects the rate of disease (e.g. age, gender, race)
Need to adjust the variable:
o Age standardized or age adjusted rates: what is the death rate for each state if the pop have identical age distributions
2 methods to calc age-standardized rates: direct and indirect
o Direct method needs
1: find age specific rates in each gp
2: find age structure of a standard pop
o ∑(rate x sample of age I/pop)
