Concept Ratio Proportion
Equation Event or People (A)/ Event of People (B)
Units None %
Use Descriptive
Pros/Cons
Rate
# events
Person- time experienced to dev p event yrs # events % # subjects available to experience event in T # with characteristic % # ppl at risk Point = specific time pt; period = time interval OR Incidence x Disease Duration # new events Person- total time (pop'n)at risk yrs # new events Person- actual time at risk years #nonfatal new cases Pop'n at risk # deaths Total pop'n # deaths # new cases #new cases pop'nat risk for limited T of obs'n Age at death -- predetermined age at death years (predetermined standard = 65)
Frequency of event
Risk
Prevalence
Burden disease in pop'n
Affected by survival; no measure risk; mix chronic/acute cases
Cumulative/Crude Incidence (CD) Incidence Density (ID)
Assume entire pop'n followed through; always smaller than ID Not include time not followed up
Disease causation
Morbidity rate
Mortality rate
Case-fatality
Attack rate
Years of Potential Life Loss (YPLL)
Premature mortality index
Research/resource priorities, surveillance trends/interventions
O E b UE
Epidemiological Study Designs. Descriptive -- Describe health events with attention to person, place, time. Generate hypothesis and resource allocation.
Concept Relative Risk (RR)
Equation Rate E Rate NE
Use Likelihood dev'ng Outcome (O) in Exposed (E) group relative to Unexposed (UE) group; ratio incidence in E vs. incidence in UE Rate of O that can be attributed to the E in the E group
Interpretation Those with [E] are [RR%] [more/less] likely to develop [O] than those with [UE]
a
q+p C p+3 Rate-Rateue
Attributable Risk/Risk Difference (AR)
[AR#] cases of [O] per [x ppy] among the [E] group can be attributed to [E]
a+b
p + 5 AR
Attributable Fraction Exposed (AR%)
100. Incidence in E = 100 RR Incidence total Incidenceue OR AR x PrevelanceE PAR 100 Incidence total
Proportion of the disease amongst the E that [AR%] of [O] amongst the [E] group can be is attributable to the E attributed to [E]
Population Attributable Risk (PAR)
Excess rate of O in the total population that is attributable to the E
[PAR]excess cases of [O] per [x ppy] in the population can be attributed to [E]
Attributable Fraction population (PAR%)
Proportion of O in total pop'n which is attributable to E; assume causation
[PAR%] of [O] in the population can be attributed to [E]
n
a+b(RR-1)+1 a/c po Ratio of odds of E in O* to odds of E in O; for Cases were [OR] times more likely than b/d= bc case-control studies. No change with O" #; no controls to have been exposed change if examine % E in O/O'. Equals RR if O rare (prevalence <5%) 1p!p Z Used for stratified case-control studies if OR Cases were [ORMH] times more likely than ni b/w strata are similar. controls to have been exposed after Lbici adjusting for [confounder] ni
Odds Ratio (OR)
Mantel-Haenszel Summary Odds Ratio (ODMH)
Causal Relationships Direct: factor -> disease;