18 - analysing pop data Flashcards
(17 cards)
What Toxicology Questions use Population Data?
- Does asbestos exposure increase risk of lung cancer?
- Does exposure to parathion (an OP pesticide) increase risk of PD?
- Does talking on your mobile phone give you brain cancer?
The ‘Browne’ hypothesis and asbestosis
Asbestosis is an inflammatory disease not asbestos exposure
* Asbestosis better correlation with Lung CA than asbestos exposure
Parathion expsoure and risk of developing PD?
MPPP use (synthetic heroin) ➔ dopaminergic cell death
* Contaminated with MPTP
* Association exists but causal relationship not proven
Does talking on your mobile phone give you brain cancer?
Very difficult to measure exposure
* Many types of tumors
* Evidence fails to support idea
Epidemic
Out-break due to exposure of a group of persons to a noxious influence that is common to the individuals in the group
Endemic
The constant presence of a disease or infectious agent with a given geographic area or population group”
Pandemic
An epidemic occurring worldwide or over a very wide area
Investigating an epidemic
- Compare cases to how many were expected
- Interview cases (date of onset, sources of food and water, previous exposure)
- Matched controls
Index case
The first case in an epidemic
Incubation period (in pop. data)
Distance from index case to middle of distribution of outbreak
Attack rates
Cumulative incidence in group over some period
* Often used to compare those exposed to various possible sources of infection
Looking at cases and controls to find source of outbreak
Subtract exposure rates controls from cases
* Looking at areas, food, interaction
Toxic Oil Syndrome Spain 1981
“Black market” cooking oil (rape seed oil denatured with aniline to prevent consumption)
* Meant for industrial purpose
* Criminals attempted to remove aniline ➔ produce new toxin
Effect of intervention in Epidemic
Spanish Gov. replaces “black market” oil with olive oil for free
Calculating relative risk
a/(a+b) / c/(c+d)
* Ratio of outcomes in exposed versus outcomes in unexposed pop.
* Cohort studies
Calculating Odds Ratio
ad / bc
* Still use contingency matrix
* Ratio odds of exposure in cases to controls (rather than disease rate)
Logistic Regression
log(P/(1-P)) = ß1x1 + ß2x2 + ß3x3 + ß0
* When multiple independent var. need to be considered
* X is the var and ß exponentiated to produce OR
