Epi Stats

Year 1 - Normal Animal

pneumonics for research questions

be FINER and work SMART

FINER

feasible, interesting, novel, ethical, relevant

SMART

specific, measurable, achievable, realistic, time-bound

null hypothesis

theory that the factor being investigated has no effect on another factor

alternative hypothesis

theory that the factor being investigated does have an effect on another factor

stages of a research project

1. planning

2. design

3. data collection

4. data processing

5. data analysis

6. presentation

7. interpretation

8. publication

type 1 error

rejecting a true null hypothesis

type 2 error

accepting a false null hypothesis

types of qualitative categoric data

binary, nominal, ordinal

types of quantitative data

discrete, continuous

binary data

data where there are two possible options

nominal data

categoric data where the options are names e.g. breed, species

ordinal data

data where the options are in an order e.g. small, medium, large

causal inference

the process of establishing and quantifying a cause-and-effect relationship between two variables, differentiating it from mere association by determining the independent impact of a cause on an effect

causal web model

way of conceptualising and representing how the complex web of multiple factors can combine to cause disease, requires knowledge about likely associations/causal structure

causes of non-causal associations

chance, reverse causation, bias, confounding

reverse causation

where the assumed outcome is the actual cause of the exposure, reversing the hypothesized causal link

uses of measures of disease occurence

shows how much disease is present, relates disease to a population at risk, gives a measure of a time period disease is occuring over

examples of basic measures of disease

number of cases, proportions, rates

disease prevalence

the proportion of animals in a defined population that has the disease of interest at a specific time

formula for disease prevalence

number of cases at set time/number of individuals at set time

point prevalence

prevalence recorded at a single point in time

period prevalence

prevalence recorded over a period of time

disease cumulative incidence

the proportion of healthy individuals that develop a disease over a specific time period out of those at risk

formula for disease cumulative incidence

number of new cases during a specific time period/ population at risk during the same period

what does disease prevalence show?

how widespread a disease is, the chance that an individual has a disease, risk of exposure if you contact an individual at random, impact on a population of a disease

what does disease cumulative incidence show?

the risk of contracting a disease

what does a relative measure of disease association quantify?

the per capita change in risk of contracting a disease caused by a risk factor/exposure

2 relative measures used

relative risk, odds ratio

relative risk formula

risk of disease in exposed animals/ risk of disease in unexposed animals

odds ratio formula

odds of disease in exposed animals/ odds of disease in unexposed animals

when to use relative risk to analyse data

cross-sectional data (taken at a single point in time), cohort (observing a population over time and seeing who gets the disease and who is exposed)

when odds ratio is used to analyse data

case-control (select cases and match them to controls and ask who had the exposure)

what do absolute measures quantify?

the impact of a risk factor on a population

2 absolute measures we use

attributable risk, attributable fraction

formula for attributable risk

risk of disease in exposed animals - risk of disease in unexposed animals

attributal fraction formula

(risk among exposed animals - risk among unexposed animals)/ risk among exposed animals

measures of average

mean, median and mode

when to use the mean to analyse data

when data is symmetric and evenly distributed

when to use median to analyse data

when data is skewed

Henle-Koch postulates criteria for an agent to be a cause of disease

present in every case of the disease, isolated and grown in pure culture, cause specific disease when inoculated into a susceptible animal

Evans’ unified concept of causation criteria for an agent to be a cause of disease

proportion of individuals with disease is higher in exposed than unexposed, exposure to the cause is more common in cases than in those without the disease, number of new cases is significantly higher in exposed than unexposed, the disease follows exposure to the cause, measurable spectrum of host responses follow exposure along a biological gradient, experimental reproduction of the disease occur more in exposed animals, elimination of the cause results in a lower incidence of the disease, preventing the host response decreases the expression of the disease