Methods to Control for Confounding
Study Designs
Statistical Tests
Interpretation of Measure of Association
Epi Equations
100

Separating sample into groups and analyzing the groups 

What is stratification

100

Assess outcome and exposure at same time
Generates hypotheses
Prevalence can be calculated
Cannot demonstrate causality 

What is cross-sectional study 

100

Dichotomous outcome, comparing 2 groups 

What is chi-square

100

Relative risk = 2.9
Exposure: Poor diet
Outcome: Diabetes

The risk of diabetes among those with poor diet is 2.9 times the risk of diabetes among those without poor diet. 

100

# new cases of disease in a time period / total person-years

What is incidence rate

200

Ensuring equal distribution of confounders in experimental/intervention studies

What is randomization

200

Population level data is utilized
Quick and inexpensive
Can't confirm correlation because individuals are not interviewed

What is ecological study 

200

Continuous outcome, comparing 2 groups 

What is t-test

200
Odds ratio = 10.2

Exposure = Bee stings
Outcome = Heart attack

The odds of bee stings among those with heart attacks is 10.2 times the odds of bee stings among those without heart attacks.

200

Number of people with a disease in a population at one point in time/ Total in the given population at same point in time

What is prevalence

300

Ensuring equal distribution of confounders and must be maintained in both design of the study and analysis. 

What is matching 

300

Exposure is first assessed
Can demonstrate temporal relatioships
Rare exposures can be looked at
Can look at multiple outcomes 

What is cohort study

300

Continuous outcome, comparing 3+ groups

What is ANOVA

300

Relative risk = 0.32
Exposure =Seeing a puppy
Outcome= Dying of cuteness

The risk of dying of cuteness among those who saw a puppy is 0.32 times the risk of dying of cuteness among those who did not see a puppy.

300

(# dying from disease / # people with disease) X 100

What is case-fatality rate 

400

Limits study participants, but may limit generalizability of the study 

What is restriction 

400

Participants chosen based on outcome
Good for rare diseases
Relatively inexpensive
Subject to recall bias

What is case-control study 

400

Continuous outcome, comparing many groups

What is linear regression

400

Odds ratio =1.00
Exposure = Pollen
Outcome= Getting a haircut

There is no association between being exposed to pollen and getting a haircut. 

400

(# dying from disease / # deaths in population) X 100

What is proportionate mortality rate

500

Controls for multiple confounders at once and involves construction of a mathematical model 

What is multivariable analysis

500

Exposure is assigned/controlled by researchers
Most powerful way to measure cause and effect relationships
Individually-based 

What is a clinical trial

500

Dichotomous outcome, comparing many groups

What is logistic regression 

500

Risk ratio = 0.56
Exposure = Rock climbing
Outcome = Head injury 

The risk of head injury among those who rock climb is 0.56 times the risk of head injury among those who do not rock climb. 

500

(Incidence in placebo group - Incidence in treatment group) / Incidence in placebo group

What is efficacy