public health - exam 2

descriptive study design

Describes only – does not examine relationships

• Surveillance data (study)

• Case reports

• Case series

analytic study design

Examines relationships / Tests hypotheses

analytic study design - observational

investigator observes

• Ecologic

• Cross-sectional

• Case-control

• Cohort

analytic study design - experimental

Investigator assigns exposure

• RCT

cross-sectional studies

• Snap-shot of population(s) at a single point in time

• Exposure and outcome data collected at the same time

• Participants selected without regard to exposure or disease status

point prevelance

from cross-sectional studies

number of people with an event at a specific time point divided by the size of the population

(disease status (top) vs exposure status (side))

disease prevelance

number of cases (diseased) at specified time/total population

exposure prevelance

number of people exposed at specified time/total population

cross-sectional studies strengths

good for generating hypothesis

temporality is not a problem for time invariant exposures

relatively low cost

does not require follow up

cross-sectional studies limitations

temporality (reverse causation bias)

subject to non-exchangeability

subject to selection bias

does not measure risk

not useful for rare diseases

prevelant cases may no be same as incident

case-control study

backwards directionality (tracing back to exposure status based on outcome)

(outcome staus (top) exposure status (side))

What measures of disease frequency can we estimate from a case-

control study?

prevalence of exposure and odds ratio (measure of association)

cases

sample (or all) people who developed outcome

controls

sample of people who did not develop outcome

case-control studies - advantages

good for rare outcomes

good for oucomes with long latency periods

study many (detailed) exposures

efficient with time and money

require smaller sample size

case-control studies - disadvantages

usually only one outcome per study

temporal ambiguity (exposure is measured after disease has occured)

inefficient for rare exposures (very few cases or controls will have rare exposure)

genealizability concerns (over-represent people who live longer with disease)

suceptible to selection and information bias (retrospective info)

cohort study

gold standard of observational study designs

follow a population of exposed and unexposed individuals at risk of developing an outcome to compare the development of hte outcome between groups

(disease (top) vs exposure status (side))

cumulative incidence

number of poeple who became newly ill during the time/number of people at risk during time

incidence rate

number of people who became newly ill during time/person time at risk

cohor studies - strengths

temporality

can estimate incidence of disease

can consider time-varying information

can study adverse exposures or those that cannot be randomzied

good for rare exposures

can study multiple outcomes/exposures

serve as the base population for other study designs

cohort studies - limitations

require long-term follow up

high cost

not practical for rare diseases

loss to follow-up

differences in those who are exposed and not can bias results

RCT study

gold standard of study designs

randomly assign people to either receieve or not receive the exposures

control group represents what would have happened without exposure/intervention/treatment

(disease status vs exposure)

intent to treat

Analysis in which data are evaluated based on the group that participants were randomly assigned, not in the group they eventually ended up in

*ITT answers the question, “What is the effect of being assigned to a treatment?

RCT - strengths

strongest evidence for causality

if blinded, less observer bias

calculate measures of disease incidence

randomization can minimize effects of confounding

unlikely misclassification of exposure

RCT - limitations

not “real life” so limitations of generalizability

high coset and very time consuming

could be inapporpriate or unethical

could be poor compliance and loss to follow-up

prevlalence ratio

prevalence of outcome in exposed/prevalence of outcome in nonexposed

What study designs can be used to calculate a prevalence ratio of the outcome?

• Cross-sectional study

• Cohort Study (after follow-up)

• RCT (after follow-up)

risk ratio

cumulative incidence of disease in exposed/cumulative incidence of disease in non-exposed

(same equation and values as prevalence ratio)

What study designs can be used to calculate a risk ratio?

• RCT

• Cohort study

rate ratio

incidence rate of disease in exposed/incidence rate of disease in non-exposed

disease and person time (top) vs exposure status (side)

What study designs can be used to calculate a risk ratio?

• RCT

• Cohort study

odds ratio

probability that an event will occur divided by the probability that an event will not occur

(#cases exposed/# cases not exposed)/(# controls exposed/#controls not exposed)

case and control top exposure statues sides

direction of measures of association

• Values = 1.0 means that there is no association

• Values > 1.0 indicate a positive association

• Values < 1.0 indicate a negative/inverse association

magnitude of measures of association

The value tells you how many times the (prevalence, risk) of the outcome the exposed group has vs. the unexposed group

confounder

associated with exposure (in base population)

associated with outcome (among unexposed)

not in causal pathway

statistical adjustment

include confounders as covariates in a regression model

stratification

analyze the exposure-outcome association within levels (strata) of a confounder

mediation

• Mediation explains how or why an exposure affects an outcome

• A mediator is a variable on the causal pathway between the exposure and the outcome (indirect)

effect modification

• Effect modification is when the relationship between exposure and outcome is different depending on a third variable.

quantitative effect modification

: Association between exposure and outcome is in same direction across strata but the strength of the association varies.

qualititative effect modification

: Association between exposure and outcome is in opposite direction across strata or is present in some strata but absent (null) in others