EPIB Exam #2

Study Design:

A plan for data collection, analysis and interpretation

• Descriptive

  • Exploratory investigations

  • Address who, what, where and when 

  • Describe occurrence of exposures and outcomes 

  • Often generate hypotheses 

• Analytic studies 

  • Design is appropriate for testing a research question 

  • Address how and why

  • Assesses associations between exposures and outcomes

General Classification of Study Designs

• Observational 

  • Exposures occur in a “non-controlled” environment

  • Researchers observe and describe cases or relationships among variables

• Experimental

  • Researchers have control over the exposure and who is exposed

• Meta-analyses

  • Combines the results of several studies that address a set of related research hypotheses

Ecological Study Design

Observational approach that analyzes the relationship between exposure and outcome at the population or group level.

Ecological fallacy

Associations at the group level may not represent the exposure-outcome relationship at the individual level. Therefore, you cannot assume evidence of link between exposure and disease at the individual level

Ecological studies: advantages & disadvantages

• Advantages

  • Quick, simple, inexpensive

  • Good approach for generating hypotheses when a disease has unknown etiology

• Disadvantages

  • Ecological fallacy

  • Data is all at group level, not individuals

  • Imprecise measurement of exposure and disease

Correlation: Pearson’s r

• Correlations can be measured using Pearson’s coefficients “r”

• R can take on values range of -1 to +1

• Smaller absolute values of r are weaker 

• Larger absolute values of r are stronger

• Positive vs negative tells direction

Cross-Sectional Studies

 studies are a type of observational study design; investigator does not manipulate the exposure

• Outcome and the exposures are measured simultaneously at a single time point

• Provides a “snapshot” of a population

• The unit of observation is the individual 

• Analysis is also focused on the individual 

Advantages and disadvantages

• Advantages 

  • Relatively quick and inexpensive 

  • Good for measuring long-term exposures and immutable characteristics 

  • Can be generalizable if based on general population 

  • Multiple exposures and outcomes can be studied

• Disadvantages 

  • Does not provide incidence data 

  • Not good for low prevalence or short-duration diseases

  • Cannot determine temporality of exposure and disease

  • Limited usefulness for inferring disease etiology

Measure of Association

• Chance of something happening can be expressed as a risk and/or odds

• Risk - chances of something happening / chances of all things happening 

• Odds - chances of something happening / chances of it not happening

Defining Odds

• Odds are the ratio of the probability of an event occurring to that of it not occurring 

• Odds - # of events of interest / # of events NOT of interest 

• Possible range is from 0-infinity (low to high odds)

Odds Ratio Interpretation

• OR = 1 implies no association 

  • The odds of [exposure] among the [cases] is the same as odds of [exposure] among the [controls].”

• OR < 1 pimples cases group has lower risk of exposure 

  • Easiest to interpret as a percent: “The odds of [exposure] among the [cases] is [(1-OR)x100%] lower than among the [controls].”

• OR < 1 = proactive factors

• OR > 1 = risk factor

Cohort Studies

Observational study designs that describe a group of people selected based on a common characteristic

Prospective Cohort Study

Measure association between an exposure and risk of disease in the population, moving forward in time

Advantages and Disadvantages of Prospective Cohort Studies

• Advantages 

  • Collection on lots of exposures 

  • Ability to establish a temporal sequence 

  • Efficient for rare exposures 

• Disadvantages 

  • Expensive

  • Time-consuming 

  • Inefficient rare outcomes 

  • Inefficient for diseases that take a long time to develop (with long induction and latency periods)

Retrospective Cohort Study

Measure association between an exposure and risk of disease in the population, moving backward in time

Retrospective Advantages

• Advantages 

  • Efficient for diseases with long latency periods 

  • Temporal sequence only when good historical records are used (not guaranteed with this study design)

  • Relative low-cost, quick to accomplish only when historical records are used

Risks

Absolute risk for exposed groups: A / (A+B)

• Absolute risk for unexposed groups: C / (C+D)

• Risk difference: (A / (A+B)) - (C / (C+D))

• Relative risk: A / (A+B) / C / (C+D)

• Interpretations 

  • Relative risk

    • [Exposed group] had [RR value] times the risk of [outcome] compared to the [unexposed group]

  • Risk difference

    • [Exposed group] had [RD value] more/fewer cases of [outcome] per [rate base] compared to [the unexposed group]

Experimental Studies

• Manipulations of variables → assigning to exposure group

• Only use when feasible and ethical

• Types: quasi-experimental and randomized control trial

Phases of Clinical Trials

• Phase 1 - tests for safety on healthy individual

• Phase 2 - tests for side effects on a larger group

• Phase 3 - tests for long term effectiveness and a wider demographic 

• FDA approval - safe for public use

• Phase 4 - can be take off market and continues to test for effectiveness

Measures of Association

• Relative risk

• RR = 1, no association 

• RR>1, treatment group more likely to develop the outcome 

• RR<1, treatment group less likely to develop

Strengths of experimental designs

• Provides the strongest evidence of association 

• Manipulation means having control 

  • Amount of exposure 

  • Timing and frequence

  • Period of observation

Limitations of experimental designs

• Artificial setting 

• Expensive

• Ethical concerns