MPH 604:

Epidemiology
CHAPTERS 7 & 8: COHORT AND
CLINICAL/COMMUNITY TRIAL
STUDY DESIGNS
Learning Objectives
Differentiate cohort studies from other study designs.
List main characteristics, advantages, and disadvantages of cohort
studies.
Describe three research questions that lend themselves to cohort
studies.
Calculate and interpret a relative risk.
Differentiate Experimental studies from Observational studies
Describe the key elements of a Clinical or Intervention Trial
Discuss ethical considerations in a clinical trial
Study Design: Cohort Study
Prospective, Cohort, Longitudinal Study

Unit of Measures Individual

Time Period Today -> Future (optional Past)

Today: collect Exposure data
Future: follow cohort to collect Outcome data

Optional Past Data Prospective Cohort Study

Past: can also collect Exposure from past
Study Design: Cohort Study

Gold Standard

Study Design
You design the study, you can minimize bias

Temporality
Start with No Outcome, only Exposure, follow into future
Examine Cause-Effect

Fewer Recall Bias Issues

Compared to Case-Control Study
Study Design: Cohort Study
Cohort - Population, group, or community with a common exposure
E.g. Birth Cohort all born in the same year (1984)
E.g. Same Locale school cohort, occupational cohort (Nurses Health
Study started in 1976)

Cohort Effect
Members of a cohort experience similar influences
E.g. Tobacco Use in USA
Pre-WWI, smoking was rare
WWI troops provided cigarettes as part of rations
After WWI, smoking prevalence increased
Subsequently, lung cancer incidence increased
Study Design: Cohort Study

Timing of Data Collection

DESIGN PAST PRESENT FUTURE
Prospective E D
Retrospective E D
Historical
Prospective E E D
Code: E, exposure; D, disease
Study Design: Prospective Study

Advantages
Gold Standard
Minimize Bias
Examine Cause-Effect
Interventions

Disadvantages
Expensive
Length of Time
Not for Rare Diseases
Loss to Follow-Up
Analysis: Prospective Study

Life Tables
Survival Curves
Relative Risk
Life Table Methods

Estimates for survival during time intervals

Cumulative survival probability at end of interval

E.g. Life tables for survival times of cancer patients in clinical trials.

Two types of life table methods:

Cohort Life Table
Period (Current) Life Table
Life Table Methods (contd)

Cohort life table:

Shows the mortality experience of all persons born during a particular year.

Period life table:

Enables us to project the future life expectancy of persons born during the
year as well as the remaining life expectancy of persons who have attained
a certain age.

Social Security Agency Period Life Table 2009

http://www.ssa.gov/OACT/STATS/table4c6.html
Wade Hampton Frost Cohort Life Table
Social Security Agency:
Period Life Table 2009

http://www.ssa.gov/OACT/STATS/table4c6.html
Survival Curves

A method for portraying survival times

In order to construct a survival curve, the following

information is required:

Time of entry into the study

Time of death or other outcome

Status of patient at time of outcome, e.g., dead or

censored (patient is lost to follow-up)
Survival Curves: Example

15 subjects followed
over 36 months; all
entered the study at
the same time.

Nine died at different

points of the study.

Deaths of two patients

caused a steep drop at
19 months.

Each step indicates the

death(s) of one or
more patients.
Kaplan-Meier survival curves for time to
hepatitis C seroconversion by type of injection-
related risk practice among 352 young adult IDUs
in Chicago, 1997-1999.
Sharing cookers
1
0.9
0.8
0.7
No cookers log-rank p=0.0095
0.6 Any cookers
0.5
0 100 200 300 400 500
Time (days)
Measures of Association
Disease Status
Incidence Rate
Exposure Yes No Totals
Yes A B A+B A/(A+B)

No C D C+D C/(C+D)

Relative Risk [A/A+B] / [C/C+D]

Incidence of disease in the exposed group
Incidence in the non-exposed group

Attributable Risk [A/A+B] - [C/C+D]

amount of risk attributable to the exposure
Relative Risk: another setup
Ratio of cumulative incidence rates (IR)
IRe X 10n
RR =
IRue X 10n

e is exposed
ue is unexposed

EXPOSED IS ALWAYS THE NUMERATOR!

Probability of an outcome (disease) for the exposed group

RELATIVE to the unexposed group
Interpretation of Relative Risk

RR = 1 implies no difference in risk of the

outcome for the exposed relative to the
unexposed group

RR < 1 implies decreased risk of the outcome for

the exposed relative to the unexposed group (i.e.,
the exposure is protective) less likely

RR > 1 implies increased risk of the outcome for

the exposed relative to the unexposed group (i.e.,
the exposure is hazardous) more likely
Observational studies
Descriptive studies:
hypothesis generation
case reports, case series, cross-sectional surveys

Analytic studies:
hypothesis testing
case-control, ecologic, cohort
Hierarchy of Study Designs:
Validity for Etiologic Inference
Study Design: Experimental Study
Community Intervention Study, Clinical Trials

Unit of Measures Individual

Time Period Today -> Future (optional Past)

Today: collect cohort, randomize into groups
Start manipulation of exposure/intervention

Future: follow cohort to collect Outcome data

Past: can also collect Exposure from past

Schematic Diagram of a
Clinical Trial
Study Design: Experimental Study
Randomization

Intervention or Drug
Prophylactic or therapeutic/curative

Outcomes of Experimental Studies

Clinical end points outcome, disease, death, or recovery
Must be measured in comparable manner between control and
intervention group
Example of a Crossover Design

Study Sample

Experimental Control
Group Group

Washout Period

Control Experimental
Group Group
Ethical Aspects of Human Experimentation

Benefits must outweigh risks.

Ethical issues:

Informed consent
Withholding treatment
Sequential designs are used as a
solution.
Monitoring for side effects
Deciding when to withdraw a patient
Protecting the interests of patients
Summary of Clinical Trials

Strengths: Limitations:
Provide the greatest control over:
Artificial setting
the amount of exposure
the timing and frequency of
exposure
Limited scope of potential
the period of observation impact

Ability to randomize reduces the

likelihood that groups will differ Adherence to protocol is
significantly. difficult to enforce.

Ethical dilemmas
Quasi-Experimental/Community Trial No Randomization!
Possible bias in conditions of study; ranked below RCT.