MI580 Lecture 3 - Intro to Study Design and Clinical Trials

Goals of epidemiology studies (2)

1. Collect information that will help characterize a disease, situation, or perception

2. Describe relationship (association) between exposure (risk factor or possible cause) and disease (outcome)

Exposure

Something that could be linked to an increased or decreased risk of the disease or outcome

Outcome

Interchangeable with disease or health related event

Observational Study

The researcher is not changing anything that happens to the subjects (other than taking measurements). Self-selected exposure.

Interventional Study/Experimental Study

Investigator causes the subject to have their particular exposure status. Randomly allocated exposure, always prospective.

Strong evidence for cause and effect.

Primary Data Collection

Data collected specifically for the study

Secondary Data

Data from previous studies, databases, or other records.

Descriptive Study

Generally collect information on one group with some similar characteristic (e.g. disease status, subgroup of interest)

No comparison group

Purpose is to look for patterns

Analytic Study

Collects information on at least two groups

Includes a comparison group for hypothesis testing

Purpose is to measure an association

Cross-sectional Study

Exposure and outcome measured at same point in time

AKA Prevalence study

Longitudinal Study

Exposure and/or outcome status are measured at different points in time, can document if the disease came after an exposure.

Retrospective Study

Longitudinal study where the exposure and outcome have already occurred.

Prospective Study

Longitudinal study where neither the exposure or outcome have occurred or only the exposure has occurred.

Types of Sampling (3)

1. Exposure-based (Cohort Study)

2. Outcome-based (Case-Control Study)

3. Neither (Cross-Sectional Study)

Ecologic Study

Individual level data is missing; Aggregate data compared (ex. state level data)

Cohort Study

Exposure status is sampled and patients are followed up for outcome

Case-control Study

Outcome status is sampled and patients surveyed (followed up) for exposure

Phases of Drug Development

1. IND - Preclinical data/regulatory application

2. Phase I - Safety, first in humans

3. Phase II - Dose Ranging 2a and Efficacy 2b

4. Phase III - Large Scale Efficacy

5. NDA and Phase IV - Post-market

Examples of Intervention Studies (3)

1. Clinical Trial - treatment/cure focused

2. Field Trial - prevention focused

3. Community Trial - uses groups rather than individual-level data

Control Definition

The standard comparison group(s) to the experimental drug

Types of Controls (3)

1. None - oncology, rare diseases

2. Historical - group from the past, secondary data

3. Randomized

Randomization in Clinical Trials

Ensures that groups will be similar with respect to characteristics that may influence their response to treatment.

The benefit of randomization increases with sample size.

Steps to Set Up a Clinical Trial (6)

1. Determine what intervention is

2. Determine what the control(s) is

3. Select subjects (Inclusion, exclusion criteria)

4. Assign to groups randomly

5. Determine if any groups will be blinded

6. Consider noncompliance, sample size, and generalizability

Defining the Intervention Criteria (3)

1. Strength/Dose Level

2. Duration of Intervention

3. Frequency of Intervention

Possible Clinical Trial Controls (3)

1. Standard Therapy

2. Placebo

3. None at all

Information Collected in Clinical Trials (4)

1. Baseline measurements

2. Follow-up measurements

3. Main outcome

4. Side effects

Types of Randomization (2)

1. Simple - all subjects randomly assigned

2. Stratified - sorted by demographic, then randomly assigned within demographic

Blinding

Blinding means that the subjects (placebo effect) and/or investigators (bias in recording) do not know which treatment arm each subject is in.

Sometimes not possible, sometimes not successful.

“Problem Subject” Categories (4)

1. Non-compliance - In the study, but didn’t get/take their treatment as assigned

2. Crossover - Sought out a different treatment

3. Withdrawal - Quit the study and told you

4. Loss to follow up - Disappeared and you don’t know why

Experimental Study Disadvantages (4)

1. Most costly

2. Most difficult

3. Generalizability can be difficult to determine

4. Many ethical concerns/responsibilities