27
EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data Epidemiologic Study Designs—a review Manya Magnus, Ph.D. Summer 2001

EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

  • Upload
    gary

  • View
    60

  • Download
    0

Embed Size (px)

DESCRIPTION

EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data. Epidemiologic Study Designs—a review Manya Magnus, Ph.D. Summer 2001. Objectives. To review epidemiologic study designs To review measures of association used in epidemiologic studies - PowerPoint PPT Presentation

Citation preview

Page 1: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

EPID 623-88 Introduction to Analysis and

Interpretation of HIV/STD Data

Epidemiologic Study Designs—a review

Manya Magnus, Ph.D.Summer 2001

Page 2: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Objectives

• To review epidemiologic study designs• To review measures of association

used in epidemiologic studies• To discuss interpretation of study

results

Page 3: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Basic study designsAnalytic

Descriptive Observational Intervention

Case reports/series Cross-sectional

Randomized controlled trials

Birth cohorts

Cohort (concurrent, non-

concurrent)

Community-based

interventions

Ecologic/correlational

Case-control (and variations)

Page 4: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Case reports/case series (1)

• Description of unique, unusual, rare events in one or several individuals

• Often used by physicians• May stimulate awareness of problem

(note: PCP, DES/vaginal clear cell adenocarcinoma), hypothesis generating

Page 5: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Case reports/case series (2)

• Analysis of case series:– Might focus on incidence or prevalence in

general (or target) population– Raw data, small numbers– If more than a few, proportions (5/10=50%)– Narrative as essence of report, plus

communication of context, why interesting

Page 6: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Birth cohorts

• Allows exploration of trends based on age or period effects

• Allows exploration of cumulative effects of exposure, latency periods, timing of exposure

• Explore cohort effects, age effects, period effects

• Interaction between time and calendar age• Looking graphically at cross-sectional data by

age group

Page 7: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Ecologic studies

• Aggregate measures• Environmental measures• Global measures• Units of observation generally geographic region,

area• Looking at number of outcome events and

predictors of interest on aggregate level• “Ecologic fallacy” issues• Evaluation of association: plotting, comparison of

rates, prevalence, adjusted-measures, etc.

Page 8: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Cross-sectional studies

• Individual-level data• “Snapshot” of exposure and outcome• No temporality provided• No causation can be inferred• Useful in hypothesis generating• Analysis: frequencies, cross-tabs, point

prevalence rate ratio, odds ratio, comparison of rates, etc.

Page 9: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Cohort studies

• Concurrent or non-concurrent• “gold standard” for observational studies• Basis for other designs• Disease-free at baseline• Follow for outcome• Analysis: frequencies, cross-tabs, relative

risk, hazard ratio, attributable risk, etc.

Page 10: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Case-control studies

• Look at those with/without outcome of interest and evaluate exposures

• Good for rare diseases• Many variations (nested, case-cohort,

etc.)• Analysis: frequencies, cross-tabs, odds

ratio, attributable risk, etc.

Page 11: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Randomized controlled clinical trials

• Cohort studies based on model of RCT• Randomize subjects to receive

intervention and follow for outcome(s) of interest

• Intent-to-treat analysis• Analysis: frequencies, cross-tabs,

relative risk, hazard ratio, etc.

Page 12: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Community-based interventions

• Similar to RCT, but randomizing communities

• Unit of analysiscommunity (note power issues)

Page 13: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Reminders (1)Diseased Not

diseased

Exposed A B A+B

Not exposed

C D C+D

A+C B+D A+B+C+D

Page 14: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Reminders (2)

Relative risk (RR) formula

[A/(A+B)]/[C/(C+D)]

Page 15: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Reminders (3)

Odds ratio (OR) formula

AD/BC

Page 16: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Interpreting published results

Page 17: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Steps to understanding published tables

1. What is the study design?

Page 18: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Steps to understanding published tables

1. What is the study design?

2. What is unit of analysis?

Page 19: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Steps to understanding published tables

1. What is the study design?2. What is unit of analysis?

3. What are predictors of interest?

Page 20: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Steps to understanding published tables

1. What is study design?2. What is unit of analysis?3. What are predictors of interest?

4. What are outcomes of interest?

Page 21: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Steps to understanding published tables

1. What is study design?2. What is unit of analysis?3. What are predictors of interest?4. What are outcomes of interest?

5. In tables, what is n?

Page 22: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Steps to understanding published tables

1. What is study design?2. What is unit of analysis?3. What are predictors of interest?4. What are outcomes of interest?5. In tables, what is n?

6. Is the table referring to subset or whole sample?

Page 23: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Steps to understanding published tables

1. What is study design?2. What is unit of analysis?3. What are predictors of interest?4. What are outcomes of interest?5. In tables, what is n?6. Is the table referring to subset or whole sample?

7. What is being presented, compared?

Page 24: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Steps to understanding published tables

1. What is study design?2. What is unit of analysis?3. What are predictors of interest?4. What are outcomes of interest?5. In tables, what is n?6. Is the table referring to subset or whole sample?7. What is being presented, compared?

8. What are the denominators? Do they differ by column?

Page 25: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Steps to understanding published tables

1. What is study design?2. What is unit of analysis?3. What are predictors of interest?4. What are outcomes of interest?5. In tables, what is n?6. Is the table referring to subset or whole sample?7. What is being presented, compared?8. What are the denominators? Do they differ by column?

9. Can you add up data from text?

Page 26: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Steps to understanding published tables

1. What is study design?2. What is unit of analysis?3. What are predictors of interest?4. What are outcomes of interest?5. In tables, what is n?6. Is the table referring to subset or whole sample?7. What is being presented, compared?8. What are the denominators? Do they differ by column?

9. Can you add up data from text? 10. Can you calculate measures of association?

Do they agree with the authors’?

Page 27: EPID 623-88 Introduction to Analysis and Interpretation of HIV/STD Data

Steps to understanding published tables

1. What is study design?2. What is unit of analysis?3. What are predictors of interest?4. What are outcomes of interest?5. In tables, what is n?6. Is the table referring to subset or whole sample?7. What is being presented, compared?8. What are the denominators? Do they differ by column?9. Can you add up data from text? 10. Can you calculate measures of association? Do they agree

with the authors’?

11. Are the authors’ conclusions correct?