Use of Evidence in Medicine: Cohort Studies

Use of Evidence in Medicine: Cohort Studies

1. Introduction to Cohort Studies

  • Cohort Studies Definition: Arguably the single most important type of research study for understanding disease causation and prevention.

  • Publication Frequency: Every week, results from cohort studies are published in major medical journals such as the New England Journal of Medicine, JAMA, Lancet, and BMJ.

  • Impact: Results significantly influence clinical practice and public health policy, affecting millions of clinical decisions daily.

2. Recap of Types of Observational Studies

  • Observational Studies: The investigator observes, measures, and records data without intervening in usual care. Includes:

    • Ecological Studies

    • Cross-Sectional Studies

    • Case-Control Studies

  • Intervention Studies (Experimental): The investigator intervenes with a pre-planned strategy (e.g., new treatment) and measures the outcome. Includes:

    • Non-randomized Intervention Studies

    • Randomized Controlled Trials (RCTs)

    • Systematic Reviews: Both for observational studies and RCTs.

3. The Basic Question in Analytic Epidemiology

  • Focus: Are exposure and disease linked?

  • Diagram Representation: E (Exposure) leads to D (Disease).

4. Observational Studies and Disease Linkage

  • Purpose: Generally seek to examine the relationship between an 'exposure' and a disease 'outcome.'

  • Examples of Exposures:

    • Risk or protective factors, e.g.,

    • Do omega-3 fatty acids reduce the risk of coronary heart disease (CHD)?

    • Does coffee drinking in pregnancy increase the risk of low birth weight?

    • Is parenting style associated with the risk of eating disorders in children?

5. Types of Exposures

  • Host-Related Factors:

    • Demographics (e.g., age, gender)

    • Biological (e.g., blood pressure, cholesterol)

    • Behavioral (e.g., smoking, physical activity)

    • Psychosocial (e.g., depression)

    • Genetic/Family History

  • Environment-Related Factors:

    • Physical (e.g., season, pollution)

    • Geographic Region (e.g., areas in the UK)

    • Healthcare Access (e.g., procedure underuse)

  • Agent/Disease Factors:

    • Infectious agents

    • One disease causing another

6. Cross-Sectional Studies

  • Design: An observational design that surveys exposures and disease status at a single point in time (a snapshot).

  • Characteristics: Exposure and disease are measured simultaneously, making this type of study exist only at one time point.

7. The Cohort Study Overview

  • Definition: Involves following a defined group of healthy individuals with different exposures over time to analyze who develops a disease.

  • Focus of Research Question:

    • Example: Does a vegetarian diet protect against colorectal cancer?

8. Design of a Cohort Study

  • Participants:

    • Group without the disease (cohort) is analyzed based on exposure.

    • Exposure groups:

      • Exposed group: Develop disease / do not develop disease

      • Unexposed group: Develop disease / do not develop disease

  • Follow-Up: All participants are monitored over time for disease outcome.

9. Example: Adventist Health Study (AHS-2)

  • Participants: 96,354 Seventh-Day Adventist participants recruited between 2002-2007 and followed until 2014.

  • Diet Categories:

    • Vegan: 7.6%

    • Vegetarian: 28.9%

    • Vegetarian plus fish: 9.8%

    • Semi-vegetarian: 5.5%

    • Non-vegetarian: 48.2%

  • Participant Breakdown for Colorectal Cancer:

    • Of 77,659 participants:

    • 40,367 vegetarians: 252 developed colorectal cancer

    • 37,292 non-vegetarians: 238 developed colorectal cancer

10. Features of a Cohort Study

  • Study Type: Prospective or longitudinal study with subjects classified by exposure status.

  • Duration: Follow-up necessary for disease outcome over an extended period.

  • Population: Generally requires a large sample size to ensure enough cases for statistical relevance.

  • Hypotheses: Often investigates multiple hypotheses and various disease outcomes.

11. Strengths of Cohort Studies

  • Disease Causation: They provide stronger evidence for causality than other observational studies (cross-sectional or case-control) because exposure is assessed before disease onset (temporality).

  • Data Capture: Unlike cross-sectional studies that only capture prevalent cases, cohort studies can capture all incident cases, including deaths.

12. Methodological Issues in Cohort Studies

  • Exposure Measurement Validity: Ensure precise methods and minimize bias.

  • Outcome Ascertainment Completeness: Confirm all new disease cases are identified through various means such as registries and repeated assessments.

  • Cohort Representation: Assess whether the selected cohort population is representative of the general population to generalize findings.

13. Measurement Techniques

  • Initial Assessment: Baseline often includes examining participants to measure exposure through various methods. Valid measures can enhance study reliability.

  • Challenges: Some measures are more difficult to obtain validly (e.g., dietary factors), leading to validation studies. Ideal measures should be precise, unbiased, and repeatable.

14. Ascertainment of Disease Outcomes

  • Follow-Up Methods: Identifying new disease cases can involve

    • National registries

    • Contacting GPs or reviewing medical records

    • Repeated questionnaires and assessments during follow-ups

    • Maintaining high follow-up rates to track participants effectively.

15. Cohort Study Population

  • Sample Selection: May be a representative sample or a specific group chosen based on occupation or region. Examples include:

    • Whitehall Study of civil servants

    • Nurses Health Study (USA)

    • British Doctors Study

    • Framingham Study (Massachusetts USA)

  • Generalizability Impact: The type of study population can significantly affect the findings' generalizability.