Appraising etiologic & prognostic study

Understanding how to critically appraise research is crucial for interpreting scientific literature.

In this session, we focus on two fundamental types of study designs: etiologic and prognostic studies. These study types serve different purposes and require distinct considerations.

Study Design Considerations

The cohort study is the most suitable design for both etiologic and prognostic research.

• Etiologic research seeks to establish a causal relationship between a risk factor and a disease.

• Prognostic research aims to estimate risk and predict outcomes over time, requiring longitudinal data.

• When a cohort study is not feasible for etiologic research due to ethical concerns, resource constraints, or long latency periods, a case-control study may be acceptable.

• Prognostic research, however, always requires a cohort study as it involves tracking individuals over time.

Key Differences Between Etiologic and Prognostic Studies

Example: Cannabis Use and Schizophrenia

Consider an 18-year-old named Mr. C, a heavy cannabis smoker diagnosed with schizophrenia. Two research questions can be framed:

• Etiologic Question: Does smoking cannabis increase the risk of schizophrenia? (A cohort study would be ideal, adjusting for confounders like age, gender, and mental health history.)

• Prognostic Question: What is the probability of a cannabis smoker developing schizophrenia? (A cohort study would assess absolute risk, incorporating patient characteristics.)

Data Collection and Analysis

• Etiologic research requires precise data collection for all determinants.

• Prognostic research should mirror real-world clinical settings to avoid overestimating predictive value.

Both rely on multivariate analysis:

• Etiologic studies control for confounders to infer causality.

• Prognostic studies assess model accuracy in predicting outcomes.

Tools for Critical Appraisal

To appraise etiologic and prognostic studies, you can use:

• Critical Appraisal Checklists: CASP, JBI, EBM, and others.
• RAMMbO Framework:
  • Recruitment: Were participants representative of the population?
  • Allocation: Was selection unbiased?
  • Maintenance: Were participants followed up appropriately?
  • Measurement of Determinants and Outcomes:
    • Determinants should be measured accurately in clinical settings.
    • Outcomes must be measured with high precision.

Understanding Results

• Prognostic studies often use:
  • Scoring systems
  • Prediction models
  • Survival curves
• Receiver Operating Characteristic (ROC) Curve: Helps differentiate between high and low-risk patients.
• Area Under the Curve (AUC): Indicates test accuracy:
  • Higher AUC = Better discrimination ability.
  • AUC > 0.9: High accuracy.
  • AUC 0.7–0.9: Moderate accuracy.
  • AUC 0.5–0.7: Low accuracy.

Critically appraising etiologic and prognostic studies is essential for making informed clinical decisions. Understanding study design, data collection, and result interpretation ensures the reliability of evidence used in medical practice.