Lecture on Relative Risk and Odds Ratio

Introduction

• Understanding Relative Risk and Odds Ratio
• Used to compare the occurrence of events between two groups.
• Aim to determine the existence and strength of an association.

Key Concepts

• Association: Relationship between exposure and outcome.
• Point Estimate: Actual number representing risk or odds.
• Statistical Significance: Indicated by P value or confidence interval.

Relative Risk (RR)

• Used to compare probability of event occurrence in a study.
• Example: Lung cancer risk for those exposed vs unexposed to secondhand smoke.
• Calculation:
  • RR = (Incidence in exposed group) / (Incidence in unexposed group)
  • Example: Risk in exposed group (0.92) / Risk in unexposed group (0.17) = RR of 5.41.
• Interpretation:
  • RR = 1: No difference between groups.
  • RR > 1: Positive association, increased risk in exposed.
  • RR < 1: Negative association, decreased risk in exposed.
  • Further from 1, stronger the association.
• Significance: Must consider P value or confidence interval.

Odds Ratio (OR)

• Used when RR cannot be calculated (e.g., case-control studies).
• Difference Between Odds and Probability:
  • Odds = Probability / (1 - Probability)
• Calculation:
  • OR = (Odds of exposure in cases) / (Odds of exposure in controls)
  • Example: Odds in cancer group = 43 / 29 = OR of 1.48.
• Interpretation:
  • Same as RR: OR = 1, OR > 1, OR < 1.
• Significance: Same test as RR for statistical significance.

Study Design Considerations

• Cohort Study:
  • Can calculate RR and OR.
  • RR is preferred when possible.
• Case-Control Study:
  • Cannot calculate RR directly, use OR.
• OR and RR are comparable only when the outcome is rare.

Additional Considerations

• OR can overestimate risk for common outcomes.
• Caution needed in interpreting OR in common outcomes.
• OR used in logistic regression, popular in medical research.

Conclusion

• Understanding and correctly interpreting RR and OR is crucial for assessing risk and association in medical studies.