How to Quantify Risk

Introduction

• Risk usually connotes a negative implication.
• High risk = potential for bad outcome; Low risk = less likely bad outcome.
• High risk may be good if the outcome is positive.
• In this context, risk indicates chance of developing a disease.
• Uses a similar 2x2 table as for evaluating diagnostic tests.

Types of Studies

• Focuses on exposure/intervention rather than tests.
• Example: Two groups—one exposed and one not exposed—to compare the risk of disease.
• Exposure can be beneficial or harmful.

Quantifying Risk: Odds Ratio (OR) vs. Relative Risk (RR)

• Both describe the same story from different perspectives.
• OR formula: (A/B) / (C/D)
• RR formula: (A/(A+B)) / (C/(C+D))
• Odds Ratio: Uses ratios.
• Relative Risk: Uses proportions.

Study Designs

• Cohort Studies: Recommended for RR because they depict natural prevalence.
• Case-Control Studies: Recommended for OR due to biased sample selection.

Mnemonics

• OR: Used in Case-Control Studies - Different letters (C and O in Case-Control).
• RR: Used in Cohort Studies - Same letters (R and R in Relative Risk).

Low Prevalence Scenarios

• OR approximates RR when disease prevalence is already low.

Relative Risk (RR) Interpretation

• RR > 1: Exposure increases disease occurrence.
• RR < 1: Exposure decreases disease occurrence.
• RR = 1: No effect on disease occurrence.

Additional Metrics

Relative Risk Reduction (RRR)

• Formula: 1 - RR
• Indicates percentage reduction in risk with an intervention.

Absolute Risk Reduction (ARR)

• Formula: Proportion of unexposed developing disease - Proportion of exposed developing disease.
• Example: 8% (unexposed) - 2% (exposed) = 6% ARR.

Attributable Risk (AR)

• Formula: Proportion exposed developing disease - Proportion unexposed developing disease.
• Example: 21% (exposed) - 1% (unexposed) = 20% AR.

Number Needed to Treat (NNT)

• Formula: 1/ARR
• Example: ARR = 10%, then NNT = 1 / 0.10 = 10.
• Interpretation: Treat 10 people to benefit 1 person.

Number Needed to Harm (NNH)

• Formula: 1/AR
• Similar logic as NNT, used when exposure increases disease risk.

Case Fatality Rate (CFR)

• Formula: (Number of deaths / Number of cases) * 100
• Example: 4 deaths out of 10 cases = 40% CFR.

Conclusion

• Understanding these metrics is crucial for quantifying risk in studies.
• Recommended to review related videos for better understanding.