M09.01.09 Double Hump Graph

In epidemiology and biostatistics, understanding cutoff points on diagnostic screening tests helps differentiate between healthy and diseased populations based on sensitivity, specificity, and accuracy. A double hump graph illustrates these cutoff points and aids in identifying the optimal values for screening parameters. Below is a detailed look at how different cutoff points affect sensitivity, specificity, and predictive values.

Graph Interpretation

Healthy and Diseased Populations Along a Screening Dimension

Cutoff Points (A – E):

• Point B: Provides the highest sensitivity, correctly identifying all sick patients.
• Point D: Most specifically, identifying only the diseased patients, minimizing false positives.
• Point C: Optimal accuracy, where sensitivity and specificity intersect, providing a balance.

Key Metrics and Concepts

Points to Remember

• Optimal Sensitivity: Found at point B, identifying all sick patients.
• Optimal Specificity: Found at point D, with minimal false positives.
• Accuracy: Highest at point C, where sensitivity equals specificity.
• Predictive Values: Negative predictive value peaks at the point of optimal sensitivity, and positive predictive value peaks at the point of optimal specificity.

Receiver Operating Characteristic (ROC) Curves

A ROC curve graphically represents the trade-off between sensitivity and specificity for different test thresholds.

Figure 2: ROC Curve Analysis

• Curve E achieves the highest sensitivity with minimal false positives, indicating it as the most effective cutoff for screening.

Bibliography

• CDC Screening and Diagnostic Testing Guidelines
• NIH: Principles of Screening Tests
• Medical Biostatistics: Concepts and Practices