MediScan, a healthcare startup, uses a binary classification model to flag chest X-rays for possible pneumonia so radiologists can prioritize urgent cases. The team is debating whether the current model is acceptable because leadership sees high precision, while clinicians are worried about missed positive cases.
| Metric | Value |
|---|---|
| Precision | 0.91 |
| Recall | 0.68 |
| F1 Score | 0.78 |
| Accuracy | 0.95 |
| AUC-ROC | 0.89 |
| Positive class prevalence | 8.0% |
| Predicted Positive | Predicted Negative | |
|---|---|---|
| Actual Positive | 544 | 256 |
| Actual Negative | 54 | 9,146 |
The VP of Product argues that 91% precision means the model is “very accurate” and should be deployed broadly. The head radiologist argues that recall of 68% is too low because missed pneumonia cases can delay treatment. You need to explain the difference between precision and recall using the model's actual results and recommend which metric should matter more in this use case.