Table 4.
Sample of probabilities and risk for TB from a prediction model utilizing Computer Aided Detection for TB (CAD4TB) and demographic data from individuals visiting TB diagnostic and treatment centers, in Karachi (Q3 2013–Q2 2015).
| CAD4TB score | Gender | Age | Predicted Probability for MTB Detection* | Risk for TB** |
|---|---|---|---|---|
| 50 | F | 51 | 0.004 | Low |
| 50 | M | 51 | 0.005 | Low |
| 50 | M | 32 | 0.008 | Low |
| 50 | F | 32 | 0.01 | Low |
| 50 | M | 21 | 0.012 | Low |
| 50 | F | 21 | 0.016 | Low |
| 80 | M | 56 | 0.051 | Low |
| 80 | F | 56 | 0.065 | Low |
| 80 | M | 36 | 0.102 | Medium |
| 80 | F | 36 | 0.127 | Medium |
| 80 | M | 22 | 0.16 | Medium |
| 80 | F | 22 | 0.198 | Medium |
| 90 | M | 61 | 0.101 | Medium |
| 90 | F | 61 | 0.127 | Medium |
| 90 | M | 41 | 0.192 | Medium |
| 90 | F | 41 | 0.234 | High |
| 90 | M | 30 | 0.263 | High |
| 90 | F | 30 | 0.316 | High |
| 90 | M | 19 | 0.35 | High |
| 90 | F | 19 | 0.411 | High |
*Predicted probabilities from multiple logistic regression model using CAD4TB and demographic information that is age and gender(Model 2).
**Arbitrary cut-offs for TB risk (Female gender, lower age, and high CAD4TB scores associated with greater risk for TB).