Table 3. False positives, false negatives, and total errors for treatment identified with SAS text mining algorithm among stage IV non-small cell lung cancer patients (n = 17,310), 2012–2014, California.
| False Positives | False Negatives | Total Errors | |
|---|---|---|---|
| Treatment Group | n (%) | n (%) | n (%) |
| Platinum doublets | 90 (0.5) | 246 (1.4) | 336 (1.9) |
| Pemetrexed-based regimens | 159 (0.9) | 140 (0.8) | 299 (1.7) |
| Bevacizumab-based regimens | 35 (0.2) | 63 (0.4) | 98 (0.6) |
| Pemetrexed and bevacizumab regimens | 114 (0.7) | 17 (0.1) | 131 (0.8) |
| Single agents | 189 (1.1) | 37 (0.2) | 226 (1.3) |
| Tyrosine kinase inhibitors | 287 (1.7) | 117 (0.7) | 404 (2.3) |
| No systemic treatment | 895 (5.2) | 642 (3.7) | 1537 (8.9) |
| Unknown systemic treatment | 473 (2.7) | 981 (5.7) | 1454 (8.4) |
Percentages (%) represent percent of total