Results of modelling the effects of changing from lower to higher quintiles of referral rates
| Quintile of age-standardised fast-track referral rate | |||||
| First quintile | Second quintile | Third quintile | Fourth quintile | Fifth quintile | |
| Obtained from data | |||||
| Sensitivity* | 42.8% | 46.3% | 48.1% | 48.8% | 50.6% |
| Specificity | 92.0% | 89.3% | 87.7% | 85.9% | 82.4% |
| Application of data to 1000 cancers | |||||
| Cancer+fast-track (true positive) | 428 | 463 | 481 | 488 | 506 |
| Cancer, no fast-track (false negative) | 572 | 537 | 519 | 512 | 494 |
| No cancer+fast-track (false positive) | 2548 | 3434 | 3943 | 4525 | 5645 |
| No cancer, no fast-track (true negative) | 29 452 | 28 566 | 28 057 | 27 475 | 26 355 |
| Total fast-track referrals | 2976 | 3897 | 4424 | 5013 | 6151 |
| Extra cancers via fast-track/extra referrals | |||||
| Moving from this quintile to fifth† | 78/3175 | 43/2254 | 25/1727 | 18/1138 | |
| Moving from this quintile to fourth | 60/2037 | 25/1116 | 7/589 | ||
| Moving from this quintile to third | 53/1448 | 18/527 | |||
| Moving from this quintile to second | 35/921 | ||||
*Sensitivity and specificity obtained with assumed cancer prevalence in symptomatic patients of 3%. A sensitivity analysis of changing this assumption is in online supplementary material 1, table ST2.
†Summing all columns in this row produces the figure of 164 additional cancers/8294 additional fast-track referrals, representing the effect of all practices behaving like those in the top quintile of age-standardised fast-track referral rate.