Models can be run directly on the native model software, or thru aggregating software that allows data input once and then runs multiple models. Aggregator software includes CancerGene, Hughes RiskApps Express, and Hughes RiskApps (Installed version). In comparing results between these approaches discrepancies can occur for 2 reasons:
- User issues: Did you enter EXACTLY the same factors into all programs? One discrepancy last week was because a different patient age was used in each program.
- Differences in conventions used: When data in entered, conventions are followed in terms of how to deal with the entered data. A different convention in the aggregator will cause a small difference from the original model software. For example, if you enter age 55, risk will be calculated from age 55. But if age is determined by birthdate and current date, the risk might be calculated from age 55.75. In Tyrer Cuzick, you can choose benign or unknown for the biopsy result. If you choose unknown, the risk will be higher than for benign.