Model based user interface design for predicting lung cancer treatment outcomes.

We have developed a web-based tool to predict lung cancer patient's survival probability using previously developed survivability prediction software architecture. Four statistical models are included in this version, three for non-small cell lung cancer and one for limited-stage small cell lung cancer. To make the software tool more accessible and convenient for doctors and patients in a clinical setting, user interfaces are developed using a model based approach. Inputs common to prediction models are placed in interface which appears first, model specific inputs later. This design approach reduced both number of entries per interface and average number of interfaces a user needs to navigate.