Guide2Treat: Software

Evidence-based guidelines have been proved to exhibit a positive impact on the quality of the treatment process. However, their influence on patient care remains limited. For the success and effectiveness of clinical guidelines, a fast and clear provision of knowledge at the point of care for all treating physicians is essential. A uniform formal representation is developed so that the knowledge extracted from the patient data is consistently presented at the human-machine interface. The gold standard for evaluation is determined using a human abstraction of 86 rectal cancer and 89 colon cancer patients. In order to represent the position of a patient on a clinical algorithm, the evidence-based knowledge from the guidelines must be linked to the patient-related data and visualized at the point of care. The goal is to automatically position a patient on a clinical path in order to derive a guideline-based therapy decision. Using this approach for therapy recommendation, the algorithm achieves a precision value of 90.56% for colon cancer and 82.76% for rectal cancer, with recall values of 91.11% and 86.58%. The identification of guideline-based recommendations from clinical notes has the potential to be used as a clinical decision support tool.