Supporting breast cancer decisions using formalized guidelines and experts decision patterns: initial prototype and evaluation

Abstract

Transparent decisions and its documentation of breast cancer patients' therapy are getting more important especially since modern therapeutic approaches favor personalized forms of treatment. The medical decisions for a treatment are very complex, because there are rules and different options for each patient. To support the decision process, we analyzed the current decision rules and implemented them in a prototype of a rule-based expert system. Thus, this system shall support the quality assurance regarding transparent documentation of individualized therapeutic decisions. For evaluating the system, we used data from a state tumor center and compared the decisions suggested by our system with expert ones. The system and the expert approach will be compared with each other as well as the differences in the treatment decisions. The first preliminary results show us that the human factor-like must be considered by creating a decision support system. The prototype delivers first results, which are restricted, but the results are promising for further developments.

Towards Distributed Healthcare Systems – Virtual Data Pooling Between Cancer Registries as Backbone of Care and Research

“…The TI is the main infrastructure of the German national e-Health agenda. 1 It provides access for patients to their Personal Health Records (PHRs) and ways for physicians to share and exchange patient data in a secure way. It is also the technological foundation for many planned national e-Health projects, such as, research databases and electronic prescriptions.…”

“…Hence, initiatives have been emerging notably in disciplines that highly depend on large amounts of data such as oncology: e.g., a federated search over several university medical centers' data stocks by German Cancer Consortium (DKTK) [26]. Anyways, there are many potential oncological information hubs, e.g., alone for breast cancer in German-language Europe 280 centers exist [1], thus resulting in an almost unmanageable but still worthwhile number of data sources -and the need for more centralized way of data usage. In the information flow of a desirable "learning healthcare system" [37, cf.…”

“…A large body of literature gives evidence to the improved diagnostic accuracy, adherence to clinical practice guidelines, and some clinical outcomes [15]; [25]. All over Germany, medical centers have established tumor boards in their clinical environment [1] not least due to German Cancer Society's initiatives promoting certification of centers for organ cancer or oncology.…”

Guest editorial: special issue on “Artificial Intelligence in Health and Medicine”

“…In this article, "Supporting Breast Cancer Decisions Using Formalized Guidelines and Experts Decision Patterns-Initial Prototype & Evaluation" the authors [12] introduces an approach to support the Decision system in medicine which can assist the experts/patients to have quality assurance regarding transparent documentation of individualized therapeutic decisions. They compare their automatically generated decision system with manually expert decisions with recommendations and their results are promising.…”