Bayesian networks are one of the most intuitive statistical models for both estimation, classification and prediction of patients outcomes. However, the availability of inference software in clinical settings is still limited. This work presents preliminary steps towards the creation of simple web-based forms that can access a powerful Bayesian network inference engine, making the derived models usable at bedside by both the clinicians and the patients themselves.

Integration issues are still one of the main concerns regarding hospital information systems (HIS). Hospitals need integration engines capable of guaranteeing that messages are properly dealt with, use standards properly (eg. IHE), are secure, promote integration of legacy systems, enable auditing, monitor failures, evolve to include new standards, have balancing capabilities and are able to compensate problems with the source-projects. This paper describes an architecture for an integration solution for HIS based on open-source components. The proposed architecture is called HealthIRA and is based on a HL7 message router(Mirth), a load balancer (HAProxy), audit-trail (Elasticsearch), HL7 sniffer (GO), message cache, clinical document online translator and databases (MySQL). HealthIRA is partially (all modules but audit-trail, cache and translate) in use at an oncology hospital. Initial results include the analysis of the gathered data and a comparison between the message workflow and the provider's' documentation which revealed important incongruities. With all modules installed, HealthIRA acts as an integrator, a standard's auditor, a translator and a secondary source of data when the primary source is not available, solving some of the integration problems. We identified still the need for an open-source orchestration module with an user-friendly interface for each service.