Event Notification Architectures
This research activity focuses on improving the efficiency of the communication between health secondary-care operators, such as physicians and/or pathologists, and the general practitioners. During their clinical story, patients are requested by their general practitioners to make examinations at certain laboratories and/or to receive specialized treatments at given hospitals. Clinical documents produced after patients have received such health secondary-care, such as examination outcomes and treatment summary, have to be promptly delivered to general practitioners so they can take the best medical decision or better follow up the clinical history of their patient.
The current practice is to adopt a paper-based communication: clinical documents are printed and given to patients, which will further bring them to their general practitioners. Such a communication is error-prone and inefficient. Our proposal is to realize an electronic delivery of clinical documents to general practitioner. Specifically, we propose a notification architecture based on publish/subscribe services for a scalable and efficient communication concerning the availability of certain clinical documents and where they can be retrieved.
Publish//subscribe services for communication of document availability
Leveraged on typical APIs provided by such publish/subscribe services, we have built a solution that is:
- organized as a topology of brokers according to a federated architecture so to welcome a large number of interconnecting applications without being affected by severe performance worsening due to its large scale;
- composed of consuming application interested to receive certain kinds of events, which contain the brief description and remote reference of the relative clinical document.
A prototype of such a solution is available here.
The interest of such applications have been formalized in proper subscription schemes, which can be of several kinds:
- Topic-based Subscription: each event to be notified is labeled with a string indicating the kind of documents it is related, e.g., blood test or urinalysis;
- Content-based Subscription: each consuming application will receive events related to documents produced for a given patient, which is referred by means of an anonymous identifier for privacy reasons;
- Hierarchy of Topics: general practitioners are able to model their topics of interest by associating them so to form a diagnostic plan or a so-called clinical guideline. This way, a general practitioner can choose to be notified not when a single document has been produced, but when a proper collection of documents, somehow related, are available.
This last subscription scheme represents a progress with respect to the current state of the art of publish/subscribe services. In addition, we have not used the traditional type-based subscription, since we have not imposed a particular format and/or structure of the exchanged events, so to leave users free to set their own format/structure of reference. Such a solution has been implemented as a collection of web services, which can be used in standalone clients, or even integrated within the context of an existing Health Information System (HIS).
- C. Esposito and M. Ciampi, "On Security in Publish/Subscribe Services: a Survey", IEEE Communications Surveys & Tutorials, vol. 17, no. 2, pp. 966-997, 2015, Impact Factor: 6.806
- C. Esposito, M. Ciampi, and G. De Pietro, "An event-based notification approach for the delivery of patient medical information", Information Systems, vol. 39, 99. 22-24, Elsevier, January 2014, Impact Factor: 1.235
- C. Esposito and M. Ciampi, "A Hierarchical Event-based Architecture for the Notification of Medical Document Availability", in the proc. of IWBBIO 2013 - International Work-Conference on Bioinformatics and Biomedical Engineering, 2013
- C. Esposito, M. Ciampi, G. De Pietro, and P. Donzelli, “Notifying Medical Data in Health Information Systems”, in DEBS 2012: Proceedings of the 6th ACM International Conference on Distributed Event-Based Systems, pp. 373-374, 2012, ACM New York, NY, USA, ISBN: 978-1-4503-1315-5, DOI: 10.1145/2335484.233552