SOA as a basic infrastructure for Pervasive Computing in Healthcare – Part 2
In my SOA as a basic infrastructure for Pervasive Computing in Healthcare – Part 1 I claimed that hospitals will need to devise an event-driven SOA as a basic infrastructure upon which the clinical and administrative processes can be designed, developed, deployed and in some cases automated. A high level of pervasiveness will help to reduce the documentation effort hence reduce process costs.
During a Swiss Health conference last week, Dr. med Ingrid Seyfahrt-Metzger claimed that, due to general market conditions german hospitals are forced to focus on process excellence. As 82% of all direct-to-patient costs are labour costs. Due to demographic changes, reduced prices and new labor laws they expect those costs to rise withan additional 40 Mio Euros. Process excellence and process automatization is seen as the most important instrument to be able to cope with those new challenges. Other presenters showed a direct linkage of process management and service quality.
At IDC’s SOA Conference in London earlier this month, Ruediger Spies, VP enterprise applications at the analyst house, said: “SOA makes it possible to automate a much higher degree of business processes.”
Together with the market situation hospitals are in, this means that SOA and process automation become tactical and incremental. The above mentioned market pressure will force SOA and BPM adoption as it leads to dramatic efficiency improvements while cutting labour costs and improves quality. Targeting process improvement holds the greatest hope for better care and lower costs.Who said that SOA is dead?
Business processes and clinical processes
In hospitals we talk about clinical & administrative processes. Clinical processes may be much more complex anddefine the core processes of the hospital, but both types of processes are transformations in which inputs are converted to outputs. They both operate in the context of constraints, regulations, and defined roles and relationships. Where administratvie processes emerge from experiences, most clinical processes are based on evidence based research materials.
Mathias Kirchmer compared in his book, “High performance Through Process Excellence“, process models and the way the represent reality to pop art paintings, which represent and capture reality in a different way. A (business) process model is a model of one or more (business) processes, and defines the ways in which operations are carried out to accomplish the intended objectives of an organization. such a model remains an abstractionand depends on the intended use of the model. It can describe the workflow or the integration between (business) processes and can be constructed in different levels.
Visual process models are descriptions of steps that take place during a process and frefrequently represented in flowcharts such as those documented in the web-based BPMN editor Oryx, developed by the university of potsdam. By offering a webbased tool, and the possibility to search existing models, they focus an area in which I see much potential. An open BPMN based wiki-process-pedia. Wouldn’t this be a great if we had this available for the BPX community?
After this very brief introduction on (B)PM I would like to go back on my previous claim and adress the case of process automation by using an event-driven SOA and pervasive healthcare technology.
Pervasive Bedmanagement – Integrating an Event-Driven SOA with smart devices
With an decreasing average length of stay (ALOS) hospitals have to optimize their occupancy rate, hence the management of hospital beds has become a core activity which involves both medical and administratvie personell. During the planing phase of bed assigments, it is important to ensure that patients are admitted appropriate to their clinical needs. In addition, while aiming for a highest possible utilization, capacities need to be assigned such that planned admissions don’t need to be cancelled due to capacity shortages and sufficient emergency admissions can be handled.
During medical treatment, tasks of bed management mainly elate to transfer of patients, e.g. between wards due to changed edical conditions, and to short-term movements of patients, eg. for specific therapy, surgery or diagnostics (x-ray, etc.).
According to a study conducted by N.A.D.Connell, T.P. Young,, approximately seven per cent of inpatients admitted to German hospitals in 2006 were transferred to a ifferent ward. For a large Hospital this means that up to 50.000 beds/patients are being transfered. According to some of my clients the actual documentation time/effort takes in average 5′ for decentral documentation, and up the 10′ for organisations that use decentral paper-based followed by central computer-based documentation. This means that the annual documentation effort takes 1000 mandays (calculated on 8 hours / day).
If staff and rooms (and potentially patients) were equipped with RFID and respective readers, the effort for dosumenting the transfer processes could be reduced considerably. If we could automate this process (as described below) we could save 900 mandays annually.
As shown in the BPMN diagram above, the events triggered by the sensor network can be integrated with respective components of the hospital application landscape. Data entry required by the medical staff could then be provided through mobile devices. Thereby the effort for non-value-adding tasks related to such a transfer could be considerably reduced. For example, the data to be provided by a nurse could be minimized to distinguishing between shortterm moves and transfers to another ward and the identification of the destination department.
Setting this process example into the context of the three features of pervasiveness – coverage, seamless integration and process automation – a number of improvements in convenience of care and performance of back-end processes an be realized based on an event-driven SOA environment. In our example we used the update encounter enterprise service that has been delivered with EHP4 for SAP Patient Management.
With regards to coverage, compared to classic bed management processes which often require a high volume of personal interaction, the integration of sensor technology for asset and patient management enables real-time access to resource availability and hence more efficient bed allocation. In addition an integrated system landscape can react to the events triggered by a transfer process for example by enabling access to medical data based on the patient’s current location/ward assignment, also improving privacy. Automation is enabled on the one hand through workflow support, e.g. through triggering events that require human intervention, hence reducing process idle times. On the other hand particularly back-end process can be fully automated, e.g. by using event-driven documentation of activities. Integrating real-world with system events and allowing the hospital staff to manipulate and access information through situation specific devices and user interfaces, efforts can be reduced through better integration of the IS/IT with their tasks.
credits : I want to thank René Fiterrer from SAP Research Switzerland, helping me building this use case, for his great support!