Interoperable Healthcare Workflows: A Vision for the Industry (Notes From A Keynote)

[This post is part of a series I am writing as a HIMSS17 Social Media Ambassador (four years in a row!) in the run up to HIMSS17, in Orlando, February 19-23. Stop by and meet me at the first ever HIMSS Makerspace, booth 7785 in the Innovation Zone!]

I recently attended the inaugural Healthcare Business Process Management Notation Workshop, in San Diego, California. It is not often one has the luck and foresight to be present at the start of something important. This felt like the start of something important.

What was the subject of the workshop? BPMN stands for Business Process Modeling Notation. It is a set of symbols and conventions for representing workflows and processes. In some cases, it is used not just to represent, in a standard format, workflows and processes, but also as a “program” that can be executed by a workflow engine in a business process management systems. This approach to application development is sometimes referred to as “low-code”, “less-code”, or even “codeless” application development. Analysts, who are not themselves traditional programmers (as in Java or C# programmers) can work together with users and domain experts to draw workflows and create screens to create fully functional software applications.

Diagrams of workflows being executable by workflow engines is still a relatively alien concept to healthcare, though workflow management systems have been prevalent in other industries for decades. I looked around for the best possible quote, describing the relationship between BPMN and workflow engines, and think it is worth including in the preface to my notes from the workflow keynote.

“BPMN 2.0 and Workflow Engines

One of the most common technologies to describe a business process is the ‘Business Process Model and Notation’ – BPMN 2.0 standard. BPMN was initially designed to describe a business process without all the technical details of a software system. A BPMN diagram is easy to understand and a good starting point to talk about a business process with technician as also with management people. Beside the general description of a business process, a BPMN model can also be executed by a process or workflow engine. The Workflow Management System controls each task from the starting point until it is finished. So based on the model description the workflow engine controls the life-cycle of a business process.” (Microservices, Verticals and Business Process Management?)

On to the workshop! It was organized by Ken Rubin (VA, Standards & Interoperability, OMG, Healthcare) and Denis Gagne (Trisotech, OMG BPMN) At one point I raised my hand, just before we joined our breakout groups, and asked Ken the following question:

“Is our aim portable workflows between healthcare organizations, passing data between organizations via workflows, or virtual workflows across organizations?”

To which Ken answered: Yes.

Shane Mcnamee, MD, VA, kicked off the workshop with his keynote: Interoperable Healthcare Workflows: A Vision for the Industry.

It was the best depiction by a clinician of healthcare current workflow state versus ideal future workflow state, intelligible and comprehendible by a workflow management and business process management audience, as I have ever heard. I took detailed notes. Here are his major points.

The following is not word-for-word. To some extent it is a direct transcription from my notes. To some extent I paraphrase Dr. Mcnamee. However I believe I did justice to the substance of his remarks.

  • Dr. Mcnamee explained his military rehabilitation medicine background.
  • As rehabilitation patients flow through the healthcare system, everything should “just happen” for them.
  • When we are sick and vulnerable, we shouldn’t have to captain our own healthcare within and across systems of care
  • It’s difficult for even physicians to navigate through these complex systems of care.
  • Health IT has worked on the interoperability problem for years. We can technically exchange data in some places. We are slowly increasing ability to exchange meaning as well.
  • But what is the purpose of the data flowing back and forth? It is so we can DO things for people. Data has to drive the process of healthcare.
  • When we step back from mere data exchange and semantic interoperability, how is true process interoperability possible?
  • Where we see success today, there are great case managers and nurse always on the phone bothering people, but you also see a lack of IT tools.
  • Dr. Mcnamee introduced a fictional, but representative, case study: globetrotting Eva is pregnant, carries her antepartum record everywhere, and is captain of her own healthcare process.
  • Dr Mcnamee feels as if he is in a dark cave with a tiny flashlight (his computer) and folks are asking for help. He looks around the cave, finds data, synthesizes, and send them off into the deepest darkest part of the cave. And hopes the everything happens as it should.
  • The better clinicians aren’t necessarily the smart clinicians, but those who know how to get things done.
  • They know which phone numbers not even bother to call. As an intern Dr Mcnamee copied all the phone numbers of the lab personnel. Then he called each one until he knew who would answer. Much of healthcare is knowing where the secret doors are.
  • Lack of assurance the right things will happen for individual patients as they flow through healthcare is troubling, to the patient and provider.
  • Every patient has their own, slightly different, healthcare pathway.
  • Compared to a factory model of healthcare, where there are consistent pathways, in healthcare there are as many pathways as there are people.
  • The VA has spent a lot of time and effort over last year and half optimizing BPMN, placing the BPM engine, itself, at the core of the Enterprise Health Management Platform.
  • We seen challenges and gaps with respect to BPMN but hope BPMN can become what we need and adopt to deliver care seamlessly across the country, with “tooling” at both ends, like a railroad has for loading and unloading equipment.
  • We are beginning to develop care pathways in BPMN. If you know anything about healthcare, there are a LOT of different care pathways, at different institutions, across institutions. There is just a ton of care pathway content out there, but can’t be shared and adopted with powerpoint L.ac of electronic tools is problematic.
  • No docs can follow an 300-page care algorithm.
  • Our biggest challenge? How do we figure out who is supposed to do what? How is what they do managed safely? These chunks of knowledge are episodic, through perhaps stack or embedded. They include decision support, medications, lab orders, consultations, finances, anything that is part of care delivery over time for a group of similar patients
  • We use Red Hat’s BPM engine. Here is a BPMN diagram for colorectal screening. Age and risk profiles drive recommendations.
  • There’s lots of discussion about how to deal with BPMN being so deterministic. We need to more flexibly cope with probabilistic behaviors.
  • What should go into a BPMN model, and how should it interoperate with the data systems beneath it, to deliver the right care to the right patient at the right time?
  • The other piece is “Who?” Where do we send tasks, how, how to detect failure to accomplish something by a specific time and then handle escalations…
  • How does the BPMN modeling we are doing integrate into actual workflow, especially in terms of integration with existing EHRs? (in terms of data access, transformation, and write back into these systems). What is the user interface and user experience? Is BPMN driven workflow embedded? Is it off to the side of the EHR? How do we introduce modeled workflows to EHR users and what we can do with these modeled workflow in a way that we don’t do what so much health IT does, which is just make things more complicated.
  • We, patients (and clinicians), want a system that understands us, knows what our needs are, monitors and makes sure that things get done over time, so that when I show up, sick and vulnerable, and the least able to to take charge, I shouldn’t have to be the captain of my of my own healthcare. We believe BPMN can move us in that direction.
  • Eva needs to be able to live her live, and have her healthcare process served up by any healthcare organization across the country.
  • We are right on the cusp of of process integration in healthcare. Major organizations and products are beginning to manage workflows, but they are proprietary.
  • What’s going to happen if we don’t put something down quickly, and move people toward it, is we will have process silos. You may have a fully process-integrated hospital, and another such hospital across the street, but the processes won’t be connected.
  • We need a virtual community of practice in conjunction with various test beds. Cleveland medical institutions are an excellent testbed. The HIMSS Innovation Center is in Cleveland. We need to build infrastructure, so that in order for our vision to actually happen, we need an environment in which it can happen. Test environments and conference rooms are insufficient. We need to study interoperability across actual institutions.

Dr. Mcnamee was the first of excellent eleven speakers. In particular, Denis Gagne gave an excellent overview of BPMN and two related notations: CMMN (Case Management Model and Notation) and DMN (Decision Model Notation). I took lots of notes.

The second half of the day was devoted to breakout section. I took a lot of notes. To me most interesting what discussion of relationship between BPMN and FHIR (???). Are they complementary? BPM engines driven by BPMN models need access to data FHIR exposes. Is there overlapping responsibility? Future drafts of FHIR may include models of task and workflow. Fascinating, especially since FHIR is so famous in health IT, while BPMN is well known outside health IT, but much less so within health IT.

I hope you will want to learn more about BPMN and the workflow tools and Business Process Management systems that use BPMN. And you may wish to attend the next BPMN in Healthcare workshop, this March, in Reston, Virginia.

Here are a couple of related posts I have written.

Relative to the subject of workflow management systems, business management systems, and interoperability in general, check out my twin five-part series.

By the way, the workshop gave away a copy of Business Process Management in Healthcare, Second Edition, for which I wrote the foreword and contributed a chapter. You can find my forward, chapter, and link to the book in ????).

You can find PDFs of the slides from the presentations on the workshop agenda page.

However, here are some of the slides I snapped with my phone.

The final diagram! There, roughly, three levels. At the bottom level are is System A and System B, each containing data and their own internal business logic. At the very top level is a well-defined, but not executable, representation of workflow across System A and System B. In between is executable BPMN. Of course, “executable” means there is something to execute the BPMN, a BPMN compatible workflow engine, implied but not shown.

FHIR was on our mind!

The interoperability stack! Process Semantics Interoperability (and above) corresponds to Task-Workflow Interoperability and Pragmatic Interoperability, in my previous series of articles.

Leave a Reply

Your email address will not be published. Required fields are marked *