Earlier this week I attended the two-day Long-Term and Post-Acute Care Health IT Summit in Baltimore (archived pdf).
A few years ago, when I worked for an EHR vendor, I investigated the state of health IT in long-term care, but didn’t find much, at all. There was some software to generate government-mandated data sets, and that was about it. So I was glad to be surprised at how much interesting health IT and workflow IT related discussion I found at #LTPAC13 (that’s the Twitter hashtag). In fact, I heard “workflow” everywhere and frequently, from the podium, in the hallways, and at lunches and coffee breaks. If I had the combined transcripts of all the attendees and counted the times “workflow” was uttered, it would have been in the thousands.
Words and Phrases Associated with “Workflow” at #LTPAC13
In fact, there was more talk of healthcare workflow at #LTPAC13 than at some other health IT conferences I’ve been to recently. I saw lots of healthcare workflow-relevant presentations about care coordination, clinical decision support, telehealth and population health management. This health IT summit, about long-term and post-acute care, was chockfull of imaginative ideas and insights about real-world healthcare workflow.
Rather than try to summarize and synthesize a veritable word cloud in which “workflow” appears over and over, I’ll focus on two white papers. They synthesize in a way I cannot; they are written by experts on long-term and post-acute care health IT and they emphasize workflow.
The first white paper is from LeadingAgeCAST, Center for Aging Services Technologies. It has an excellent overview of the necessary steps to document current workflows and envision how HIT can help, and avoid hurting, the work of clinicians and support personnel. I’m an industrial engineer (or at least have an MSIE) and I entirely understand ends that drive the above suggestions and the means by which they are achieved.
But I will have one observation and one question.
EHR FOR LTPAC: A Primer on Planning and Vendor Selection
[begin quote]
3.2.3 Workflow and Process Redesign
3.2.3.1 Importance of Workflow and Process Improvement
Change due to HIT needs to be managed not only to help individuals overcome their concerns and adopt the technology well, but also to ensure that the change brought about by the technology is the right change for the organization.
- Workflow and process changes must be understood and managed. Workflow and process changes should be made to take advantage of HIT, but must also point out the control points that must be present or should be added to the systems for optimal results.
- Professional judgment must be applied when using any tool, including all forms of HIT. Computers can crunch data tirelessly and remind humans who may be forgetful in the rush of their everyday lives. However, they are not substitutes—no matter how well programmed— for the experience that trained workers bring to health care. Workflow and process changes should aid professionals, but are not substitutes for them.
3.2.3.2 Process Mapping as a Means to Manage Change
Process improvement experts suggest that the best way to help potential new users of information systems overcome resistance to the change is to demonstrate the quality efficiencies and patient safety of computer systems and to engage users in making their own changes.
3.2.3.3 Workflow and Process Mapping
Process mapping is a fairly well-defined science, with a number of tools and techniques that can be deployed to understand current workflows and processes as well as identify opportunities for improvement.
3.2.3.4 Mapping Current Workflow and Processes
The following steps should be used to map current workflow and processes:
- Identify processes to be mapped, those that will be impacted by the HIT being acquired.
- Use individuals who actually perform the process; they know it best and they need to own the impending change.
- Instruct persons on process mapping, why it is being done and how it is done.
- Map current processes. Avoid identifying opportunities for improvement now, because critical controls built into current processes may be overlooked.
- Validate maps to ensure they reflect current processes, all variations, and all the information needed. This is the time to be candid about how processes and workflow really occurs in the current state, including workarounds, so that the new HIT-supported processes can be developed in ways that are most productive and helpful.
- Collect all forms and reports that are part of processes to be automated through HIT.
Obtain benchmark data to define expectations for change and for use in benefits realization studies.
3.2.3.5 Workflow and Process Redesign
Map how the workflows and processes will be performed with HIT. Identify potential problems in current workflows and processes and determine their root cause. Things to look for include bottlenecks, sources of delay, rework due to errors, role ambiguity, unnecessary duplications, unnecessary steps, long cycle time, lack of adherence to standards, lack of information and lack of quality controls.
[end of quote]
The reason there is so much emphasis on getting workflow right before implementing EHRs and health IT systems is that it is so doggone hard to change the workflow after implementing these systems. My question is this, if workflow is so hard to change, and this requires so much time and energy and cost getting it right, and most of the time organizations get it wrong, and then are stuck with relatively unchangeable and unimprovable workflows, why not use a different technology, one in which changing and improving workflow is easy, not hard, to do?
In fact, just think how much easier to do all the above, supremely rational, steps to improve workflow (what I call meta-workflow), if only the EHR or health IT system used workflow technology. Workflows could be not just diagramed, but executed too (What’s So Special about EHR Workflow Management Systems?). Data to improve workflow would not have to be collected or made up, but come directly from the workflow log (Clinical & Business Intelligence, Meet Process Mining).
The second to last session was called Talking EHR with the Nurse Executive Council and the CIO Consortium. The presentation, itself, was short. Most of it was given over to taking questions and comments from the audience. And guess what? Most of the questions revolved around workflow. The vendors don’t understand my workflow (often true). EHRs often disrupt my workflow (unfortunately true). Are there alternatives to EHRs? (I’d argue that EHR workflow systems are the best alternative, at least with respect to workflow issues with current workflow-oblivious EHRs.)
I got up, three times, and made the following comments. (Yes, word for word, this is what I said.)
Comment 1:
Regarding workflow, there are such things as workflow systems, workflow management systems, business process management systems and dynamic and adaptive case management systems. They’re very popular in other industries. So, the question I have, is, if workflow is such a big problem, and I hear it mentioned frequently in presentations and in the hallway, why not use workflow technology to help solve it?
Comment 2:
The basic problem is that electronic health records, as they are now constructed, is they are not malleable. You can’t change them easily after you implement them. You have to go back to the C# or Java programmer back at the vendor. They are not instrumented, so you have analytics so you can improve cycle time, throughput, and consistency. And they do not have a means of systematically improving those statistics. All those things that I just described are classic attributes of software called workflow management systems, business process management suites and dynamic or adaptive case management systems. These systems are based on structured workflows that can be easily changed and easily improved. Current EHRs are based on structured documents. They force the user to be the workflow engine because there is no workflow engine in the EHR. If you want a workflow engine in your EHR you need to use a workflow management system. There are two ways of doing that. You can reengineer and retrofit that EHR or you can build it on top of workflow management system, with good workflows, workflows you can change and adapt to users’ needs and preferences.
Comment 3:
I read the white paper. It’s fantastic. From my perspective, coming from workflow management and business process management in healthcare, you’ve nailed it.
[end of my comments from the audience]
Yes, Electronic Health Record (EHR) Solutions LTPAC Providers Need Today (the second whitepaper I referred to earlier) did indeed nail it.
First of all, the word “workflow” appears 36 times in a 27-page paper (with lots of diagrams and white space!).
Here’s the concordance plot. You see that ‘workflow’ is mentioned mostly toward the end, in the concluding remarks.
N. of Hits = 36, File Length (in chars) = 65109
Let’s pull out all of the instances of ‘workflow’, sorted alphabetically on the next word after ‘workflow’. Take a look at the phrase “workflow platform”, which occurs seven times. You’ll need to scroll a ways. Be sure to checkout some of the other interesting contexts in which “workflow” occurs.
| real-time participation in care processes leveraging configurable | workflow | and checklists. |
| Sections above address the | workflow | and interactive decision support capabilities needed |
| secure cross provider | workflow | and passing information through standards based document |
| clinical environment requires rapid innovation in design of care | workflow | and processes within and across changing care settings. |
| the existing EMR modular configuration causes disruption in | workflow | and results in development of “workarounds”, thus negatively |
| In our post-Facebook world, the benefits of socialization of work and | workflow | are becoming better understood. The ability to “share”, “like” |
| It requires a platform approach that ensures that all system | workflow | can be consistently developed, implemented, managed, and imp |
| variability in function specifics, process, protocol and | workflow | . |
| While | workflow | guide task- and process-level actions and checklists organiz |
| Rich extensible | workflow | , integration and coordination |
| It also delivers a CRM aware platform with database, |
workflow | , interface, portal, social, and mobility building blocks |
| Clinicians strive to “fit” their care delivery process or | workflow | into the EMR modules. |
| The care delivery process embedded in clinical | workflow | is often characterized in terms of a pattern of actions |
| enforce the protection of extracted information, or include the | workflow | necessary to make it easy for the user to add the security |
| source and pre-analyze the “big data” needed to “prompt” |
workflow | . |
| pply that expertise to the workflow “prompt” step in order to enhance | workflow | outcomes will be a key differentiator of effective EHR system |
| Socialization adds an accelerator dimension to all aspects of the | workflow |
platform described above. |
| The same extensibility platform, our | workflow | platform, if you will, informs multiple dimensions of extensi |
| community of providers. In fact, the value enabled by the |
workflow | platform is primarily found in the network effects |
| exchange, or a cross-entity coordination all reflect the same basic | workflow | platform requirements. A core patient-centered EHR and secure |
| securing compliance for complex, high impact processes. The |
workflow | platform should be able to richly leverage the checklist |
| An effective | workflow | platform should be able to accommodate development and sharing |
| capturing of action/decision may be pertinent. The effective | workflow |
platform will provide rich support for developing unit proces |
| The “prompt” aspect of the unit | workflow | process brings evidence to bear on system outcomes. |
| Protocol-Driven and Contract-Based Workflow The system |
workflow | process should be driven from a defined and customizable set |
| were developed, little attention was paid to supporting the clinical |
workflow | process that guides professional clinical care within and |
| These systems were not designed to support the clinical |
workflow | process that guides professional clinical care. |
| new ways of visualizing information is a key factor in assuring that | workflow | processes can be defined in ways that are highly intuitive |
| ernal and external sources (big data) and apply that expertise to the |
workflow | “prompt” step in order to enhance workflow outcomes |
Management
|
|
|
| accelerate time to market, while achieving the efficiency and |
workflow | requirements outlined in the paper. |
| A platform approach can provide support for configurable | workflow | , rich data analytics and visualization, checklists and proto |
| Historically, | workflow | systems viewed roles statically, with processes flowing in |
| Protocol-Driven and Contract-Based | Workflow | The system workflow process should be driven from a defined |
| Clinicians rely on clinical care delivery processes ( | workflow | ) to provide care (reference Appendix 2.0). |
| The | workflow | unit process described above can be oversimplified as trigger |
Words and Phrases Associated with “Workflow” at #LTPAC13
Yes, there’s “platform”. Maybe “process” and “processes” should be combined to rival “workflow.” Or, since process and workflow can be synonymous, maybe I should have deleted those too. But look at the rest of the words.
- Configurable
- Prompt (as in to prompt a user)
- Checklists (a hot topic)
- Coordination
- Protocol-driven
- Outcomes
- Expertise
So, clearly, workflow is at the center of a number of extremely important health IT topics that are on lots of minds these days. And there it is: “platform”. As in “workflow platform.” Seven times!
I have, in fact, written a lot about workflow platforms. They include workflow management systems, business process management suites, and adaptive case management systems. For example, this is what I wrote about “platform” in an EHR workflow context:
An EHR workflow management system is the workflow management system used to generate a specialty-specific EMR workflow system. It is a platform for launching and managing EMR components just as an operating system is a platform for launching and managing user applications and software services. Both are “an agreement that the platform provider [gives] to the software developer that logic code will interpret consistently.”
(http://en.wikipedia.org/wiki/Computing_platform)EMR workflow systems are extensible by virtue of being generated by EHR workflow management systems. Extensibility “means the system is designed to include hooks and mechanisms for expanding/enhancing the system with new capabilities without having to make major changes to the system infrastructure.”
(http://en.wikipedia.org/wiki/Extensibility)An EMR “software extension is a computer program designed to be incorporated into another piece of software in order to enhance, or extend, the functionalities of the latter. On its own, the program is not useful or functional.”
(http://en.wikipedia.org/wiki/Computing_platform)
Workflow technology and extensible platforms are not only complementary, it is the workflow platform that makes extensions possible. Much of the rest of Electronic Health Record (EHR) Solutions LTPAC Providers Need Today also describes qualities the most EHRs lack, but workflow platforms do very well.
From Ten Years Ago, Dallas HIMSS: Landmark Presentation on BPM EMR Workflow Groupware (by the way, 13 years ago the EMR/EHR was called the Computerized Patient Record, or CPR):
[begin quote]
The best way to present our own comments on the importance of modular and extensible EMR component architecture is to highlight two slides along with their slide handout notes (still available on the HIMSS website and archived here). They refer to Microsoft’s COM objects, not today’s Web services, .NET components, plugins, and so on, so I’ve supplied an updated translation. These slides are also animated, so be sure to click on them.
(I admire the WordPress plugin system that allows me to extend my blog to publish these Flash videos based on ten-year-old PowerPoint slides. The average WordPress blog has five plugins. This one has fifteen. Works great. It’s similar to the point I’m making about extensible EMRs–ironic!)
Two slides from our presentation (animated):
[kml_flashembed fversion=”8.0.0″ movie=”/video/EMR-pediatric-clinical-workflow-groupware/extend/CPRs-should-be-extensible2.swf” targetclass=”flashmovie” publishmethod=”dynamic” width=”540″ height=”418″]
[/kml_flashembed]
From PowerPoint slide notes:
“Allow developers to customize the workflow-enabled CPR, since one size does not fit all. Here Microsoft’s Component Object Model (COM) is critical, since it allows a developer or VAR to add their own screens as options for selection by the workflow engine.”
Updated translation:
“Allow users and developers to customize EMR workflow groupware systems, since one size does not fit all. Here Microsoft’s .NET, Web services, plugins, and other modular means to extend EMR functionality are critical, since they allow a user, developer or reseller to add their own screens and functionality as options for selection by the workflow engine.”
[kml_flashembed fversion=”8.0.0″ movie=”/video/EMR-pediatric-clinical-workflow-groupware/architecture/component.swf” targetclass=”flashmovie” publishmethod=”dynamic” width=”540″ height=”418″]
[/kml_flashembed]
From PowerPoint slide notes:
“This slide is for the programmers in the audience (please explain it to the non-programmers). Decide to work with a workflow-enabled CPR that relies on a COM architecture possessing a set of publicly accessible APIs, so you can assemble a best-of-breed component solution and customize to your users or market.”
Updated translation:
“This slide is for the programmers in the audience (please explain it to the non-programmers). Decide to work with an EMR workflow groupware system that relies on a modern modular architecture possessing a set of publicly accessible APIs (application programming interfaces, check out the WordPress plugin API), so you can assemble a component solution customized to your users and market.”
After our 2000 presentation we continued to emphasize the relationship between clinical group workflow requirements and modular componentized EMR platforms:
“Workflow management systems are usually highly componentized, in that the workflow engine does not need to know much about the applications that it executes (just the prerequisite circumstances for execution and what context information to supply). This componentization provides a route … to introduce new EHR functions or ways of accomplishing them (such as a new decision support module or data display) into work-a-day … settings.” (“EHR Workflow Management Systems: Essentials, History, Healthcare”, TEPR Conference Proceedings, 2004, Fort Lauderdale)
[end of quote]
Let’s take a look at more items on this white paper’s wish list. There’s a section, starting on page 20, titled Extensible. I already covered workflow technology’s uniquely extensible properties above. But then there is the following…
“Every unit process includes a trigger, a particular data view, a recommended action (prompt, alert, decision support, etc.), an execution-communication-documentation of action including any new data associated with that action, and an update of relevant data.”
and a diagram, from left to right…
About which they say: “The workflow unit process described above can be oversimplified as trigger-> information -> action -> information. Finding rich new ways of visualizing information is a key factor in assuring that workflow processes can be defined in ways that are highly intuitive and compel user engagement and effectiveness.”
In other words, a defined workflow is triggered and then executed while interacting with data and users, leading to an action. This is not an uncommon vision of EHRs becoming proactive members of the virtual care team. However, what has been lacking (at least within the confines of healthcare and health IT) is an effective, efficient, and flexible workflow platform for making this vision real. I argue we (healthcare and health IT) must import from, and partner with, the workflow management systems industry, now usually called the business process management industry, to finally bring this vision to reality. I wrote about this in Clinical Intelligence, Complex Event Processing and Process Mining in Process-Aware EMR / EHR BPM Systems. Here’s the outline from that presentation (on-line via the previous link).
Malleable and systematically improvable workflows, mediating the trigger to action loop, will be necessary to realize the vision described by this white paper. And the best representatives today of workflow platforms with malleable and systematically improvable workflow are called business process management systems or suite. See my EMRs and EHRs Need to Solve “The BPM Problem”: Why Not Use BPM to Help Do So? in this regard.
Later on the same page, in the section titled Checklists & Protocols is this…
“Increasingly in the healthcare world checklists are being seen as a strategy for defining and securing compliance for complex, high impact processes. The workflow platform should be able to richly leverage the checklist pattern both for development and for implicit user experience.”
I write about precisely this in response to recent call from Atul Gawande (a famous checklist proponent) at the recent Health Datapalooza in DC. See My Second Blog Post Inspired by Health Datapalooza: Workflow, Communication, Tasks, and BPM.
: need to take work back from providers–workflow, communication, task management. "That's when you get sudden adoption"
— Lauren Hatch Fifield ()
About which I write:
I agree!
If EHRs and health IT systems need
- workflow,
- communication,
- task management,
where can we find the technology to provide this functionality?
And then I simply search in Google for workflow+communication+task+management. The ten links I get back are all workflow management systems related, so take them one at a time. I pull out the abstract or intro that seems most relevant to Atul Gawande’s call to action. He’s calling for the same thing that Electronic Health Record (EHR) Solutions LTPAC Providers Need Today calls for: True EMR and EHR workflow platforms.
Finally, on the next page, in a section title Social Collaboration is this …
“In our post-Facebook world, the benefits of socialization of work and workflow are becoming better understood. The ability to “share”, “like”, “follow”, “group chat” people, tasks, events, objects, records, and documents can significantly enhance any work team’s cohesiveness, effectiveness and productivity. While there are particular security concerns that need to be accommodated, care teams are no exception. Socialization adds an accelerator dimension to all aspects of the workflow platform”
Again, BPM has what healthcare and health IT needs: the “Simple, no-training social interface.” From one of my trip reports to a BPM vendor’s conference…
The “no-training social interface” referred to is the now almost classic activity stream. Folks are used to Facebook, LinkedIn, and Twitter, why not use a similar user interface? It takes less training and is naturally social to boot. Since each platform doesn’t require a programmer to create a different program, and therefore different user interface, consistency across platforms is achieved. So, not only does previous user experience with social media transfer to the EHR, training on one EHR platform (say, desktop Web) transfers to another (say, mobile, or, increasingly, from mobile to Web and desktop)
Essentially, what Electronic Health Record (EHR) Solutions LTPAC Providers Need Today describes are EMRs and EHRs implemented on a modern workflow platforms commonly called business process management suites.
Earlier is a section titled Call to Action for Software Providers and the following …
“This EHR whitepaper proposes a vision for the delivery and use of an integrated, secure and optimized on-line system for the LTPAC care team, care delivery and interaction with the patient. This on-line system does not exist today.”
No, it does not. But it can, if we combine the clinical content and functionality associated with traditional EMRs and EHRs with the context-aware, effective, efficient and flexible workflow of modern, extensible, BPM suites, especially as they add social workflow, move to the cloud, and deploy to a variety of devices without having to write a separate app for each mobile platform.
So? What are we waiting for?