Month: March 2009

Short Link: http://j.mp/4Io0Ya

Someone contacted me with a challenge, “What’s so special about EHR workflow management systems and why can’t it be added to an existing EHR?”

Here is my answer:

Many EHRs are cumbersome, inflexible, and difficult to optimize with respect to their process workflow. Implemented correctly, an EHR workflow management system is graceful, flexible, and optimizable. If you understand the reasons for these advantages you will also understand why other kinds of EHRs cannot easily fix their problems. By the way, this is not to say that EHRs have not added important task management capabilities in recent years. However, this task management is typically based on “frozen” workflow.

By “cumbersome” I mean that most EHRs require expensive human users to do what should be done by less expensive combinations of software and hardware. Physicians don’t want to become data entry clerks; they just want to click once and as Star Trek’s Captain Picard says, “Make it so!”

By “inflexible” I mean there is no way for the user to easily improve task workflow in order to spend more time on value-added tasks directly benefiting patient and user. If an application requires a physician to click five times, four clicks of which are effortful and extraneous, there needs to be some sort of workflow or process editor to eliminate the four non-value added clicks that are wasting the physician’s time.

By “difficult to optimize with respect to process workflow”, I mean there is no way to systematically guide changes in EHR workflow so as to maximize patient satisfaction, clinical outcomes, and practice profitability goals. In contrast, processes driven by EHR workflow management systems can be analyzed through use of business process management (BPM) tools that suggestions ways to improve workflow processes. Workflow engines create workflow logs. These are step-by-step records of who clicked on what, when, where, and why. BPM tools can analyze these logs (through a technique called workflow or process mining) and suggest better workflows that will minimize non-value added tasks. This in turn frees human resources to be reallocated to value-added tasks that contribute to happier, healthier patients and greater take home income.

OK, fair enough, EHRs will need to incorporate WfMS technology. Why can’t EHR vendors just add workflow engines, process definitions, workflow logs, and process mining to existing EHRs? The word “workflow” has certainly been an EHR industry buzzword for the last five years. However, in the marketing din “workflow” has become almost meaningless. Yes, addition of messaging facilitates person-to-person coordination; interfaces make application-to-application coordination possible; and patient tracking is about coordinating the most important resource of all, patients. While these added capabilities can improve workflow, they aren’t workflow management. A workflow management system by definition requires a workflow engine that consults workflow or process definitions to drive tasks to people and applications. Tacking on messaging, interfaces, and tracking is a lot easier than picking up an EHR that does not rest on a workflow management system and then inserting beneath it a fully fledged workflow management system foundation, with its powerfully directive workflow engine, powerfully customizable process definitions, and powerfully analyzable workflow log.

Here is one way to think about “adding” a workflow management system to an existing EHR. When you look at a non-WfMS-based EHR you are looking at screens that are the result of a human programmer creating areas that will contain buttons and menus and so on, then placing these buttons and menus in these areas, and then connecting those buttons up with various functions and procedures that have also been created by a human programmer. This is why we need programmers in the first place; if it could be done more easily or less expensively we’d do it that way instead.

This is in fact what a workflow management system allows a non-programmer to do, to directly edit application workflow. Who creates the areas for the buttons, and then the buttons, and then connects the buttons with what happens when the buttons are pressed? The workflow management system does. You should now see why a workflow management system cannot simply be added to an existing computer application. The existing application was created by a human programmer. In order to add a workflow management system foundation you will have to replace the programmer with a workflow management system to regenerate the application. Therefore existing systems will need to be rewritten (by the very definition of how workflow management systems operate).

Several years ago we were giving a demo to a visiting physician who had happened to have taken some programming courses in college. During the demo the physician said, wait a minute, I thought you were going to demonstrate integration with my favorite patient questionnaire application. To which the user said, “Oops!” And then proceeded to pop up the process workflow definition editor within which he added the questionnaire task, set a couple of properties of that task, dismissed the editor, and then gave the demo again. This time the questionnaire screen popped up automatically at the intended step in the workflow. To which the visiting physician said, “I get it!” An EHR workflow management system is a development environment that lets non-programmers to create and edit their own EHR workflow systems. Exactly.

Adding a workflow management system to an existing EHR application would be like adding a foundation to a standing skyscraper or a hull to a floating ship. Many current EHRs will have to be rebuilt on top of workflow management systems foundations if they are to become the graceful, flexible and optimizable EHR systems that healthcare needs in the long run.

I have (somewhere) a copy of a several hundred page user manual for a typical document management system-oriented EHR. It has a hundred pages of chapters about workflow. These chapters tell an human user what to click, in what order, and in what circumstances in order to perform a variety of tasks. If you look at chapters about workflow for an EHR workflow management system, you will see that they are about editing EHR workflows so that these tasks happen automatically after a user just clicks the button and “fires and forgets.”

EHR workflow management systems are built on a foundation that can “Make it so!” This is the EHR workflow management system unique selling proposition *and* its barrier to entry for EHRs that are not workflow management systems.

Short Link: http://j.mp/86lspq

In 2003 and through 2005, I contributed to, and updated, the General Workflow Management Criteria portion of the “Workflow Management Survey: Ambulatory EHR Systems” survey, conducted by Andrew & Associates, which appeared in Advance for Health Information Executives. It’s been four years WfMS technology has been creeping into the EHR mainstream, so it’s worth an update and a raising of the bar, so to speak. 

Here are the originally described EHR WfMS features and functions:

1. System displays selected worklists for active cases.
2. System permits employees to view and complete work items that have been assigned to other employees.
3. System creates reminders for work items that have not been completed when due.
4. Users can selectively modify assigned work items.
5. System maintains a record of various changes made to work items.
6. System maintains various records of completed activities.
7. Users can selectively correct or modify records of completed activities.
8. System has a workflow engine that automatically creates work items based on a workflow definition using various defined criteria.
9. Users can customize workflow definitions to match their internal processes using various defined criteria.
10. Users can define roles and resources that will receive task assignments.
11. Users can edit workflow definitions using a graphical user interface.
12. Software vendor provides workflow definitions for individual medical specialties.

A good start six years ago, however they actually barely scratch the descriptive surface of a fully functional EHR WfMS. (Although, if one had to prioritize, criteria 8, 9, and 11 are arguably the most important and the basis for my Litmus Test for Detecting Frozen EHR Workflow post.)

I’ve tentatively expanded these 12 criteria almost fourfold on a special page I’ve added to this blog titled, Electronic Health Records Workflow Management Systems Features & Functions Survey. If you fill out the survey, whatever comes in over the transom I’ll summarize back to the blog. By “tentatively” I mean that this survey is more of a rough public draft intended to attract comment, critique, and suggestions for improvement than anything that could be considered a polished final product.  I’d rather put out something imperfect now so it can be improved by critique than hold onto it while making successive smaller and smaller improvements.

This EHR WfMS F&F Survey is *not* intended to be an exercise in systematically comparing EHR WfM and EHR WfM-like systems. It could be used that way, however my primary goal is to raise awareness about the details and importance of EHR WfMS features/functions, and to create and improve a conceptual framework for understanding EHR WfMS and EHR WfMS-like applications.

You are welcome to submit data anonymously or not, to complete all or only some items, or to only provide comments on specific items. I am particularly interested in comments from workflow management systems and business process management professionals who would like to see this technology and way of thinking brought to healthcare in general and electronic health records in particular.

I’ll make minor revisions to the page in place. However, for larger revisions I’ll release a new version of the survey and archive the old one.

Yes, I know my terminology may be somewhat at variance with that of the WfMS industry. And yes, I know that in some cases I essentially ask the same question more than once in different ways. This is partly due to the old survey stratagem of asking the same question in different ways as a form of validity check and partly due to my observation that, in the electronic health records industry, workflow management systems terminology is not yet, in fact, typically standard with that of the workflow management systems industry. In other words, an EHR professional may use (and therefore recognize) other words than that adopted by the WfMS industry and I am trying to be accessible to both EHR and WfMS professionals.

I eventually plan to add business process management oriented material, however while WfMS ideas and terminology have become relatively standard, BPM is a moving target, so I’ll focus first, here, on WfMS aspects of EHR functionality. That said, implications of BPM for electronic health records are fascinating and important, so I’ll certainly post on the topic.

My hope is that improving these survey criteria might eventually become a community effort. Perhaps some sort of open source model based on ideas from the creative commons might apply. I’m thinking of a downloadable version coded in XML to facilitate sharing and co-development. (Is there an XML questionnaire schema that might be appropriate here?) I’m open to suggestions.

Short Link: http://j.mp/8xfEPJ

A friend of mine is Dave Hubbard, the motivational speaker, an All-American collegiate athlete who played professional football in the 1970’s for Hank Stram of the New Orleans Saints and then the Denver Broncos. Dave is extremely familiar with healthcare both professionally (serial entrepreneur) and personally (broke his back jumping out of a perfectly good airplane!).

We’ve had a series of interesting conversations about, of all things, football plays and EHR workflow. We co-wrote the following post to see if we might enlist you in telling us if you think we’re on to something, or not.

–Chuck & Dave

Medical offices are like football teams: an offensive line moves a patient encounter forward while a defensive line seeks to create chaos and stop encounter progress. This analogy is more productive than you might think!

Medical office staff members interact in ways that are similar to a football team. For example, they have an offensive line whose responsibility it is to efficiently, effectively, and flexibly move an encounter from waiting room to checkout. There’s a quarterback who calls plays. Sometimes it’s the physician who directs staff to administer a vaccination or auditory test; sometimes the plays are called automatically based on the reason for the patient’s visit, such as “well child” versus “ear ache.”

Tasks are “passed” among team members, such as a nurse gathering vitals and checking medications and allergies before passing the assessment and treatment tasks to the physician. “Dropping the ball” results in inefficiency that slows the encounter and ineffectiveness that affects patient care and physician revenue.

The defensive line may be less obvious, but it consists of threats to the accomplishment of efficient, effective, flexible workflow. It is the offensive line’s responsibility to protect this workflow. For example, the phone nurse blocks defensive line interruptions that would otherwise distract the physician from maximizing use of the most important and constrained resource in the practice, his or her time. Anyone (or anything) who contributes to the hassle factor of practicing medicine is part of the defensive line.

Similarities between a medical team and a football team are more than an amusing analogy. All teams are cognitive systems, and their study is called team cognition (with contributions from distributed cognition). Shared mental models, workspace awareness, radar views, and teams of experts versus expert teams are topics of team cognition that apply to all teams, including those in medicine and football.

Using this football metaphor (and some ideas from cognitive science), we encourage you to think (and comment!) about office processes from the perspective that to win, the ways in which the plays are being run must be examined. Doing so will allow people to express what they are most proud of, but also to critically evaluate performance problems in a constructive way, one in which everyone is committed to success.

Questions to consider:

How is your medical practice similar to a football team?

What position does each employee play? Who is offense? Who is defense? Are there any special teams?

Who is the quarterback? The coach? Does everyone know their position?

What about the referee? The coach? The patient? The fans?

Who owns the team? Is there an owners’ association? A players’ association?

If your medical practice were a football team, what would be your version of the following: Holding? Tripping? Unsportsmanlike conduct? Unnecessary roughness? Running versus passing? Huddling? Incomplete pass? Field goal versus touchdown? Memorizing key plays? Time-out? Substitutions?

Suppose you could review game films with your staff. What are examples of plays you’ve run to achieve major yardage or touchdowns?

What are examples of plays where you’ve thrown for a loss, fumbled the ball, or suffered interceptions? Why did they occur and what can you do to keep them from happening again? How do you define victory?

Do different styles of medical practice lend themselves to different sports analogies? Soccer? Golf? Which do you suggest and why?

Short Link: http://j.mp/4X17lj

The recommended length of a blog post is between 250 and 1000 words. Since a picture is worth a thousand words, I am posting the following graphic without further comment (except to ask “What I am trying to represent?” and that I look forward to your comments).

Short Link: http://j.mp/79mGR7

If a first year medical student says that a patient has a temperature, his or her attending may say something like “Of course your patient has a temperature, all patients have a temperature! Is their temperature normal or abnormal? Elevated or subnormal? What, exactly, is the patient’s temperature?

In the same sense, all software including EHRs, have workflow. The question is whether the workflow is good or bad and whether you have the means to adapt it to your purposes. EHRs whose workflow cannot easily be modified by a non-programmer have what might be called “frozen” workflow.

Most EHR applications require the user to navigate between screens to enter data. An EHR workflow management system can be configured to drive common input screens in any order that makes sense. This is most commonly based on the visit reason. In other words, an EHR workflow management system can deliver customized workflow around a practice’s current workflow process for gathering data. For example, a practice may currently use a paper based workflow system that they like. This not only involves what data they gather, but when and who collects the data during the workflow. An EHR workflow management system can be configured, through use of its process definition editor, to precisely match this pre-existing workflow.

Some EHRs do increasingly deliver some workflow capability; that is, they can contribute to an increase in productivity by doing for the user what they would otherwise need to do themselves (navigate to the next screen, inform the next person what they need to do, trigger an external application such as a digital ECG, and so on). However, these EHRs rely on “frozen” process definitions. Their workflow cannot be easily changed to adapt to circumstances not foreseen by the programmer. Programmers are not clinicians, or even more important, not intimately familiar with each and every minute step of dealing with a patient in medical practice. Only users can claim this degree of familiar knowledge. This means that users cannot easily change what EHR screens occur in what order, who gets passed the task baton when, or what external application to fire up automatically.

Only a full-fledged EHR workflow management system provides additional user interfaces that allow users, not programmers, to bend EHR workflow behavior to their specific and evolving needs. For example, a user should be able to easily instruct the EHR to insert a new data gathering step between vitals and chief complaint, or to easily add the automatic playing of a preventative care video at the optimal point of a well visit.

Apply the following litmus test designed to detect frozen workflow:

The simplest test of whether an EHR is built on a workflow management system is to ask for a live demonstration of the following:

1. Ask to see an encounter from beginning to end. Focus on the sequence of screens.
2. Ask to see the process definition that controls the sequence of screens just observed.
3. Ask to see a small edit in the process definition using the EHR process definition editor, such as the deletion or reordering of several steps.
4. Ask to see the same encounter again, while focusing on whether or not the changes in the process definition have indeed resulted in the appropriate changes in screen sequence.
5. If the screen sequence changes in just the way that would be expected if an EHR workflow engine is consulting the just edited process definition, then you are likely looking at an EHR workflow management system.

If an EHR cannot demonstrate steps (1-5), then the EHR lacks the capabilities of a workflow management system. It does have workflow (because all software applications have workflow). It may even have workflow that is good for a particular task and context. However its workflow is frozen.