Building your #Workflow Glossary

Hi fellow Clinical #Informatics and other #workflow enthusiasts, 

Happy 2019! While I continue to work on compiling the business case for Clinical Informatics, I thought I'd take a minute to talk about #workflow terminology. 

A. THE BACKGROUND :
Simply put - words matter. Any bilingual person who has ever tried to translate the phrase 'scram' or 'hit the road' into another language knows that a word-for-word translation does not always work. (Really? Hit the road..?) One might try to translate it as 'it's time to leave', but even that fails to convey the certain informal, vernacular quality that the phrase 'hit the road' conveys so well. So my advice to anyone working in a translational role - Do your best, but always translate at your own risk. 

In healthcare, we have a number of terms that people generally understand, but their exact definitions may vary from organization to organization. They include such common terms as : 

• Order
• Order Set
• Protocol
• Policy
• Procedure
• Guideline
• Standing Order
• Clinical Pathway
• Documentation
• Templates
• ... and more!

While almost all clinical staff have a general sense of these terms, their true understanding may not be exactly the same - And so, with regard to the term ‘protocol’, for instance, they may quietly have overlapping circles of a common understanding :

The problem is that these differences in understanding may result in dramatically different expectations about how exactly a 'protocol' works, and what it can do to help their workflow : 

• Can a protocol be used to allow a Registered Nurse to titrate an IV heparin drip?
• Can a protocol be used to allow a Registered Nurse to give a pneumonia vaccination?
• Can a protocol be used to allow a Respiratory Therapist to titrate the settings on a ventilator in the ICU?
• Can a protocol be used to allow a Registered Dietitian to modify a diet for an inpatient?
• What is the difference between a protocol and a standing order?

To increase the amount of common understanding, it's helpful to look at your federal and state regulations, along with your own safety and operational needs, to see if they offer any definitions that help clarify the answers to these questions :

After all, once there is a clear definition - then you can create a standardized template, development procedure, and staff education to give everyone on your team a greater, more standardized understanding of the tool and what it can do. Remember - It all starts with the definition.

B. THE PROBLEM :
Healthcare faces some challenges in harmonizing this terminology - What a protocol can do in some organizations is different than what a protocol can do in others. And despite CMS regulations which refer to the use of protocols, many federal and state regulations use these terms interchangeably - See this 2013 letter from the Centers for Medicaid Services (www.cms.gov), page 4 : 

Standing orders: Drugs and biologicals may be prepared and administered on the orders contained in pre-printed and electronic standing orders, order sets and protocols (collectively referred to as “standing orders” in our guidance) only if the standing orders meet the requirements of the medical records CoP.

And this, from the Interpretive Guidelines §482.24(c)(3) on page 78 : 

There is no standard definition of a “standing order” in the hospital community at large (77 FR 29055, May 16, 2012), but the terms “pre-printed standing orders,” “electronic standing orders,” “order sets,” and “protocols for patient orders” are various ways in which the term “standing orders” has been applied. For purposes of brevity, in our guidance we generally use the term “standing order(s)” to refer interchangeably to pre-printed and electronic standing orders, order sets, and protocols. However, we note that the lack of a standard definition for these terms and their interchangeable and indistinct use by hospitals and health care professionals may result in confusion regarding what is or is not subject to the requirements of §482.24(c)(3), particularly with respect to “order sets.” 

Making it even worse is when Informatics professionals then have to compare this with their state regulations :

... which may have slightly different understandings and definitions of these terms.

Fortunately, there are some very talented medicolegal and compliance experts out there, who can help an organization to develop a strategy for navigating these regulations, while planning their workflows, both before and after an EMR implementation. One of the best I've seen is Sue Dill Calloway, BSN MSN JD, who has a fantastic series of lectures on the importance of this terminology, for regulatory, financial, and patient safety reasons.

But in the absence of a simple, standardized national glossary, with good functional definitions of these tools - It can be very hard to develop the templates, development procedure, and education you need for your team. 

C. THE SOLUTION :
Given the lack of clarity about these terms, what's the average CMIO, CNIO, or clinical informaticist to do? Fortunately, there is a strategy you can employ, and that is expanding upon a fairly simple template for functional definitions : 

[ Term: What It's Called ] - [ Functional Definition: What It Does ] 

This simple template is helpful in separating terminology for tools that have slightly different functions, e.g. : 

Term1 - FunctionalDefinition1
Term2 - FunctionalDefinition2
 ...and so on...

So if we can accept this simple template for separating terminology and function, we can then start to draft a 'conceptual map' for these common terms in healthcare (click the image below to enlarge) : 

(REMEMBER - THIS GRID IS JUST A DRAFT AND IS NOT COMPLETE!)

As you start to do this exercise, you'll see that there are some terms which have very similar functions, and other terms which don't : 

• Guidelines and Policies initially look like they might have similar functions - until you consider that policies might result in root cause analysis and disciplinary action, and guidelines don't. (Policies=rules, guidelines=suggestions).
• Protocols and Standing Orders seem to have very similar functional definitions, so we need to figure out if they are true synonyms, or if there is some kind of a difference between them.
• Procedures and Plans also have similar definitions - So we will need to figure out how to separate them. In this case, I've taken the liberty of separating them in time, suggesting that procedures describe current tasks, and plans describe future tasks. 

Given the similarities between protocols and standing orders, it's helpful to separate them by considering their risk - and thus their initiation/triggering mechanisms, FOR EXAMPLE: 

• Standing Orders = Used for common, LOW-risk clinical scenarios in which the benefit to the patient of rapid evaluation and care outweighs any known risks. Standing orders may be initiated ('triggered') by a clinical POLICY (e.g. 'All clinic patients will be screened and potentially administered for pneumonia vaccination, according to the Standing Orders for Pneumonia Vaccination.) All orders and outcomes of standing orders will be attributed to the attending provider.
• Protocols = Used for common, HIGH-risk clinical scenarios in which the benefit to the patient of improved care standardization outweighs any known risks. All protocols must be initiated ('triggered') or discontinued by an ORDER (e.g. 'Initiate Ventilator Liberation Protocol' or 'Discontinue Ventilator Liberation Protocol'). All orders and outcomes of clinical protocols will be attributed to the ordering provider.

While you undergo this exercise, it's important to look at your regional, state, and federal regulations, and to speak to experts (like Sue Dill Calloway, BSN MSN JD as I mentioned above). If there are no regulations to guide you in this grid, then you and your clinical and administrative leadership will have to make local decisions about how your organization wants to define these tools.  

As you work on these definitions, keep in mind other things you can do to improve safety and clarity, e.g. "Orders are documented instructions [ that ] must be signed within 24 hours."

As you start to build out this grid for your own organization, talk to people who use these tools, and you'll start to better understand the form, function, and other issues related to their design. And once you think your grid is complete? Bring it back to your senior leadership for review, discussion, and formal approval. Voila! You now have your own organizational glossary that will help you develop the templates, procedures, and education that create a greater understanding, and improved standardization, predictability, and efficiency, for both your clinical and administrative teams. 

Hope this is helpful in guiding you to build your own workflow glossary! If you have any other tips, suggestions, or comments, leave them in the comments section below!

Remember - This blog is for educational discussions only. Do not use any of these definitions without formal review and discussion with your own informatics, legal, administrative, and clinical teams. Have any other clinical terminology tips you'd like to share? Feel free to leave in the comments below!

Raising a Well-Supported Workflow

Hi fellow Informaticists and other clinical leaders,

Long time no post - But I'm glad that I had some time this weekend to catch up on my blog. Have several pieces in the works, but today's is a roughly 18-minute video I put together on "Raising a Well-Supported Workflow".

Having workflow challenges? Not sure what the impact will be if you change a document? I'm hoping this video will help. Remember, workflow design isn't hard, once you see the big picture. It's a lot like propping up a tent, pole-by-pole - The trick is to know what poles you will need to prop up your tent.

So with that, I'd like to offer up this video for your consideration. Special thanks to Charles Webster, MD from ChuckWebster.com for his awesome definition of "workflow"!

Hope this was helpful to you, our future healthcare leaders! If you have any thoughts or comments, please leave them in the comments section below!

How to spot and fix a Frankenform

Hi readers - As I write this last post for 2015, I'm going to explore one of the most common questions that Informatics professionals get asked : "Why can't you just make this paper form electronic?" - 

The most common answer that Informatics professionals respond with is, "Well, you just can't", usually followed by some sort of an awkward smile and an answer that sounds like, "You know those computers, they are always so difficult." 

But there is a real reason, with a better answer. To help explain that answer, I'd like to introduce a new concept - The Frankenform.

A. FIRST, SOME COMMON DEFINITIONS

Before we go on, I just wanted to review some DRAFTED definitions of a few common archetypes we all use in healthcare. They include : 

1. POLICIES - Tools used to describe an organizational standard
2. PROCEDURES - Tools used to describe a series of actions conducted in a certain order or manner
3. GUIDELINES - Tools used to educate staff about how to achieve a desired outcome
4. ORDERS - Tools used to document an instruction to deliver a defined type of care to a defined patient at a defined date/time in a defined manner (sometimes for a defined reason)
5. ORDER SETS - A collection of ORDERS used to standardize and expedite the ordering process for a common clinical scenario
6. CLINICAL PATHWAYS - A collection of ORDER SETS used to standardize the care for a common clinical diagnosis
7. PROTOCOLS - Tools that allow a nurse, pharmacist, or other licensed medical professional to start/modify/stop a patient care order on behalf of a licensed physician.
8. DOCUMENTATION - Tools used to record and transmit information
9. STAFF/PATIENT EDUCATION - Tools used to educate staff/patients about a particular topic
10. STAFF SCHEDULES - Tools used to determine who is responsible at a particular date/time
11. BUDGETS - Tools used to allocate resources for a project or initiative

These definitions will be helpful in spotting Frankenforms - Forms that often combine these functions.

B. THE FRANKENFORM

So what exactly is a "Frankenform"? It's a term that Informaticists sometimes use, loosely, to describe paper forms that combine more than one of the above functions, or are used by different people in different scenarios. If I had to write a better definition, it might be described as : 

Frankenform (n.) - a form or document that is designed with more than one archetype, role, or scenario in mind.

Frankenforms existed all over in the paper world. While they are often convenient (having everything in one place), they generally don't behave well in the electronic world because : 

1. They contain documentation recorded from two different roles. (e.g. the dietitian and physician both signing off on one TPN order)
2. They contain two different archetypes (e.g. part-policy-part-order-set, or part-policy-part-guideline, or part-documentation-part-order-set, etc.
3. They contain documentation from two different scenarios. (e.g. the sometimes-seen, all-encompassing "antibiotic order set" which contains antibiotics to cover all scenarios, from pre-op antibiotics to treatment of sepsis) - These all-encompassing tools are sometimes also called "pick lists", because they can be used in almost any scenario.)

Electronic medical records generally will not allow you to build Frankenforms into their systems because of these three reasons :

1. They enforce legal-grade authentication - So a form used by two different people must be re-engineered to find out who-is-filling-out-what-part-of-the-same-form
2. They are engineered to enforce archetypes - Generally, order sets are found in the order set section of the software, documentation is found in the documentation section, and guidelines and protocols may not be contained in the software at all. So while you can link from your clinical documentation TO your order set, or link your order set to a set of clinical guidelines, you can't actually put documentation and orders in the same part of the software.
3. Taking advantage of good, electronic Clinical Decision Support (CDS) generally depends on a clear, linear workflow.

Interestingly, most EMRs will still allow you to make orders, order sets, or documentation to address two or more different scenarios - While it's quite not as unorthodox, it still can lead to very lengthy documents/order sets with poor decision support that can frustrate users in the long run because they require a lot of clicking to complete them.

So let's now look at these three different types of Frankenforms in a little more detail.

C. THE "TWO DIFFERENT ROLE" FRANKENFORMS

A common workflow challenge is when two different roles are involved in ordering/documenting in the same workflow. The classic example of this is the Dietary TPN order, which is usually one order with many fields - Some are filled out by the physician, and some are filled out by a Registered Dietitian. 

You can often spot these forms because they have multiple signatures on them. While it's important to have both signatures before processing the order (often for safety/billing reasons), having two different signatures can cloud the workflow that led to the completion of the order - Who filled out which field? Did the dietitian enter the potassium, or did the physician? If the potassium needs to be raised, who does a nurse call? The physician? Or the dietitian?

Fixing these, to make them electronic, can be very complicated - and often means a good deal of workflow analysis and redesign. The electronic solution will typically have electronic orders and documents with electronic co-signatures, but will often mean a more rigid workflow (e.g. having to decide who starts off the workflow, the doctor or the dietitian? And who cosigns the order? And how do they attribute the cosignature to the right person?)

D. THE "TWO DIFFERENT ARCHETYPE" FRANKENFORMS

This is the scenario where a paper form with one signature actually contains components from two different tools, e.g. an "ED Nursing Protocol" which is part-documentation, part-orders, and part-guideline. Again, these were very convenient in the paper world, because you could have all-the-information related to the workflow in one place. 

While these are sometimes easier to build in an electronic environment (provided they really only have one stakeholder), they still generally require separation of the tools into their electronic components - E.g. the documentation in one place in the software, LINKED to the guidelines for review, LINKED to the orders that get activated. 

Often, while dissecting these Frankenforms into their separate components, workflow questions arise which must be looked at to ensure safety and regulatory compliance. Again, this is usually not hard to overcome, but it should be expected that converting these paper forms to an electronic workflow will take some additional time and resources. 

E. THE "TWO DIFFERENT SCENARIO" FRANKENFORMS

(Also sometimes referred to as "pick lists") - While this Frankenform looks fairly innocent (who wouldn't want all of their antibiotics on one order set?), it's generally a sign of a larger workflow issue. These Frankenforms, seeking to address many-different-clinical-scenarios-with-one-tool, can require the most time to redesign because they usually raise larger workflow questions, bigger than the form itself.

For example, having a broad "ED Antibiotics Order Set" means you are missing opportunities to develop disease-related, evidence-based order sets, which often involve more than just antibiotics. While in the short-run, physicians may like having all of their antibiotics in one place, they may get frustrated looking for other medications related to disease management, and/or miss other quality indicators. 

The solution to this type of Frankenforms is generally the construction of a larger library of disease-related, evidence-based order sets, focused on common disease pathways that doctors are responsible for initiating after they reach a diagnosis. (And so you might even want a separate library of order sets used to work up common chief complaints, to help them reach this diagnosis!)

While creating (and maintaining!) this larger library can take time and resources, it generally results in shorter, disease-related order sets, which are focused on the total management of the patient in an evidence-based manner, with better decision support, better provider satisfaction, better quality compliance, and better overall time savings. 

F. HOW TO FIX FRANKENFORMS

If 'going electronic' means that you will need Informatics resources to identify and fix existing Frankenforms, then budgeting for a successful EMR implementation generally means : 

• Conducting a complete review of all current clinical documentation and forms.
• Developing a good understanding of your current workflow issues by estimating the number (and type) of Frankenforms currently in use
• Planning and budgeting for the informatics resources necessary to fix (and maintain!) your solutions.

And so the first step in solving these issues is finding an experienced Informatics or workflow professional, and asking them to do a good current-state and needs analysis. The answer will help determine your success and satisfaction with your new EMR implementation!

I hope this post has been helpful in creating understanding and clarity. Thank you so much for reading my blog, and many happy wishes for 2016!

Have any thoughts about workflow redesign and optimization? Leave them in the comments section below!

Cutting Healthcare Costs by Making A Better Brick

1. THE BRICK

An interesting question I get asked is, "Why don't all order sets look the same at every hospital, even for the same disease?"

This question can be posed in several other ways, including :

1. "Why can't we just use order sets from someone else?"
2. "Why can't we just use canned order sets without editing them?"
3. "What do your policies look like?"
4. "Why can't we just use canned policies?"
5. etc...

But all of these basically ask the question, "Why isn't this standard?" and, of course, the follow-up question is, "Why is every hospital re-inventing the wheel?".

For inspiration to answer this question, I'd like to first explain a little bit about a wonderful thing : The brick.

According to the Wikipedia article, bricks have been in use since about 7500 BC to help construct things. (It's actually a really interesting article - If you appreciate human civilization, the brick has played a big role in building our streets, aqueducts, houses, walls, etc...)

The reason the brick is such a useful thing, from a design standpoint, is because it has two features :

1. A brick has a fairly predictable shape that allows you to easily arrange them to connect two or more places in space.
2. A brick is designed to withstand a particular load.

You'll notice these two features are helpful when designing any system - Having basic units engineered in a predictable manner, which can be assembled to make a bigger, more complex system that achieves a certain goal.

Hospitals contain systems that are also made of smaller units - Order sets are made of orders, clinical pathways are made of order sets, charts are made of notes, policy manuals are made of policies, etc. 

Unlike a physical brick, however, all of these basic units are conceptual - They are mostly complex documents - not physical objects - so it's a little harder to tell how predictable they are - e.g. to know if they're not engineered exactly the same.

For example, with physical bricks, it's much easier to tell if one hasn't been engineered exactly the same as the others :

You can immediately and obviously see the entire system is off, and locate the offending brick quickly.

But real bricks are not 100% predictable - They all have some degree of imperfections between them -As a kid, I occasionally ran into bricks behind our grade school, or around landscaping projects - I think I could only stack about 10 of them on top of each other before they start to fall over. (Legos are about the only bricks I can think of that are engineered to such a high standard that you can stack virtually hundreds of them on top of each other and still have practically a straight wall.)

But because regular bricks in the real world have subtle imperfections, humans have developed a tool we can use to compensate for these imperfections : Mortar, or cement.

Mortar/cement only works, however, to help straighten out the system when our brains get involved - We see the system leaning, so we set up guidelines/markers to help determine what is straight, and we put down the mortar/cement to compensate and correct the system. Again, the compensation depends on our human brain.

Because documents are not physical objects, we may not see the system leaning - But it can lean the same way in a conceptual manner. Fortunately, our brains can still compensate for a lot of conceptual leaning. 

So in my job in clinical informatics, I look at the standards by which the "document" bricks are built - To determine just how standard they are, and how much people's brains are compensating.

And this is why hospitals all have order sets that look and behave slightly differently - Because there aren't national standards by which their bricks are engineered. Fortunately, human brains are filling in the mortar.

What would it take to get all hospitals to engineer exactly the same bricks? The same engineering standards at all hospitals. 

And what would it take to get all hospitals to engineer bricks as well as Legos? An engineering process that was detailed enough to ensure that every brick looked virtually identical.

So why don't all hospitals have the same engineering standards and engineering process? Because documents, unlike bricks and Legos, are not as easily understood/studied/observed as physical objects. And because, for better or for worse, human brains can compensate extremely well - So there is little pressure to engineer them exactly the same. 

So as a result : Every hospital is re-inventing the wheel when it comes to their processes and their documents. In almost every hospital, they are looking to achieve exactly the same goal (in brick terms, "bear the same load") - Delivering standardized but customizable, evidence-based, high-quality care. But because the engineering standards and processes are not defined nationally, their tools are all ever-so-slightly different, and so virtually every hospital has to engineer them slightly differently.

2. A NEW IDEA ON CUTTING HEALTHCARE COSTS

The number of people employed to re-engineer all of these tools is remarkable. And it doesn't just include order sets - It includes every document in a hospital, virtually everything I mentioned in the CMIO's Checklist - Order sets, policies, protocols, documentation/forms, templates, etc. And all of these tools have to be continuously updated to reflect best practices, new technology, new evidence, etc.

These operations have become part of the price of healthcare - Continuously re-engineering all of these tools, so that the front-line doctors and nurses have the best and most up-to-date :

• Policies and Procedures to learn from and operate by
• Orders to take care of patients and deliver care
• Order sets to standardize and expedite the ordering process for a common clinical scenario
• Clinical Pathways to standardize care for a common clinical diagnosis
• Protocols to standardize and automate care for a common clinical scenario
• Guidelines to help standardize outcomes for a common clinical scenario
• Documentation tools to record and transmit data about care, and guide their thinking
• Templates to help them standardize and expedite the documentation process
• etc...

Unfortunately, keeping up with all of this information, and managing it, is very challenging for most hospitals. The professional industry term for this is "document management", and hospitals that do this well will probably have an easier time in the next five years than hospitals that do this poorly.

So to help hospitals struggling with this, I have often wondered why nobody publishes national, standard definitions of these tools, so that we could at least have the same engineering standards, and maybe then the same engineering processes - So that the order sets would all look the same - And you could truly use the same order set at any hospital...?

I suspect the reason that no regulatory body currently wants to offer these standard definitions is this : Anyone who tries to introduce these definitions and engineering standards nationally will have a big operational and financial challenge : 99% of hospitals would suddenly have to re-tool all of their documents to meet these national standards. Imagine the cost to healthcare nationally.

But so then I wondered - could we do something like this on a much smaller basis, in a much slower, more controlled manner?

Again I'll mention our Interstate 91 Informatics project - Where a bunch of volunteer Informaticists along the Interstate 91 Corridor here in New England are starting to meet regularly to talk about our common informatics issues. As I mentioned in a previous post ("Can we do better than SOAP?"), we are going to start talking about ways to standardize our documentation on a regional level. I'm interested to hear people's responses because it could be very interesting if, in the next step, we looked at standardizing our definitions, engineering processes, and engineering standards. Would it allow us to share resources that ultimately saved all of our hospitals money, and reduced the cost of operations and care in the entire region?

Stay tuned!

My belief is that we are all both teachers and students our entire lives - So I love to hear people's feedback and thoughts. Feel free to leave comments or questions - Always glad to entertain any new or interesting ideas!