College of American Pathologists
Printable Version

  Brewing a high-octane laboratory, step by step






August 2007
Feature Story

At Lab InfoTech Summit 2007 in March, Steven H. Mandell, MD, shared the experience of the laboratory at the University of Michigan in implementing Lean and Six Sigma. Dr. Mandell is assistant professor of pathology and director of MLabs, University of Michigan Medical Center, Ann Arbor. An edited version of his presentation, “Polishing Lab Core Functions with Lean and Six Sigma Tools,” follows.

The auto industry has been putting pressure on us in southeastern Michigan to change the practice of health care, saying, “We know the services you are offering us are not as efficient as they can be.” They know Lean and Six Sigma offer significant advantages in quality, service, and efficiency within manufacturing environments, and they are learning that these same techniques can be applied to health care. These corporations influence our boards of directors and are stimulating program development of Lean initiatives at our hospitals. As lab leaders, we should not have to wait for our hospital administration to turn to us and say, “We need lean labs as well.” We must proactively convince them that Lean and Six Sigma are the proper strategies for our laboratory and that funding these initiatives is financially sound. How can we do this?

First, you must find your burning platform—a compelling reason that convinces administrators that dramatic change is needed. At our campus, the opening of a cardiovascular hospital was a few months away but we already had our clinical laboratory working at capacity. No provisions had been made for the increased workload our clinical laboratories could expect as a result of this expanded program. We asked for resources to be directed to Lean efficiency initiatives for our core laboratory operations rather than additional staffing or laboratory space.

When you talk to your administration about getting Lean, it’s important that they don’t think lean and mean. We’re talking about improving processes and efficiency, and nobody is losing their job in this process. If you start cutting full-time employees in implementing Lean and Six Sigma, your buy-in will fail miserably and you will sacrifice any long-term gains you hope to achieve.

Lean focuses on eliminating waste within processes, not eliminating people. Six Sigma, in contrast, focuses on reducing variability and eliminating errors within processes. While they are different disciplines, they have some things in common. Both are focused on what is important to the customer—in our case, the patient. Both base decisions on facts and data rather than opinion. Both help transform difficult problems into manageable ones; they create low-hanging fruit out of high-hanging fruit. And both programs increase employee satisfaction and empower employees in a number of ways.

First, these programs are owned by the employees who do the work. It provides them with a disciplined approach to problem solving. Many of us are effective at brainstorming, but this allows us to take that next step and make it rocket-powered, if you will. It teaches employees to see waste, savings, and opportunities, and it eliminates staffing shortages and overwork. These, combined over time, make for a much happier work environment.

In becoming lean, you can’t just apply the tools of Lean; you also need to adopt the philosophies to ensure success. Lean tools help to change processes. Lean philosophies, however, help to change people within a culture to become identifiers of waste and practitioners of continuous improvement. Change—even change for the better—can be a disturbing thing to some folks. Managing the human aspects of change is the most difficult challenge to implementing Lean.

The tools we used to identify waste and rebuild processes to the proper specifications can be described with a few examples. The best, global concept is found in the parable of a glass partly filled with liquid as being half-empty or half-full depending on one’s point of view. In process engineering terms, neither is true. What is true instead is that the glass—as it was designed—is exactly twice as large as it needs to be to contain the amount of liquid we need it to hold. The questions I have for you are, within your laboratory environments, how many of your processes are the wrong sizes for the work you need to do, and how can you determine where that misfit may be? That’s what Lean tools are all about. One of the core tools is value stream mapping. To create a map, all of the stakeholders within a process, from raw materials to the end product, need to participate. So if we’re talking about a lab specimen, participants should include the physician who’s placing the order, the physician (or even the patient) who’s getting the result, the phlebotomist, the lab assistant receiving the specimen, the technologist who’s running the test, the IT folks who manage the LIS, and so on.

I like to think of value stream mapping as process mapping on steroids. Here’s an example: Special stains at a site have been delayed and they’re arriving in an untimely fashion. To map this process, the first step is to identify the customer and define the customer’s desires. The customers here are the pathologists, and they want their special stains done the same day if they were ordered before 2 PM. We are often unaware of, or ignore, what the customer is asking for—we assume we know so we don’t ask. In most examples, it is critical to have the customer in the room, even if that customer is a patient.

Next we identify who is supplying the raw materials. As in many cases, here the customer and the supplier are the same: a pathologist who fills out a Post-it note with a special stain request and hands it to his secretary.

Once we have identified the general steps at each point—what is done and who does it—we add the steroids. And this is what differentiates a value stream map from a process map: We begin to make what had once been invisible in our processes now visible. The things we look at are not only the actual processes themselves but also what is between the processes.

Where are things waiting to happen? These are called inventory steps. What technologies are being used, and how reliable are they at each step in this process? How are tools being used, and are they being used effectively as people perform their jobs? Where is information flowing and is it where it needs to be, how it needs to be, when it needs to be there?

We add additional data for each of our processes. “PT” refers to process time. How long does it take to complete that task? “WT” is wait time. If someone gets interrupted or the task can’t be completed, how much waiting is there at that particular point? “%C&A” stands for “percent complete and accurate.” Some people refer to this as “FTQ,” or “first time quality.” This is how many times out of 100 this activity can be done from beginning to end without encountering a problem or an error.

We can then take these data and describe our components numerically. The process from order to getting a stain out could take about 43 minutes if we did it all in sequence and uninterrupted. But within this example we see there might be wait time that lasts between four and 69 hours. The accuracy rate of this entire process, if you multiplied out the accuracy at each component step, is about 75 percent. Moreover, we look at how often the customers’ demands are met and see that we’re meeting our customers’ expectations only 80 percent of the time. In this case, that would be a failing grade.

We quantify both the value-added and non-value-added components of our process. Value-added is anything that transforms the shape, form, or function of a material and information into parts or products. It’s something the customer, the patient in this case, is willing to pay for. Non-value-added, in contrast, consumes resources but does not contribute directly to making the product.

Once we develop a value stream map, we then challenge the stakeholders to create a new process that eliminates waste in and between steps. Rather than thinking of waste as an ill-defined concept, we counsel them to break it down into the components identified in the Toyota Production System: re-work done to fix errors; overproduction—making more than is necessary or more than the customer requires; excessive or unnecessary motions of people—by travel, walking, or searching for materials or items; material movement—unnecessary handoffs, transfers, or movements of material and/or information; waiting—including people waiting for machines or information or, conversely, information or material waiting in a queue for a machine to process it; and overprocessing—performing redundant or unnecessary work. An example of this last item is filling out a paper requisition, followed by transcribing that requisition into our LIS. There may be downstream benefits to our billing department, but this act is really a form of duplicate order entry—better to place an order once and only once.

We developed a value stream map for our phlebotomy sweeps where specimens are collected bedside, then taken to our central processing area, which then passes them to our chemistry area. Applying Lean concepts to reduce waste, we’ve identified opportunities to achieve a 57 percent reduction in total throughput time. At the same time, we reduced our human and specimen wait times and distances traveled to perform the same work we were doing before. We already had pretty efficient in-lab processes, so much of the waste we identified was in the transit of specimens to the lab.

This presentation is being made at the Infotech Summit to inform you that software support for Lean and Six Sigma activities is readily available but not being marketed to health care. Instead, these products have been limited to manufacturing industries. My advice: Steal shamelessly. Investigate Lean and Six Sigma in manufacturing industries and use their software, because it’s going to perform very well in your environment.

One tool we use is called “eVSM,” which stands for “evaluating the value stream map.” This is based on the Microsoft Visio product. This product takes the data within your Visio value stream map and exports it to an Excel spreadsheet to analyze and create graphic presentations.

Another tool we’re evaluating is Timer Pro, which allows you to do value stream mapping in digital movies that, as you play back the images, permit you to analyze and record different activities as they’re going on in real time. The beauty of this is not only does it collect the time and activity data for you, but it also allows you to edit waste out of your processes simply by editing wasteful activities out of the movie. As each video segment is also captured in numerical and graphical format, it enables you to re-edit your desired practices simply by manipulating their graphic or data components.

U of M has a relatively complex lab, processing about 6 million specimens a year. Over time we’ve developed silos of expertise within our laboratory to the point that we have very little cross-coverage between lab sections. One of the things we’ll need to look at in the future is consolidating workstations—like automation that exists separately now in chemistry and hematology. We are not there now. But to be able to model process changes virtually by dragging and dropping activities back and forth to different environments to plan what a more efficient future state might look like is not only powerful but easy—and a tremendous training advantage. Timer Pro has software out on PDAs, so process measurements can be taken on the fly wherever you are in the lab.

Several other companies are doing similar things, and the technologies can be relatively inexpensive—generally a couple hundred bucks.

We have a new hospital coming online, and knowing it’s going to take us time to train ourselves on these tools, we asked ourselves, is this something we could do effectively in the time frame we had? To get Lean and Six Sigma implemented within our clinical laboratories would have taken between three and five years and we only had a few months. We recognized we needed help, so we shopped for consultants to help us with the implementation. At the same time, to secure funding for the consultants, we needed to know how much this was going to cost and what the return on investment would be. So we sought help to assess our opportunities and determine if we could justify spending to get lean in the time frame we needed to.

In assessing our lab, our consultants looked at four major areas—I refer to these as “the four questions.” First, is standard work occurring? That is, are processes standardized to best practices, and are the practices you’ve developed based on actual data?

Second, are unnecessary activities taking place in the testing process? This is an identification of process wastes.

Third, they looked at the operators: Are there excessive motions? And if there are activities going on with operators, are they purposeful and properly applied at the right place at the right time?

Finally, they looked at our inventory supply and management to ask, is it excessive, and if it is, how much space is it taking and what are the costs of maintaining that inventory and space?

So after looking at these four items, the consultants said, “University of Michigan, you are a wonderful lab, but you have some areas for improvement.” One of the things you have to keep in mind is that upon hearing this, a lot of people will be resistant. When somebody tells you, “You know what? You’re messing up and you need to change,” your initial reaction is to become defensive. Part of the Lean philosophy, though, is to celebrate finding things you’re messing up because these are truly opportunities for improvement. If we don’t know about them, we can’t fix them.

I’ll give you examples of what we found. The first one is a spaghetti diagram from our histology laboratory. We followed one operator around the laboratory and in only 40 minutes or so, in this pretty confined space, this employee covered a circuitous route of about 530 feet. Not only was the operator traveling an excessive amount, but she had several barriers to her progress—chairs, counters, people. We all have this going on in our laboratories to some degree. We just didn’t know before how to use a spaghetti diagram to see it or measure it.

Another opportunity for us was batch processing. In our lab, there are sorting, accession, and centrifuge stations, and finally a transporter that takes the specimens to the testing laboratory. Specimens arrive in batches. Each specimen takes about a minute to process, but you can’t pass them to the next station until you’re done with your entire batch, so getting a batch through one station can take around 20 minutes. The first unit in a batch gets passed at about 20 minutes and the last unit in that batch also takes 20 minutes.

Batching specimens increases the time that the first specimen touched has to wait in inventory and increases the bench storage space we need to perform processing. In processing large batches, problems become less apparent. Problems are more apparent in handling one specimen at a time. If we’re dealing with things in batches and something goes wrong, like our centrifuge breaks down, that means we have problems with maybe 50 specimens at once, as opposed to just one at a time.

By converting to a single-piece-flow operation, each specimen still takes a minute to process at each station. But once that person is done with that specimen, they then pass it on to the next operator. There is no confusing multitasking; rather, single tasks occur in sequence. The next person in a sequence no longer waits for a batch to arrive at her station; she becomes active relatively quickly in this process as the first specimen is passed to her station.

In the previous process the batch that took 20 minutes can now be done in eight minutes using single-piece flow. Better yet, that first specimen, rather than coming out in 20 minutes, now comes out in four minutes. Each new specimen comes out a minute later. So within a flow process you begin to sense a smooth rhythm to the work being performed rather than the big waves we were experiencing before.

Another problem we’re addressing is our excessive, disorganized inventory and supplies. We have materials and supplies that may have been in drawers or cabinets for two or three years and we didn’t know it. These are dollars sitting in our drawers and cabinets. We had no first-in, first-out order of inventory in some of our operating areas. I spoke recently to the Michigan Society of Histotechnologists, which represents about 90 different sites across Michigan. I asked them, “Before your past CAP inspections, how many of you would go through your laboratory to try to clean out everything that was messy and eventually find expired reagents?” Every hand in the place went up. In January, CAP went to unannounced inspections. So I asked them, “How are you going to deal with this situation in the future?” And you know what their answer was? “They’re going to catch us.” They had no solution.

Well, the solution is Lean applied to inventory management. Getting organized is part of the Lean concept. The way you can re-supply your inventory and keep it organized, and use a first-in, first-out technology without passing expiration dates, is by using something called Kanban systems or signals. Signal cards might sit in the middle of a test-tube supply rack. As the amount of test tubes dwindle when phlebotomy carts are stocked, if you reach the card within that rack, you pull the card out. Then, as you’re walking out of the room, you place the card in the manager’s cardholder. Now the manager knows what needs to be replenished and later in the afternoon places that order. Signals like this can be used in large warehouses or in very small operations like our phlebotomy cart stocking area.

The consultants also said our workstations were cluttered, and that presented a risk for errors. There is a Lean solution for this too, using what we call “visual factory” and “5S.” 5S comes from five Japanese words that are interpreted to: “Sort, Set in Order, Shine, Standardize, and Sustain.” The way this works is you ask, “What do you need to do your work today, or during your shift?” Get those items only and get everything else out. Once you have the things you need to work on your shift, what sequence do you need them in and where do you need them to be organized ergonomically? Let’s put your tools in those places and set up a strategy to keep them there and always be there for you when you start your shift, your work for that day. This concept can be summarized as: “Shelves without doors, nothing in drawers, a place for everything, and everything in its place.” And the visual component is “Don’t just know where it goes, but show where it goes.”

What is Lean going to mean for your institution?

I want to tell you what our preimplementation assessment showed, and our findings are pretty typical for any laboratory environment that has implemented Lean. Lean allows you to reduce throughput time for testing by between 30 and 50 percent. It allows you to increase your productivity by 30 to 40 percent. It reduces your space required for testing by greater than 20 percent; we expect this may be upward of 40 percent for parts of our laboratory. It reduces your supply inventory 40 to 50 percent.

For our institution, just in our core lab facility—chemistry group, hematology group, central processing, and phlebotomy—we expect to eliminate overtime expenses for those areas to the tune of about $400,000. So we were a well-managed, good laboratory, but we had room to grow that we didn’t even know about until we learned to see opportunity, waste, and savings within our operation by using Lean techniques.

We talked to our consultants and said, “We need to do this; we don’t have a choice. How much is this going to cost us?” Their answer: several hundred thousand dollars. But here is how it broke down: Half of the figure was going to construction to move those benches that got in the way, put equipment closer together, cut out the travel distance between places. Some of this money went to the consultants. And some of it went to replenishing FTEs; they said, “We’re going to take six FTEs out of your laboratory and make them Lean experts so that in about four or five months, you’re going to have embedded Lean expertise within your own environment. We will go away, but you’re going to need to backfill these FTEs while they are in training.” So there’s about $100,000 in backfill of FTEs. Still they said, “No worries. You’ll get a return on investment within 11 months of us completing our project.” Not bad. I think any of your administrators would be pleased with that.

Process improvements will allow us to re-deploy six FTEs. We’re not going to be laying off anyone—that is a key tenet of our Lean implementation—or we would lose credibility and momentum. In our old processes, we suffered significant opportunity costs; there’s a whole bunch of things we would love to do within our laboratory environment that we have not had the space or personnel to do. Now, by creating this space and freeing up FTEs, we can address the initiatives required to maintain our role as a leader in academic medical center care and research.

What about the long term? The hospital wants our laboratory location for clinical use, and we’ve outgrown it, so we’re planning a new building. It will be ready in about 2012, and if you implement Lean design or design for Six Sigma within a new facility, you can save between 20 and 30 percent of your initial construction cost. If you’re planning an $80 million building, the investment in Lean looks quite paltry compared to the savings of millions of dollars down the road.

So my take-home messages to you are the following:

arrow Lean and Six Sigma are no longer optional. They are mandatory because of the gains we can achieve and because of the demands being made upon us.

arrow The tools themselves work, but leadership and dealing with the human responses to change are going to be the keys to your success.

arrow Do not automate bad or unnecessary processes. Do not put in automation thinking it’s going to solve all your problems. Get Lean first and then automate the processes that are routine and repeatable.

arrow Consultants may help smooth and shorten the time frame to a Lean transformation of your laboratory.

arrow Specimen flow, value stream mapping, 5S, standardization, and root-cause analysis form the foundations of any of the improvements you’re going to undertake.

arrow This is a continuous learning process, so learn as you go. Develop a model line and investigate the tools and software as you’re learning this process to see what works for you.

arrow Implementing Lean may indeed be costly initially, but you need to begin thinking long term, because this will come back and reward you financially. A good idea is to begin these projects slowly and let the success of each project pay for the next one.


Related Links