Process redesign in the clinical laboratory: A success story.
Our veterans' administration hospital, like many other hospitals, has been in the process of transitioning from an inpatient facility with a few outpatient clinics to a hospital with far fewer beds and a strong outpatient focus. Ten years ago, our laboratory test mix was 25% outpatient and 75% inpatient. Now those figures are reversed.
Before the process redesign
Although we believed we were meeting the needs of our providers and veterans, we received many complaints about turnaround and patient wait times. Our waiting rooms and halls were full of patients waiting for phlebotomy services. One clerk was assigned to processing all patients. Patients typically arrived for phlebotomy, completed a form with their name and social security number, and then sat in a crowded waiting room, only to be told an hour later to return to their clinic to retrieve missing lab orders.
We blamed the medical administration service for the absence of orders, but it was the patients who suffered, regardless of whom was at fault. We were stuck in a mode of providing results in about 4 hours when our providers needed the information immediately. Every delay reduced the capacity of the clinic to schedule and see more patients. Surveys (Press Ganey) indicated strong patient dissatisfaction with our wait times. Several thousand veterans were unable to access our healthcare system because our service could not keep pace with demand. We were so busy putting out fires that we didn't have time to fix the system. We had done studies and collected data, but the results and implications were disheartening, so we stopped. We needed significant reductions in cycle time, measurable improvement in customer satisfaction, and improvement in our quality indicators to better provide for our veterans, yet we had extremely limited resources for additional staff and equipment.
Process redesign: Planning and preparation
Our new hospital director recognized that standard measures for improvement were not working and introduced a revolutionary system called process redesign. The initial focus was on the primary care area because of its broad reach. An outside consultant introduced us to the lean production approach pioneered by Toyota, which dictates that an organization should do only those things that create or add value; all other activities should be eliminated.
The first step was to bring all employees on board. Numerous 15-minute sessions were held with all staff members to convey these new ideas. We learned about the concept of a "value stream," which is a picture of the flow of a process, including value-added (contributing to a service or product) and nonvalue components. Nonvalue elements of the process are also called "muda," a term coined by Taiichi Ohno (1912-1990), which means "futility" or "uselessness." Muda, practically speaking, is "waste" and includes all activities that absorb resources without creating value, such as mistakes that require correction, unnecessary movement of people or goods, and idleness of workers because a product is not ready for them to work with.
Our next task was to outline all the steps that lead to our "product" (ie, an available test result). We labeled each step as value-added or as muda. Anything labeled muda required a list of possible causes. Our value stream for a patient's lab test went something like this:
* At 8:12, Mr. Patient goes to the registration desk on the first floor and waits 15 minutes to be signed into the hospital system.
* At 8:27, he walks to the elevator and rides it down to the lab on the lower level.
* At 8:32, he arrives at the laboratory phlebotomy desk and hands the clerk the paper he received previously. He takes a seat and waits.
* At 8:45, the clerk takes the paper from the pile and puts it on the scheduling board.
* At 8:55, the phlebotomist walks to the front of the phlebotomy area and calls the patient's name.
* At 8:58, she draws his blood.
* At 9:05, the specimens are sent to the processing area.
* At 9:15, the specimens are registered in the proper lab areas.
* At 9:20, the results are available in the hospital computer.
Because we considered all wait and travel time to be muda, the value stream made evident that our entire process needed revamping.
While the need for change was clear, we had to reassure the staff that this examination of the process was designed to increase productivity and stabilize employment--not eliminate jobs or people. We emphasized that we sought to make each employee work better--not faster--so that the hospital could meet its goal of providing better service to its veterans.
One important step in the process was to eliminate the "batch and queue system." Large hospital laboratories used the batch system for years because of presumed efficiency, but our consultant had us perform a simple exercise that forever dispelled the idea that batching is better. We divided into groups of four. Each member of each foursome had a task. Member #1 would take a piece of paper, write a name on it, fold it, and hand it to member #2. Member #2 would unfold the paper, add a signature, refold the paper, and hand it to member #3. Member #3 would review the paper, stamp it, and hand it to member #4. Member #4 would fold the paper and put it in an envelope. The process was timed and repeated with variations. The fastest time was achieved when each member completed the entire procedure him or herself.
A laboratory strike team was developed to launch a complete redesign of the workflow of laboratory services, including phlebotomy, processing, hematology, urinalysis, and chemistry. Lab supervisors and the consultant chose team members based on knowledge of their particular area and willingness to work toward innovative solutions. The team ultimately had 9 members including the hospital director, the union president, a primary care nurse, the laboratory administrator, and a supervisor from engineering. Participation by the hospital director sent the message to all involved that this project had management's complete support.
Cycle time, percentage of time wasted, and minutes wasted were measured. Specific targets for improvement were defined. If the target for a CBC result was 30 minutes and 1 patient's test took 15 minutes while another's took 45 minutes, the latter was unacceptable because the 45-minute test would create a delay for the provider that could back up patients for the rest of the day.
Because the process redesign system is objective and data-driven, the team had to develop and implement measuring devices to facilitate continuous assessment of the entire process. The team was instructed to establish an operating system--a framework to analyze, improve, and monitor processes--and to attack rapidly and aggressively each process step to make improvements. The team, including the hospital director, actually worked benches in the laboratory to become completely familiar with the existing system and its inherent weaknesses.
The team was also trained to recognize and tackle obstacles. Employees were trained to identify, verify, and analyze root causes of muda, which is accomplished by asking "why" 5 times and then asking "why" 5 times more. They were trained to design cells, or work areas with machines in close proximity to each other to minimize movement and encourage multitasking. They developed standard work instructions--detailed descriptions of positions essential to the process--and distributed them to the staff for feedback. These standard work instructions incorporated contingency plans for computer downtime; staffing shortages because of breaks, lunches, and vacations; and equipment problems. They outlined "footprints" on floors where machines and supplies belonged. Tape outlines on countertops defined the location of staplers, scissors, calculators, and other necessary tools so that absolutely no time would be wasted looking for these essential items. Furniture, equipment, signs, and other items that were not essential to the process were relocated. Lab staff was encouraged to make recommendations throughout the process because once the standard work instructions were in place, no deviations were permitted.
Process redesign: Implementation
Once the plans were complete, it was time for implementation, which was scheduled for the week of April 17, 2000. Overnight, the engineering department moved counters, equipment, power lines, phone lines, computers, and plumbing. The hospital director made support services available to work as much overtime as required to accommodate the strike team's recommendations. Additional staff members were hired to help in the processing area because one of our major problems was that a hospital phlebotomy station was located some distance away, and we had no dependable method of transporting specimens delivered to the main lab. A clerk was added to our staff to find doctor's orders that had not been transferred to the hospital computer. We also received extra help for centrifuging and delivering specimens to the laboratory sections for testing.
All steps were monitored, and all data was analyzed. We scanned barcodes of incoming specimens to record transport times, and we even scanned the specimen runner's barcode to ensure timely delivery of specimens (115 trips per day!). We captured instrument test times on our analyzers daily and were able to identify where we needed to concentrate on eliminating wasted time.
To illustrate how our new system works, Mr. Patient now arrives at our laboratory, takes a number, and signs his name and his social security number on a piece of paper, which he hands to our new clerk. She finds his orders in the computer, or makes the appropriate telephone calls to obtain the orders. Within 2 minutes, a phlebotomist calls his name, accessions his orders, prints barcode labels for his tubes, and takes his blood. He is then on his way to his clinic appointment. Our runner makes a stop at all phlebotomy stations every 5 minutes to pick up specimens and deliver them to the processing site, where they are scanned into the computer, centrifuged if necessary, and delivered to the appropriate lab section.
Reaping the rewards
The results of this week long process were astonishing. The average CBC cycle time, counted from when the order is accessioned to when the results are available in the computer, decreased from 73 to 36 minutes (see Figure 1), and by the end of June approximately 80% of CBC panels run met our target TAT (see Figure 2). The time to get an average urinalysis (UA) test result went from 140 to 92 minutes (see Figure 3), and by the end of June about 50% of all UAs performed met our target TAT (see Figure 4). Finally, the average BMP (basic metabolic panel, chemistry) TAT decreased from 125 to 61 minutes (see Figure 5), and the percentage of BMPs run that meet our target TAT is slowly but surely increasing (see Figure 6).
Further decreases have materialized as the process continues. CBC TAT is currently 28 minutes, BMP TAT is 57 minutes, and UA TAT is 35 minutes. Staffing, equipment, and computer problems do occur and temporarily increase cycle times slightly, but never to the levels experienced previously.
Providers are pleased that laboratory results are ready when they need them, and patients are surprised at the short wait times. The laboratory strike team continues to meet every 2 weeks to review data, make changes when necessary, and monitor and maintain progress. Data and results are posted daily so that staff members remain informed and involved. One of the most important changes turned out to be the elimination of batching; continuous flow has been the key to success.
Our strike team members are dedicated employees who worked many extra hours to make this process redesign happen. They read literature, collected and studied data, and worked to find ways to work smarter (not harder). They also worked to enlist the support and cooperation of coworkers and supervisors. Their insight, expertise, and hard work are what made this project a success.
The process redesign project will now move to include other areas of the hospital, but we are proud and happy to serve as a shining example of how successful the project can be.
Linda Wright is a supervisor of hematology, and Donna Ferguson is a supervisor of the evening and night shift, at the Pathology and Laboratory Medicine Service, North Florida South Georgia Veterans Health System, Gainesville, FL.
Fritz-Unrue R. 2000 Systems, Inc.: An Operating System to Analyze, Monitor, and Improve Processes. In press.
Fisher K. Lab reduces processing dine by 50%. The Flag Newsletter. 2000;2:8-9.
Womack JP, Jones DT. Lean Thinking. New York, NY: Simon and Schuster; 1996.
|Printer friendly Cite/link Email Feedback|
|Author:||Wright, Linda; Ferguson, Donna|
|Publication:||Medical Laboratory Observer|
|Date:||Jan 1, 2001|
|Previous Article:||Nematodes: Worms of the world.|
|Next Article:||All in the family.|