Connect Process Flows to Become Lean
Careful mapping of the process and a consistent emphasis on the voice of the customer enable lean production of an external heart defibrillator/cardiac monitor
For more than 55 years, Physio-Control Inc. (Redmond, WA), a unit of Medtronic Inc. (Fridley, MN), has been a manufacturer of portable external heart defibrillator/cardiac monitors, which are considered Class III medical devices. Its Lifepak brand is known for quality and rugged design, and is one of the most-recognized brands in medical devices.
The medical industry is generally known for its stringent quality control and reliability processes, but continuous pressure from global competition and the recession have required companies to keep improving cost structures, while also focusing on customer needs and delivering the best patient care. These factors led Physio-Control Operations Engineering to undertake a project to improve manufacturing processes by looking at the current methods used to manufacture Lifepak 20 and Lifepak 20e defibrillator/monitors, with the goal of reducing variability in meeting customer order lead times.
The Lifepak 20/20e monitor is designed to be used in hospitals by Basic Life Saving (BLS) and Advanced Life Saving (ALS) personnel. These systems are offered in multiple configurations, including basic monitoring and defibrillation, and additional advanced monitoring features—such as noninvasive pacing, ECG monitoring, blood oxygen levels, or pulse oximetry. Each of these hardware configurations comes with literature kits in more than 16 languages. This wide range of offerings allows us to meet the specific needs of our global customers, but requires a manufacturing process that can support a low-volume/high-mix set of customer requirements.
The defibrillator/monitor is manufactured on a mixed-model line that completes the assembly, test, and configuration operations to manufacture a device. Production areas are designated as unit subassembly, configuration, accessory-kit assembly, and shipping.
The first process is the unit subassembly line, which assembles and tests the six generic hardware options offered. After each subassembly is completed, it’s sent to the configuration line, which installs the country and language-specific components, and software that enables the unit to display readings, offer guidance, and deliver audio prompts specified by the customer. Finally, the configured device is moved to the shipping area, where it’s packaged with accessory kits assembled in the kit area.
The first step in the improvement project was to create a baseline value stream map (VSM). From analysis of the VSM, we realized that the voice of the customer (VOC) was not well-defined, and was broken into many unconnected segments. The obvious indicator of this issue was that, even with large amounts of inventory between processes, downstream operations did not have the correct mix of parts to complete customer orders.
One of the most serious outcomes of this issue was that the configuration line had to wait for completion of specific unit subassemblies to complete a sales order, which prevented orders from moving through our shipping line and being invoiced. The accessory kit process presented similar challenges. We had days of inventory between it and the packing line, while we were still scrambling to make the "correct" accessory kits needed for the next sales order. Although every unit shipped with more than one accessory kit, it was the literature kits that caused the biggest problem, because there are multiple versions of the kits available, depending on the specified language.
Now that we had a current-state VSM that identified the areas that had problems, it was time to fix them. We organized a series of shop-floor Kaizen events using cross-functional teams of production and support-staff personnel to refine the issues in each area and come up with a solution.
The Kaizen methodology we used is based on the traditional Plan-Do-Check-Act Cycle (PDCA) that was scaled to meet the scope of the challenge. The basic recipe we used was: three parts Plan, one part Do, two parts Check, and one part Act.
Once we had identified that the four process flows were unconnected, the first Kaizen event effort focused on how to link them to the voice of the customer. The team quickly identified the need to create a pacemaker that would set the pace and sequence of production for the whole manufacturing process based on the VOC. It was decided that the configuration line would make the best pacemaker, as that’s the location closest to the customer, and the last place we configure a unit to the customer specification. Setting just one location as the pacemaker would prevent the other areas from becoming misaligned from actual customer needs.
To translate the VOC to the pacemaker, the team designed a build-card system to signal the configuration line, the quantity, and the sequence of devices to configure based on the build plan generated by the sales-order planner. The individual build cards contain all the information the team members needed to pick the correct subassembly from inventory, and what sales order to configure the unit to. In addition, the card was designed to travel with the configured unit, and told the shipping line to which sales order the unit belonged. This information was essential—it greatly reduced the time and effort required by the shipping team to track the progress of shipping sales orders. Now that the pacemaker was established, it was time to connect the processes to the unit subassembly and accessory kits.
Based on the success achieved by implementing the build-card replenishment system on the configuration line, the team developed a second build-card system between the pacemaker and the first station of the unit subassembly line. The build card was initiated when a unit subassembly was removed from the between-process inventory, or "supermarket," by the configuration line. The build card signaled the unit subassembly line what version of subassembly to assemble to replenish the supermarket. This removed a number of planning interactions that the production support had to perform, to keep the correct mix of unit subassemblies flowing through the system.
One obstacle in setting up the replenishment system was determining the design of the supermarket that was required to meet the demands of the pacemaker. The goal was to have a replenishment cycle that could rapidly meet the need of the pacemaker, but do it at the minimum inventory level. The inventory aspect was the biggest challenge due to the large difference in cycle time between the unit subassembly process and the configuration process. Our team reviewed two different supermarket designs: pull and sequential pull.
The advantage of the pull-style supermarket was its responsiveness to changes in customer demand, because it had all parts available in quantities that would cover the replenishment time needed by the subassembly process. It also required small amounts of production planning. On the other hand, a pull style requires high levels of inventory, which ties up money in inventory.
At the other end of the spectrum, a sequential-pull system, which builds product in the exact sequence needed, would allow the supermarket to carry a minimum of inventory. The disadvantage is that it would break down when there were changes in customer demand, or if the subassembly needed to be sent to troubleshooting or repair during testing (steps that require large amounts of production planning).
Based on the advantages and disadvantages of the two designs and analysis of the historical and projected usage rates, the team determined the best approach was to establish a hybrid of pull and sequential pull. The high-usage subassemblies were set up as pull, while the lower-use subassemblies were set up as sequential pull. To add stability to the sequential-pull parts, a small buffer inventory was established to lessen the effect of subassemblies falling out in test, or changes in customer demand.
The accessory-kit area presented a couple of different challenges besides just being aligned to the pacemaker. One of the biggest was the large number of different kits it had to produce.
There are two styles of accessory kits produced in the kit area. The first is a hardware kit, which is a collection of cables, electrodes, and sensors. The second is a literature kit, which is made of a collection of localized operational instructions, and other device-setup documentation. In hardware kits alone, there were many different versions, which grew even further when adding in the number of literature kits. In addition, the shipping line would be consuming two to four kits every six minutes. For the kits area to be efficient, workers had to be accurate in producing the kits in the correct mix, but also do so in a manufacturing time under two minutes.
To create the optimum volume of accessory kits, the Kaizen team analyzed the process using standard work sheets (spaghetti diagrams) and percent-load charts to create a baseline for the kit-building process. The team found time inefficiencies due to the operators having to get up and find items required to build the kits. As a result, the team created a new layout for the area that brought most of the items within arms reach of the operator. When the new layout was tested, the team found that the operator was more consistent, and was less prone to making assembly errors. This outcome led to a cycle-time improvement by removing a second point of verification that had been previously put into place.
In addition to improving the process lead time, quality, and alignment, the new layout and replenishment methods required 10% less floor space, eliminated 800' (244 m) of walking per day, and reduced kit inventory by 26%. This success allowed the team to reduce the height of the flow rack for the transported parts to the shipping line to less than 5' (1.5 m), which made communication more efficient between the accessory kit and shipping-line team members.
To create the correct mix, the team performed an analysis of kit usage and found that the best solution again involved creating a hybrid replenishment system made up of pull and sequential pull. Because the hardware kits had the least number of versions and made up the bulk of the volume, they were set up as a full-pull system between the kit area and the shipping line. To create the build signal for the pull system, an empty Kanban bin was sent back to the kit area. Just like the build cards, the Kanban bins were labeled with all the information necessary to replenish the parts.
A sequential-pull replenishment system was implemented for the literature kits, as they had such a large number of variations available. To signal the literature kits to be built in the correct sequence, the system was connected to the pacemaker through a card system. Again, the card had all the information the kit area needed to create the work orders, and would travel with the kits to the shipping line.
The original manufacturing process used a number of computer reports to draw constant attention by the production support groups. These reports ensured the areas were aligned to build and configure the product in the correct order and qualities. This procedure consumed a considerable amount of time and energy, and the information was outdated the minute the report was printed. By implementing visual replenishment systems, everyone who works in the process, not just the individuals who have access to the reports, knows what to build and when. This visibility significantly reduces the number of interactions people needed to perform to keep items aligned, therefore reducing process variability.
We were able to establish standard quantities of inventory between each process, which we call standard work in process or SWIP. The importance of SWIP is that it allows the ship-kit operators and other support personal to identify the existence of abnormal conditions based on whether an excess of empty bins is waiting to be filled. We also installed Andon lights on the configuration and accessory kit area, to signal to the rest of the production area if there was a situation that needed immediate attention. This arrangement was a stark difference to the situation that existed when the operators were just pushing what they thought were the "correct" parts to the shipping line.
Although it was not directly related to improving the Lifepak 20/20e device manufacturing processes, we developed some key tools needed to use lean manufacturing in a Class III medical environment. These tools related to optimizing how we determine and document the impact to our process validations due to layout and process changes. To those who have not worked in the medical-device industry, this may seem trivial, but at times it can take a substantial effort, which causes people to shy away from implementing quick-moving lean manufacturing techniques.
Using lean tools and implementing additional lean manufacturing techniques within our manufacturing processes, we created a production system with reduced process variability, aligned to the VOC, and can detect abnormal conditions before they impact our ability to meet the needs of our customer, thereby improving our ability to meet customer needs.
For more information on Medtronic, go to www.medtronic.com.
Contact Lorin Jovag by e-mail at firstname.lastname@example.org. ME
This article was first published in the May 2011 edition of Manufacturing Engineering magazine. Click here for PDF.