SEEING RESULTS:
Utilizing LEAN and Simulation Modeling to Balance Workflow within an Optical Laboratory
Ed Gensert, Richard L. Roudebush VA Medical Center
Craig Alvis, Richard L. Roudebush VA Medical Center
Abstract
The demand in eyewear for the veteran population has increased beyond current capacity of the laboratory manufacturing these orders. The VISN 11/12 Optical Laboratory has seen a seventy-five percent increase in orders over the last two years. This increase in orders has not been met with an increase in staff or equipment. This unbalanced ratio of workload to staff has lead to longer cycle-times within the Optical Laboratory. To compensate for delaying the orders by up to four times the set standard, the Optical Laboratory has resorted to increasing the number of orders that are outsourced as well as increasing the hours of overtime.This paper will describe the use of Lean Six Sigma techniques to diagnosis and reduce primary sources of constraint within the Optical Laboratory. These techniques were applied to reduce cycle-time and optimize the workflow within the Optical Laboratory.
1.Background
In July, 2010, the VISN 11/12 Optical Laboratory requested that the Indianapolis VAMC Systems Redesign department perform an evaluation of their current processes and make recommendations on potential areas of improvement. This request for evaluation developed due to the projected increase in eyewear orders for FY11. The national mandated turnaround time for a patient’s eyewear is 5 days. The Optical Laboratory’s process to manufacture the eyewear was taking up to 7 days. This turnaround time was aided by outsourcing nearly a third of the orders and working overtime. Employee morale was negatively affected due to the increased pressure to perform.
The project team consisted of the manager of the Optical Laboratory, a lead technician of the Optical Laboratory, an industrial engineering student, and a business student. The team spent four weeks collecting historical data as well as conducting process observations within the Optical Laboratory.
2. Experimental Method
The team followed the VA’s internal, Six Sigma DMAIC Methodology inspired, TAMMCS model in its investigation of the Optical Laboratory. The VA’s TAMMCS model is defined below:
Team/Aim – Team and problem are defined. The purpose and scope of the project team is established.
Map – Process is made visual with the use of Lean tools such as: Process mapping and Spaghetti Diagrams. Potential areas of constraint are identified.
Measure – Data is collected to validate current process through Process Observation Worksheets and establish a baseline before implementing any changes.
Change – Lean and Sig Sigma methods are utilized to provide analysis of data collected during Map and Measure stages. Small changes are made and studied.
Sustain – Process control strategies are utilized to ensure implemented changes are performing within expectation.
1.Team/Aim Phase:
Beginning phase of project in which business case, problem statement, and goal statement are created. The team limited the scope of the project to include only processes once an order for eyewear was received by the Optical Lab to when the eyewear was released to contract shipping firm. The Aim of the project was to reduce turnaround time to meet national standard without requiring more than ten percent outsourcing and reducing overtime by seventy-five percent.
2. Map Phase
A basic process flow map of the Optical Laboratory’s process was created and is shown below in Figure 1. The process map was validated and verified by production staff of the Optical Laboratory. The process map encompasses the entire production process once an order is received until the order has been declared shipped. The production staff was asked to identify perceived areas of constraint. The staff’s perceived areas of constraint were compared with the initial data collection findings. During the process mapping and process observation stage of the project the team discovered that production staff members are not assigned to a specific workstation, but rather flow with the orders. This was explained to the project team as an effort to help eliminate congestion during the production process. Also noted during process mapping and process observations were the under utilization of two machines: one Generator and one Eclipse Blocker. The under utilization of these two pieces of equipment are a result of lack of staff.
Figure 1: Optical Laboratory Process Flow Map
Figure 2: Optical Laboratory Spaghetti Diagram
Once a basic process map was established, the team was able to define the current state of the process.Current state is defined as the following:
Approved Production Staff = 15 (12 @ 80% utilization + 2 @ 50% utilization +1 @ 30% utilization)
Percent of Orders Outsourced = 33%
Machine layout = current
Process = current
Shift = Monday – Saturday
Unique Orders Received = 350/day
Turnaround time = 6.4 days ±0.6 day
3. Measure Phase
Each step of the process map was observed and its process time recorded. Baseline data collected through process observations was analyzed to determine areas of constraint within the process. The team used Pareto charts to prioritize areas of focus. This is shown in Figure 3 below. This data was compared to historical data collected through automated reports generated by the manufacturing equipment. In both data sets the same primary source of constraint was identified as the Edging equipment. The Edging process step accounted for over seventeen percent of the total cycle time. The constraint at the Edging process step was drastic enough that the two steps prior to Edging, First Inspection and Second Blocker, were able to operate at fifty percent utilization and still maintain enough production to sustain the Edging process.
Figure 3: Average Task Time by Process Step
Figure 4: Sample Process Observation Worksheet
3. Results
1. Change Phase
Due to limitations associated with both purchasing machinery and hiring of staff, the team determined the most cost efficient method of testing potential changes would be the utilization of simulation modeling software. The project team created a simulation model in ProcessModel.A Current State model was developed based on metrics collected during process observations. Current State process model results were validated by automated production reports and historical data collected directly from production machinery. The simulation model was set up with 1440 hour (60 days) warm up time and 336 hours (14 days) run time. Process times were set up with triangular distribution. Each simulated scenario was replicated 30 times to increase the statistical significance of output results.
Figure 5: Snapshot of Simulation Model
Current State results are shown below in Figure 6. The results confirmed the team’s initial findings of an unbalanced flow of work. The Edging step of the process directly affects the flow of orders throughout the Optical Laboratory.
Figure 6: Resource States – Current State
The Aim of the project was to reduce percent of outsourced orders to ten percent, production hours worked to standard Monday through Friday, and turnaround time to national standard of five days. The project team set up asimulation model scenario, “What If”, to reflect current staffing and machinery levels performing within the desired structure of ten percent outsourced orders and no overtime. Figure 7 below shows results of the “What If” scenario. This scenario depicts a new primary constraint located at the Generator step. The “What If” scenario results in a turnaround time of 29.7 days ± 0.4 day.
Figure 7: Resource States - “What if” Scenario
Deviations were made to the “What If” scenarios in an attempt to achieve the project team’s desired Aim. Each of the variations applied focused on Lean ideologies with intent on eliminating waste.Variations such as: removing manual data entry, implementing use of pre-cut lenses for Single Vision prescriptions, and automation of the inspection process were trialed within the simulation model. None of these variations resulted in desired output of less than five day turnaround time.
The project team shifted its focus to reducing the amount of movement by the production staff. This was achieved by increasing the production staffing level by four full time employee equivalents (FTEE). The project team added staffing to areas previously determined to be areas of constraint, Edging steps and Eclipse Blocker, and an unutilized Generator. An increase in workforce and utilization of all available equipment resulted in a relatively balanced flow of work. See Figure 8 below.
Figure 8: Resource States – Future State
By increasing resources to primary areas of constraint, the team was able to achieve a turnaround time of less than 5 days without outsourcing more than 10% of the orders and without the use of overtime. The following depicts Future State settings and the results of the simulation model.
Future State
Production Staffing = 19 (18 @ 80% utilization + 1 @ 30% utilization)
Percent of Orders Outsourced = 10%
Machine layout = current + 1 generator and +1 eclipse blocker*
Process = current
Shift = Monday – Friday
Unique Orders Received = 350/day
Turnaround time = 0.9 days ±0.1 day
* Does not imply purchasing new machines, but utilizing current equipment.
Figure 9 shown below depicts time savings for each process step from Current State to Scenario 4. Increased staffing provided the most notable time savings gain in the Edging queue.
Figure 9: Average Task Time of Each Process Step
4. Conclusion
Appling Lean Six Sigma methodologies allowed the project team to validate the need for additional Optical Laboratory staffing. Utilizing these tools provided the project team with quantitative metrics for which the affects of additional staff could be measured and compared. The results of this project were presented to the VISN 11/12 Optical Laboratory Steering Committee from which they were able to grant the addition of four FTEE.
5. Biographical Sketch
Edward Gensert, Bachelor of Industrial Engineering, Bradley University, Peoria IL, Industrial Engineering Honors Society President(APM)2009-2010
Craig Alvis, Associate of Business Science, Indiana Wesleyan University, Marion, IN.