Lean Sigma in the lab

From LabAutopedia

A LabAutopedia invited article

Lean Sigma in the lab
Authored by: Dr. Michael Allen, Lean Sigma Lab

Lean Sigma is a combination of two distinct process improvement techniques. Lean (sometimes called Lean Thinking or Lean Transformation) focuses on eliminating waste and delivering customer value in the shortest possible timescale. Six Sigma focuses on understanding variation (especially of quality of product or service) in a process and driving to progressively higher levels of consistency. Lean Sigma (sometimes called Lean Six Sigma) combines the two techniques to deliver increased quality at an increased speed and at a reduced cost. Simulation allows you to analyse your lab in a virtual environment and test process improvement ideas.

1 A roadmap for Lean Sigma
2 Lean Sigma in the lab

The benefits expected from application of Lean Sigma include faster speed, higher throughputs, increased capacity and productivity, fewer errors and improved utilisation of resources (people, equipment, facilities).

Figure 1.

Lean Thinking is a development of the pioneering work done by Toyota on "Just In Time" manufacturing. Six Sigma is a development of Shewhart's statistical control charts and Demming's "Plan, Do, Check, Act" method of process improvement. Both techniques trace their histories back more than 60 years and both continue to be developed today as well-proven methods of improving quality and efficiency. Many people have found that the best results are achieved when the two techniques are combined. Simulation has been used for almost 30 years for testing ideas and plans in a virtual environment.

A roadmap for Lean Sigma

As Lean and Six Sigma are distinct methods, each comes with a different methodology. The practitioner of Lean Sigma must first decide which is appropriate in any circumstance. As a general guide quality improvements are tackled using the Six Sigma methodology while efforts to enhance speed or reduce waste are tackled using the Lean methodology.

Figure 2.

Lean Sigma in the lab

Lean and Six Sigma both originated in manufacturing industries. The techniques have since grown in popularity in a wide range of environments from healthcare through to office work. We believe that while these techniques are invaluable it is important not to assume a "one size fits all approach" - a laboratory environment is not the same as a manufacturing environment and there are significant differences between different laboratory environments (some use large scale automation, while many use small scale bench-top equipment). Below are 10 key Lean and Six Sigma techniques that may be used in laboratories, but please bear in mind that not all techniques are applicable to all environments.

1. Structured problem solving ("DMAIC")

At the heart of Six Sigma process improvement is a structured method that guides team through improving a process:

DEFINE: The first step ensures that the team share an understanding of what problem needs solving.
MEASURE: The second step ensures that sufficient data is collected to understand the problem.
ANALYSE: The third step identifies the root cause of the problem.
IMPROVE: The fourth step identifies ways to solve the root problem.
CONTROL: The final step ensures that improvements are embedded and maintained.

2. Value stream mapping & waste identification

At the heart of Lean Thinking is the concept of value. Value is anything that your customers, clients or patients value (e.g. the result of a diagnostic test) but excludes aspects that offer no direct value to the customer. Lean Thinking provides a framework to help identify non-value adding activities (they frequently become so embedded that they are assumed to be a normal essential part of the process). Lean Thinking encourages us to look for seven types of waste (remembered by the acronym "Tim Wood"):

TRANSPORTATION: movement of material
INVENTORY: use of space and funds to hold inventories of material
MOTION: movement of people
WAITING: queuing or scheduled delays
OVERPRODUCTION: producing more than customer, client or patient wants or needs
OVERPROCESSING: excess bureaucracy or checks in process
DEFECTS: work that needs correcting or repeating

A Lean analysis of a laboratory will identify the value that is being created and will map the process used, highlighting where the seven wastes may be found in the process. A "future-state map" may be drawn up highlighting what the ideal process should be and then continual improvements are made (often involving many small changes) to progressively move towards the ideal future state map.

3. Process simulation

Laboratory environments are often complex; identifying the key resources that will enhance productivity is often hard; significant investment in equipment may be made only to find that capacity is unaltered because of another limitation (perhaps number of people). Process simulation allows a virtual model of the laboratory (or the part under investigation) to be made. "What if?" questions may then be asked and tested in this virtual environment, questions such as "what if we invest in more equipment?", "what if we operated an extended day?", "what if we can eliminate these non-value adding steps?". Process simulation can take much of the stress and costs out of process improvement allowing a large amount of work to be done in the virtual environment before applying the successful changes to the real laboratory environment.

4. Lean layouts

A key aim of Lean Thinking is to ensure that work flows through value-adding steps with little if any delay. When processes are divided between two locations a delay due to transportation (one of the seven wastes) is introduced. The delay caused by the movement alone may be minimal, but whenever movement is required it becomes natural for workers to perform their work in batches and to complete all that is required before the work is moved to the next step. This waiting for other members of the batch to be processed frequently adds up to the large majority of the total cycletime of the work. Lean Thinking therefore encourages continual reduction in batch sizes accompanied by changes in layouts to allow these smaller batches to be processed as efficiently as the larger batches previously were. An important component of this Lean flow is reducing delay by putting as many process steps close together as possible, often using multiple small pieces of equipment rather than one large piece of automation. Moving process steps close together also encourages simple communication between people, replacing e-mail with talking.

A number of techniques exists to ensure that the layout chosen is the best for the work to be done (any "one size fits all" approach must be avoided).

5. Workflow control

Any laboratory manager has conflicting aims: ensure the greatest productivity possible and ensure that results are delivered quickly. A few years ago a British newspaper ran a headline story complaining that our fire-fighters spend most of their time playing cards. The journalist had obviously not given too much thought to what he would like the fire-fighters to be doing should his house catch fire — would he like the fire-fighters to be busy putting out another fire or would he like them to be playing cards and free to attend immediately to his own house fire? This battle between efficiency (using the fewest people as possible to put out as many fires as possible) and effectiveness (to put out as many fires as possible before life and property are lost) exists in almost every environment, including the laboratory. Various Lean methods of workflow control exist to ensure the best possible balance between efficiency and effectiveness. Examples of workflow control include the following (with appropriate solutions chosen as appropriate):

CONSTRAINED WORK IN PROGRESS (ConWIP): ConWIP caps the total quantity of work your laboratory aims to complete (this is not possible in all environments); the usual level chosen is to cap work so that the laboratory works at 70-80% maximum capacity (which may be determined by simulation in complex environments).
MIXED SCHEDULING & EVERY PRODUCT EVERY (EPE): Mixed scheduling aims to continually increase the frequency of different types of work, minimising the delays caused by waiting for the next scheduled slot. EPE is a measurement of how frequently different types of work are scheduled.
KANBAN: Kanban is a signalling system to indicate when work is required; kanban may be used to control work itself or (more commonly in laboratory environment) replenishment of laboratory materials.
COMPLETION PRIORITY: When the bottleneck is a shared resource cycletimes may commonly be reduced by giving priority to work that has progressed further. The aim is to always work to clear work out of the system rather than to prioritise pushing work into the system.
"BUCKET BRIGADE": In bucket brigade a person carriers on with work in a process until they are able to pass it on to someone else (work should never be put down); the worker then goes and takes over a job at an earlier stage in the process.

6. Statistical process control

Six Sigma quality is achieved when there are fewer than 3.4 defects or errors per million in a process or laboratory method. A bed-rock of the Six Sigma approach is to continually monitor quality and continually drive down variation. The first step is to ensure that the process (method) is predictable and under control; once a process is under control then variation should be continually reduced. Statistical process control continually monitors the process and highlights as soon as the process appears to be coming out of control (outside of the normal operating limits); early action may then be taken to bring the process back under tight control. Additional tools help the laboratory manager understand where variation in results come from: for example how much is inherent in the method and how much is introduced by the method being performed slightly differently by different people. Analysis of the sources of variation allow the laboratory staff to target the correct part of the methodology to continually increase quality.

7. Design Of Experiment ("DOE")

Traditional method development keeps all parts of a method constant while adjusting one parameter. Occasionally two aspects may be varied to examine potential interactions. In DOE very efficient designs are produced that allow scientists to alter many aspects of a method simultaneously to rapidly identify the critical factors while also allowing interactions to be uncovered that are missed in methodologies that alter one component at a time. DOE aims to develop methods that give optimal quality and are resistant to variation.

8. Total productive maintenance

A modern laboratory is dependent on the reliability and quality of its equipment, whether it is multimillion pound automation or small bench-top equipment. Total productive maintenance (TPM) programmes aim to keep equipment in excellent working order so that scientists can trust that equipment will work, and work within specification required to deliver excellent results.

TPM programmes must be tailored to the type of equipment used but share a general approach. For any piece of equipment the laboratory manager should understand the reliability: what can go wrong, what does go wrong (and with what frequency), how should it be fixed (at a non-technical level, e.g. is calibration required, is a replacement part required, or is a service engineer required?), and how quickly is it being fixed. TPM can then focus on eliminating or reducing equipment unavailability caused by the most common problems.

TPM should also focus on quality of equipment: what specification limits are required to ensure reliable results, and how is the equipment currently performing? A TPM programme should give laboratory workers clear confidence that any equipment they use will perform with the required specification limits to ensure good quality.

9. "5S" & the visual workplace

"5S" has two main aims in the laboratory: the first is to ensure that a laboratory worker can find anything they need instantly (everything has a place, and everything is in its place), the second is to ensure that expensive laboratory environments are used for laboratory work and not mass storage of years of unused consumables and equipment.

5S has five steps:

SORT: sort through the laboratory and keep only that which is used
SET IN ORDER: identify a place for everything
SHINE: clean all the surfaces
STANDARDISE: ensure all workers use the space in the same way (e.g. have only one place for a particular type of laboratory consumable that everyone uses)
SUSTAIN: have procedures in place to ensure the laboratory remains impeccably organised (e.g. 10 min clear-up session per day, weekly inspection and score)

The visual workplace is an extension of the 5S environment. The aim of the visual workplace is to ensure that laboratory managers know the state of everything in the laboratory by simple displays within the laboratory. For example QC checks of equipment should be clearly displayed on or close to the equipment (not in some Excel sheet hidden on a server somewhere), a measure of the quality of laboratory tests should be prominently displayed and updated daily or weekly. The visual factory has several advantages: (1) it draws the manager into the laboratory, rather than managing from an office or PC, (2) it ensures that any drop off in performance is highlighted rapidly to all and all can help bring the laboratory back to peak performance, (3) it ensures important information is easily available at the point that it is required.

10. Change management

Any improvement process is dependent on two key factors: (1) the technical achievement of the solution (how well does the solution solve the problem) and (2) the quality of the implementation of that solution. A good solution that is well implemented is frequently better than a brilliant solution poorly implemented. As performance improves then improvements become harder to successfully implement as the "if it isn't broken, don't fix it" mindset creeps ever increasingly in. Fear of failure drives resistance to change. Change management is a critical part of Lean and Six Sigma improvements; people must be bought into the changes, risks must be understood, accepted and mitigated against (e.g. by trialling solutions before being implemented into critical processes). Most importantly, a culture of striving for the best must exist and be supported by management. Six Sigma offers many tools aimed at giving all stakeholders sufficient confidence to drive forward these improvements.

List of main tools used by Lean Sigma Lab

(But please remember the aim is not to use as many tools as possible!)

Lean tools	Six sigma	Production mechanics	Change management
Value stream mapping	process mapping	Simulation	Stakeholder analysis
Seven wastes	Graphical analysis	Theory of constraints	Project management
Lean scheduling methods	Capability analysis	Factory physics	Project selection
Lean layout design	Descriptive statistics	Queueing theory	Project charter
Voice of the customer	Statistical process control
Visual management	Comparative statistics	Optimization procedures	Creative tools
5S	Cause & effect analysis	Design of Experiment	De Bono's 6 hats
Mistake proofing	Measurement system analysis	Linear programming	Triz
Total productive maintenance	Sampling statistics	Non-linear optimization	Brainstorm
5 Whys	Risk analysis	Heuristics	Pugh matrix
	Hypothesis testing
	Design for Six Sigma