STATISTICAL PROCESS CONTROL (SPC) AND FOOD SAFETY MANAGEMENT
A PASSION FOR FOOD INDUSTRY COMPLIANCE
FREE RESOURCES
A number of useful Whitepapers and other references are available free from our website – safefood360.com/free-resources/
WHAT IS STATISTICAL PROCESS CONTROL (SPC)?
WHAT IS SPC?
Statistical Process Control (SPC) is an industry-standard
methodology for measuring and controlling quality during the
manufacturing process.
Quality and food safety data in the form of Product or Process
measurements are obtained in real-time during manufacturing.
This data is then plotted on a graph with pre-determined control
limits. Control limits are determined by the capability of the
process, whereas specification limits are determined by the client's
needs
GENERAL BENEFITS OF SPC
• Reduce variability and food safety failures
• Improve productivity
• Reduce costs
• Uncover hidden process personalities
• React to process changes
• Inform decision making and improvements
WHY STATISTICAL PROCESS CONTROL (SPC)?
ISO 22000 – ‘PROCESS APPROACH’
Food safety standards are moving more and more towards a process based
approach to developing, implementing and improving food safety
management systems
This represents a greater alignment and integration with compliance and risk
based models for business in general
The recent publication of the new ISO 22000 Standard for Food Safety
Management Systems is further evidence of this trend
The PLAN-DO-CHECK-ACT Cycle is being cited by the ISO
Use of Statistical Tools is ideal for meeting this requirements
ISO 22000 – RISK & COMPLIANCE MODELS BASED ON PDCA
The following is the Plan Do Check and Act cycle addressed in the ISO 2200 Standard
PLAN DO
Establish the objectives of the stem and its processes, provide Implement what is planned
the resources needed to deliver the results, and identify and
address risk and opportunities
CHECK ACT
Monitor and (where relevant) measure processes and the Take actions to improve performance as necessary
resulting products and services, analysis and evaluate
information and data from monitoring, measuring and
verification activities, and report the results. SPC is a validated
set of tools for achieving this.
SQM CODE EDITION 8
SQM CODE EDessentially
Requirements 8 call for SPC to be utilized
THE FATHER OF SPC
W. EDWARDS DEMING – AMERICAN ENGINEER & STATISTICIAN
THE FATHER OF QUALITY October 14, 1900 –
December 20, 1993
• American engineer, statistician, professor, author, lecturer, and management consultant.
• Helped develop the sampling techniques still used by the U.S. Department of the Census
and the Bureau of Labor Statistics
• Best known for his work in Japan after WWII where he promoted the use of statistical
tools in the rapid development of the Japanese economy and manufacturing industry
• Advocated that an understanding of overall processes involving suppliers, producers, and
customers of goods and services was essential for quality
• Argued that a knowledge of variation, the range and causes of variation in quality, and use
of statistical sampling in measurements was a powerful tool for businesses
PROCESS VARIATION
PROCESS VARIATION
W. EDWARDS DEMING “Uncontrolled variation is the enemy of quality”
American engineer & statistician
WHAT IS PROCESS VARIATION?
• All processes have variation.
• Variation is the inherent tendency of a process to deviate, move, shift or vary around a
mean central point.
• Variation can be observed throughout nature – human height and weight, atmospheric
temperature etc.
• Impossible to eliminate variation.
• The objective is to understand and know the degree and nature of the variation.
• Once this has been done it is possible to control variation
• Every process has a typical or normal distribution around the mean. Most readings will fall close
to the mean while fewer readings will fall out towards the tails, producing a typical distribution
curve – see opposite.
• Once you know what this looks like for a process you know if your process is in control…
UNDERSTANDING YOUR PROCESS
PROCESS DATA
W. EDWARDS DEMING “Without data you’re just another person with an opinion”
American engineer & statistician
WHAT IS PROCESS DATA?
• All processes and food safety systems “give off data”.
• In order to understand the degree and nature of variation you need to collect data.
• In food processes and safety management data comes in many forms:
• Process parameters – times, temperatures, pressures, pH, acidity, specific gravity etc
• Audit results – pass, fail, ratings etc.
• Non-conformances – numbers, types, criticality,
• CCP monitoring data
• Sanitation and cleaning – pass, fail, recleans, ATP, TPC etc.
• Data may be Variable (72 degrees for 2 minutes) or Attribute (Pass / Fail / Acceptable /
Unacceptable etc).
• Much of the data collected by food businesses is for evidence of compliance. Very little
is used for understanding the processes or driving improvement.
• MISSED OPPORTUNITY – hidden in this data is an ocean of insight which if unlocked can
help to improve processes, conformance and worker experience.
MEASUREMENT AND MONITORING
W. EDWARDS DEMING “You can’t manage what you don’t measure”
American engineer & statistician
WHAT TO MEASURE?
• In order to understand and therefore control your processes you need to collect data.
• This means you need to monitor and measure. Principle of HACCP.
• There is an almost infinite number of things in a food business which can measured. Just
because you can measure and monitor something does not mean you should.
• In food safety you will monitor based on two key requirements:
• Compliance – “I have to measure this because the law or my customer says I should”
• Risk – “I have to measure this because there is a significant risk contained within the
process”
• Risk assessment is the key and is reflected in most food safety standards.
• For example – thermal treatment process to kill pathogens – minimum time and
temperature required. Necessary to determine if at the set point the process is capable on
achieving the required kill based on variation of the process. Only measuring and collecting
sufficient data and then analysing it will allow you to fully determine the process capability.
USING DATA TO UNDERSTAND YOUR PROCESSES
W. EDWARDS DEMING “If you can’t describe what you are doing as a process, you don’t know what you are doing”
American engineer & statistician
USING YOUR DATA
Describing your Process
• Once you collect your data you have a few options:
Process : Tight variation –
• Simply determine if each data point meets the specification – most companies do high capability
this.
• Analyse the data (using standard statistical techniques) to describe the process in
terms of capability to meet the specification – consistently.
• Barriers: Process : acceptable
• Data needs to collected, collated and organised capability
• Tools are required to help with this process
• SPC skills required to interpret the data
• Above requires time, money and resources
• Describing the Process: Process : wide variation –
• Once collected and using standard statistical tools, the data will tell a story poor capability
• Is the process ‘normal’ – is the distribution of results showing a normal bell-curve?
• Is the variation capable of meeting the specification – standard deviations within
+/- 3 standard deviations?
• Is there any bias or skewing in the distribution of results?
• Are there any trends above or below the set point which may indicate a shift or a
possible loss of control?
SPECIFICATION VERSUS CONTROL
W. EDWARDS DEMING “Meeting specifications is not enough.”
American engineer & statistician
WHICH IS BETTER?
• There are two ways to measure the performance of a process:
• “Is the process within specification?”
• “Is the process in-control?”
• In Specification
• This approach is more basic and involves simply taking a sample, measuring it, comparing it against the specification and deciding if it
passes or fails.
• E.g. Specification pH: 5.5-5.9 Observed Result: 5.7 Decision: In specification
• It is a valid approach and sometimes can’t be avoided, however, it is based usually on one data point and does not take account of
variation. There may be out of specification results not detected by the sampling plan.
• In Control (SPC)
• This approach is usually based on a detailed capability study designed to determine the process variation and its capability to stay
within limits.
• It takes account of the probability that the observed mean and range of a sample size is as per the normal process and therefore in
control.
• In short, this approach will tell you if the process has shifted away from it normal variation and likely to produce out of specification
results.
SPECIFICATION VERSUS CONTROL
OUT OF SPECIFICATION
• Example of a simple chart plotting the result against specification
SPECIFICATION VERSUS CONTROL
OUT OF CONTROL (SPC)
• Example of a data charted using SPC techniques.
• Here we can see the data provides more insight into the actual
distribution of the data being collected and the capability of the
process to produce consistently quality or safe product.
STANDARD SPC TOOLS
STANDARD SPC TOOLS
CAPABILITY STUDIES
Histograms and Setting of Control Limits
Control : Data used to Analysis : Data used to Capability : Data used to Histogram : Graphical
determine if the process is in determine the mean process measure the capability of the representation of all the
control point, standard deviation and process to meet the required observed results stacked to
to set control limits specifications show the mean point and
spread
STANDARD SPC TOOLS
CONTROL CHARTS
Variable Data
X-Bar : Chart used to plot the Range : Chart used to plot Individual : Chart used to
average of a sample of the range between the plot the process where
measurements against limits highest and lowest readings individual samples are used
of a sample of measurements
STANDARD SPC TOOLS
CONTROL CHARTS
Attribute Data
P and nP : Charts used to plot
the number or percentage
defects for attribute data
STANDARD SPC TOOLS
IMPROVEMENT
Parato Charts
Parato Charts: Charts used collate and organise
data based on magnitude – from the most
frequent to the least frequent occurrences. Helps
to focus the attention of your resources on the
areas where maximum return can be achieved.
APPLICATION OF SPC IN FOOD SAFETY
MANAGEMENT
EXAMPLE 1 – CCP CONTROL
STANDARD SPC TOOLS
EXAMPLE 1 - CCP CONTROL
• In this example, we see the use of an individual control chart to monitor
a temperate control.
• As the CCP monitor completes their checks and records their data, this
can be plotted on a control chart which clearly indicates the upper and
lower control limits.
Analysis : Trained CCP monitors can more effective Distribution : The data collected from frequent Control Chart : CCP Monitors can see immediately
in understanding the state and level of control for CCP checks can be plotted to determine any if there recorded result is in control or if there are
their CCP with detailed data analysis change in process capability any warning trends indicating a shift in the process
mean
APPLICATION OF SPC IN FOOD SAFETY
MANAGEMENT
EXAMPLE 2 – PARATO CHARTS
STANDARD SPC TOOLS
EXAMPLE 2 - PARATO ANALYSIS OF CORRECTIVE ACTIONS
Parato Charts – Corrective Action by Source
• In this example, data relating to the source of corrective actions can be
organised to easily see where the majority of corrective actions are
originating.
• We can see that AUDITING is the main source of all corrective actions
while CONTAMINATION is a minor contributor.
• Using this data, the management team may decide to put in place a
specific improvement project to research further the main areas of
auditing causing the high level of corrective actions.
• This could also involve further Parato analysis of the non-conformances
arising from audits. In this example we can see that the Daily Hygiene
Inspection is the main suspect
APPLICATION OF SPC IN FOOD SAFETY
MANAGEMENT
EXAMPLE 3 – VALIDATION OF SANITIZING &
CLEANING METHODS
APPLICATION OF SPC IN FOOD SAFETY MANAGEMENT
EXAMPLE 3 - VALIDATION OF SANITIZING & CLEANING METHODS
CODEX “Obtaining evidence that a control measure or combination of control measures, if properly implemented,
Definition of Validation is capable of controlling the hazard to a specified outcome”
• Food businesses by law need to clean and sanitize their facilities
• This is required to produce safe and legal food products and remove hazards such as pathogens, allergens and chemicals
• Food businesses employ Standard Sanitizing Operation Procedures (SSOP) to achieve this
• In order to ensure that these methods are in fact capable of removing the hazards and reducing the risk to an acceptable level – validation of the
cleaning procedure is often required. Particularly for High-Risk unit operations.
• Simply setting up a procedure and checking it a few times (e.g. ATP, swabbing and culturing, residual testing etc) is not sufficient to validate a
cleaning method as it will not account for the possible variation from employees, chemicals and other parameters.
• Validation of cleaning methods can be conducted statistically using a capability study
APPLICATION OF SPC IN FOOD SAFETY MANAGEMENT
EXAMPLE 3 - VALIDATION OF SANITIZING & CLEANING METHODS
Capability Study Steps
• Setup and document the cleaning
procedure
• Train employees in cleaning method
• Conduct cleaning according to
procedure
• Test samples – ATP, TBC, Allergens,
Residual Chemicals etc.
• Enter results into the Excel Tool
• Repeat the above process for a
significant number of cleaning runs
• Determine capability of the cleaning
method to meet the cleaning limits.
• Modify the cleaning method if required
and repeat until validation is confirmed
APPLICATION OF SPC IN FOOD SAFETY MANAGEMENT
EXAMPLE 3 - VALIDATION OF SANITIZING & CLEANING METHODS
Capability Study
• Following each cleaning run, conduct microbiological, allergen or other tests as
required on food contact surfaces and other sample points.
• Ensure the test method is defined in advance and repeated for each test.
• The person conducting the test should also be trained and competent to
conduct the test.
• All data should be recorded accurately and retained for analysis later on.
APPLICATION OF SPC IN FOOD SAFETY MANAGEMENT
EXAMPLE 3 - VALIDATION OF SANITIZING & CLEANING METHODS
Analyse Results
• Collate the results obtained over the validation
period and conduct appropriate statistical analyses
• Determine the variability in the efficacy of the
cleaning SSOP’s
• This is your Capability Study
• It will estimate the process average and variation,
relative to the specified tolerances for the standard
cleaning procedure
• Go to https://safefood360.com/free-resources/ and
download the free cleaning validation tool and
whitepaper
QUESTIONS?
