Quality Assurance Process for Measurement

Quality Assurance Process for Environmental Measurement

By Henry Nowicki and Barbara Sherman

 

Chemical measurement data are often the basis for critical decisions on vital matters, including the health of individuals, environmental protection, and the production of safe, reliable, and useful products and services.  As a result, the demand for measurement data is ever increasing, and therefore, measurement data must be with reliable with unequivocal evidence to prove accuracy and precision.  The philosophy and procedures by which data users can obtain proof of the reliability accuracy of measurements is commonly referred to as quality assurance.

 

The environmental industry in particular requires a higher level of quality assurance compared to most other industries.  Environmental products require millions of dollars in research and development, and most never make it out of the pipeline and actually come to market.  Adding further to the quality assurance task, the environmental manufacturing process is highly regulated and often, due to cost concerns, dependent upon outside services for vital business needs. Environmental product manufacturers also rely on complex manufacturing and analytical tools to make chemical measurements. They need to .defend the decisions they make based on such measurements to the FDA and, potentially, in court against plaintiffs. Therefore, the environmental industry must make extraordinary efforts to assure quality.

 

In order to achieve this goal, analytical chemists in the environmental sector need to think and work more like process engineers. Bringing analytical chemistry to the manufacturing production line is one of the best ways to help assure quality. Testing end products is useful, but automated testing methods, such as Near-IR, should also be performed early in the manufacturing process.

 

To be successful, chemists and engineers must work together to identify and meet quality needs environmental manufacturing processes.  Since engineers work on the process units used in manufacturing, chemists can best communicate with the engineers by breaking down their measurements into process units.

 

QC + QA = Quality Assurance

Quality assurance has two separate but related activities: quality control (QC) and quality assessment (QA). Quality control is a fairly well established practice, while quality assurance is a relatively new concept. However, to truly achieve measurement quality assurance, both QC and QA must be operational and coordinated.

 

Quality control is the overall system of activities used to control the quality of a product for service so that it meets the requirements of users. The aim of QC is to provide quality that is satisfactory, adequate, dependable, and economic.

 

An example of a QC task: the mass spectroscopist calibrates the mass spectrometer before running samples to assure proper mass assignments, relative abundances, resolution and sensitivity of the instrument before running the samples. This calibration is done by running a known standard compound on the instrument. The operator

documents that the instrument is working within an acceptable range before running a sample.

 

Quality assessment is the system of activities used to ensure the overall quality control function is being performed effectively.  It involves a continuing evaluation of the products produced and of the performance of the production system. An example of a QA task: management provides known standard samples for the instrument operator to

check on their performance and adequacy of QC to get acceptable performance on the submitted known sample. Management has two options when they prepare the known QA sample: blind or blind-blind also call double-blind samples. We recommend using both options on defined schedule to test the proficiency of a measurement process. A

blind sample is useful because the operator knows it is a check on the measurement performance. A good operator will do their absolute best on a blind sample, it makes him think. In order to avoid bias and false security about the QA program blind-blind samples are submitted to the operator. A blind-blind sample is a check sample which looks like a regular sample. Thus the operator handles and reports the blind-blind sample in a batch without knowing it is a known standard. Blind-blind samples can be prepared by spiking real samples or submitting the same sample with different sample identifications.

 

Quality assurance is the system of activities used to provide the producer or user of a product or a service the assurance that it meets defined standards of quality with a stated level of confidence. It is the end result of effective quality control and quality assessment programs.

 

Chemical Analysis as a System

In order to truly achieve quality assurance, measurements need to be established as a process analogous to a manufacturing process. Instead of having raw materials converted to finished products, the analytical chemist brings in relevant samples and converts them to defensible reports and valuable information. The chemical measurement process must be brought into a state of statistical control, i.e., the individual measurements must be defined by a statistical distribution, where characteristic precision and accuracy can be assigned to the data output. The analytical chemist needs to compare the precision and accuracy of the methodology to what the project needs to assure the proposed method meets requirements. If precision or accuracy needs to be improved, the sooner the problem is recognized, the better. Clearly, quality control relates to all that is done to attain and maintain the state of

statistical control. Statistical control always has to be demonstrated before samples are run. Statistical control is demonstrated by running known standards, ideally with the  same matrix and concentration range of interest to the project.

 

It is clear that chemical analysis is more than a single process and is better characterized as a system consisting of a series of interdependent processes as indicated by the flow diagram in Table 1.

 

Table 1,

View Measurement as a Process

 

Problem Specifications

A detailed exact statement is needed of the question or situation to be solved.  It is most important that we work on the RIGHT things.

 

Correct Model

A state-of-the-art description of a theory or system that accounts for all of its known properties is known as a correct model. The model should be updated periodically as you get new data.

 

Quality Assurance Plan

A quality assurance plan is a detailed road-map to assure a successful project.  It is often the size of a text book when completed. It is a critical but labor intensive document if done correctly.

 

Sampling Plan

A sampling plan is one of the most important parts of the overall planning process to obtain defensible and useful data. The plan should address types and number of samples to be taken, the details of collection, frequency of sample collection and the procedures to be followed.

 

Sample Transport and Storage

Transport and storage pf samples needs to be validated by a stability study for the analytes in your measurement process. Proper sample containers, glass or plastic, and chemical additions to samples to stabilize the analyte are important to measurement success.

 

Analyte Isolation

Chemical measurements are often like finding a needle in a haystack. Separation science is used to isolate the analyte to avoid false positives or false negatives, when we make chemical measurements.

 

Instrument Calibration

Instrument calibration is the comparison of your results from running a standard on your instrument with acceptable result.  The goal is to eliminate any deviations in the accuracy of the two results by having the operator make necessary adjustments to the instruments.

 

Quality Control

Quality control is the summation of all of the things an analyst performs to assure that data user is satisfied. The standard operating procedure (SOP) documents all of the steps the analyst must follow. The SOP contains directions for the analyst to follow when unexpected measurement problems arise.

 

Concentration

Often the sample must be concentrated or diluted to obtain analyte detection. Detection limit is the smallest amount of analyte that is measurable at a stated level of confidence.

 

Fractionation

Fractionation is used to get the purest possible product. In today’s market being able to fractionate the optically active forms of drugs is important. One optical form may be beneficial to the patient and the other form contributes negative side-effects. Fractionation can be the difference between market success and failure.

 

Derivitization

Some analytes need to be derivatized before they can be analyzed, because the physical-chemical properties do not allow the measurement. Derivatization changes the properties of the analyte and can also be used in structure elucidation, i.e., a mass shift.

 

Separation

The data validation process should be separated and independent of the original data providers. Management needs to obtain external performance evaluations on a defined regular schedule.

 

Qualitative Identification

Evidence for correct identification of the analyte is critical. Data providers need a document which specifies identification criterion and handling of chemical interferences.

 

Quantitation

Quantitation is the analysis of a substance that determines the amounts or proportions of its chemical constituents. The degree of uncertainty must accompany the reported quantitative values.

 

Report Record Keeping

The archival process must meet the needs of all stakeholders. The organization of reports and records must have a SOP and they must be in a readily accessible repository.

 

Validation of Data

Data should be validated before it leaves the data providers facility. Once the data user receives the information it is again checked to assure its quality for their applications.

 

Continuous Improvement

After each project or task is completed, ask yourself if we can improve the next time. This does not include finger-pointing and negative behavior as the main focus. We want a culture which brings out the best in all.

 

Table 1 footnote:  There are many sequential and internally related parts to a measurement process. It is important to take the time to list the sequential most important parts of individual processes. List the parts in a measurement on a lined piece of paper. Leave enough empty lines between each item in the initial list. This space will be filled in with more thought. Colleagues and outside consultants need to add parts to the list. Getting all stakeholders involved early shows good it is very important to develop trust and teamwork as part of the working process culture.

 

Turning the Problem Into a Model Solution

Chemical measurements are undertaken to answer questions necessary in solving problems. It is important to clearly and precisely define the problem. Much like when speaking with a teenager about their problems, it is important to practice active listening to correctly understand the real problem. The original problem expressed is often not the real problem that needs to be understood. If an effective solution is to be found you must get the specifications for the problem. One of the most important steps in the process is to ensure the correct actions are being taken to reach an effective solution. Correctly defining the problem and its possible future iterations is the key to achieving this goal.

 

A problem must be represented by a model that sets forth the questions that need to be answered and defines the data requirements and how they will be used to arrive at a solution. Based on the model, a measurement program can be planned to address sampling, calibration, the methodology to be used, and the quality assurance

procedures to be followed. The design and implementation of the chemical measurement system is often an iterative process, and every measurement program should include a feedback mechanism, which provides for identification of deficiencies and the initiation of corrective actions for all parts of the system as required.

 

It is important to take the attitude of continuous improvement. This does not mean the Stakeholders are continuously changing their mind, but rather that before the next measurement is performed, the stakeholder revisit the plan with the attitude of further refining the measurement process and making it better. History has shown that measurement improvement is required by regulators and users. Therefore, a system of continuous improvement is important. Continuing education is an important part of keeping up with the latest advances in chemical measurements.

 

Conclusion

All components of the system are critically dependent on each other and are implicitly or explicitly present in every measurement activity. Failure to recognize the importance of each component and to properly design an appropriate system is a major defect of many analytical measurement programs.

 

The process used to obtain analytical data should also be used with necessary modifications by engineers and human resource management services. Having a valid plan and process which solves the problem is most important. Most of use are facing global competition, such as from China, but an area which the United States presently has a competitive advantage is management processes. We all need to get involved in the process designs for our work if we want to stay competitive.

 

About the authors

Henry Nowicki, Ph.D. directs the laboratory testing and consulting services for PACS, a provider of environmental measurements and products and services for the activated carbon industry.  Dr. Nowicki has been an expert witness in over 40 cases and is a senior instructor for the PACS short course program. He can be reached at hnpacs@aol.com  or 724 457-6576.

  

Barbara Sherman, M.S. and MBA directs the PACS short course program, conferences, and expositions. She can be reached at barbpacs@aol.com or web-site: www.pacslabs.com  for more information.

 

PACS will host the 18th international Activated Carbon Conference and Courses program at the Crowne Plaza Hotel, near the Pittsburgh International Airport. The Conference will be on October 19-20,2006, and the courses will be provided October 14-18, 2006 and October 21-27, 2006. This event will be of particular interest to those involved with water and air purification and other carbon applications. PACS has available short courses for scientists, engineers and managers.

 

PACS has provided short courses for 27 years.  PACS has over 57 1-3 day courses designed for scientists and engineers to help make them better at their professional jobs. Courses are provided monthly in major cities and at the clients time and place.  On-site classes have the advantage of customization for the client’s interests.  To get more information about PACS:  Testing, Consulting and Training, go to www.pacslabs.com or call 724. 457. 6576