Quality | Should Your Lab Perform Gravimetric Reduction Before Point Counting?
By David F. Gallup, EMLab P&K Chairman
The term "Quality" usually means different things to different people. For example, a story about IBM relates that at one point IBM decided to have some parts manufactured in Japan as a trial project. In the specifications, they stated that they would accept three defective parts per 10,000. A letter accompanied the first shipment stating, "We had a hard time understanding North American business practices. But the three defective parts per 10,000 have been separately manufactured and have been included in the consignment. Hope this pleases you."
A thorough discussion of quality would take far more space than we have here and probably consume more time than you have to read it. Consequently, what follows is by necessity a simplification that I hope will still be useful.
What do we mean when we say "Quality?" Quality of what? When people talk about high quality IAQ labs, the implication is usually about 'high quality' data. Generally, measures of analytical quality include 'accuracy' and 'precision.' Accuracy is a measure of the ability of a process to provide the 'true,' or 'real,' value. Precision is a measure of how repeatable the results are. It is possible to be accurate, precise, neither, or both. For example, a process that is precise but inaccurate would provide consistent results, but the results would not be the true or actual value. A process that is accurate, but not precise, would have an average that was close to the true value, but have a large variance in the measured result (More on this can be found at Wikipedia). What IAQ professionals would like is to have confidence that their lab provides results that are both accurate and precise.
Let's take a look at some of the factors that impact the accuracy and precision of spore trap analysis, specifically:
Experience and Training of the Analyst
Accurately identifying fungi requires an understanding of the fungi themselves, months of training, and an ability to recognize and understand the implications of subtle differences between particles. Perhaps another frequently overlooked characteristic is the personality of the analyst. Not all people are capable of and willing to sit at a microscope and maintain the high level of attention required to make correct fungal identifications on a consistent and ongoing basis.
Quality Control Systems
Let's be honest. Everybody makes mistakes. So, the question isn't "Are mistakes made?" but rather "How often are they made?", "How many go undetected?" and "How are they handled when detected?" You'd like your lab to have robust systems and processes in order to actively detect errors when they occur, notify clients when necessary, formally investigate errors to find root causes and then implement corrective actions to reduce the chances of similar errors happening again. Such a system would result in an error rate that declines over time and 'quality' that improves over time. These processes and systems need to be methods that are actually applied and used, not just some system to be shown during site visits from the AIHA. For example, the laboratory analysis program that we use, LabServe, automatically issues all the QA (Quality Assurance) and QC (Quality Control) samples. This provides assurance to our clients that these types of samples (10% of all samples) are actually sent out to the analysts, analyzed, and monitored. Manual systems, on the other hand, both cost more and are easily overridden when it's busy, a time when, arguably, these systems are needed the most. Some labs have claimed to process 10% QA samples; but, really didn't. LabServe helps us have confidence that QA and QC samples are both actually processed and that they are processed efficiently so we can focus on improving quality rather than administration.
In some ways this is obvious, in others more subtle. For example, let's say employees receive penalties for QC errors. This would lead to a culture where there is an incentive to not bring up quality-related issues and, perhaps paradoxically, would result in inferior quality as observed by the client. A culture of the laboratory that emphasizes competing on being a low cost provider will result in similar effects. The company culture is largely generated from the top. Generally, you can take a look at upper management to get a sense of the culture. As a lab, you automate and make processes as efficient as you can; but, at some point and in certain dimensions, quality and cost compete. Which direction will the lab culture push you? Is the management team truly committed to the quality of analysis? Are they committed to compliance with ISO17025 and AIHA, NVLAP, or other accrediting policies or do they ignore requirements that reduce profits?
Quality of the Sample
It needs to be pointed out that the quality of the sample also affects the results, both in terms of accuracy and in precision. An overloaded spore trap sample adversely affects both. Sampling pumps that are not calibrated will affect accuracy. A poor sampling strategy, or inattention to material details such as the state of the HVAC system or access to the outside air impact the quality of the sample in terms of appropriateness for evaluating the hypothesis being tested, even if the lab is flawless. The expert in the field plays an important part generating relevant, high quality data.
Standard Operating Procedures
Perhaps this is easy to overlook. People understand that protocols are relevant to the analytical process. However, the entire process needs to be controlled, from receiving samples through reporting. If the person who normally receives samples is out sick one day, it's important that whoever fills in for them has been formally trained and signed off on the process and that they have written standard operating procedures to follow so that they're not relying on memory that a certain sample type must be placed in an incubator within a certain time frame. While events such as these, individually, may be unusual, when you stack up the hundreds and thousands of individual steps and processes that occur within a lab, many unusual events occur. Having good training systems and formalized procedures result in a process that produces higher quality data. Notably, many analyses in this industry do not have prescriptive methods dictated by the AIHA or any other regulatory agency. In these cases, it is critical that you have experts on staff to develop the best SOPs for each analysis.
OK. So, we quickly reviewed some of the processes that impact the quality (precision and accuracy) of your results. How can you tell that these are in control within your lab? It's admittedly tricky. At a minimum, use an accredited lab. While accreditation is not a guarantee, it is a sign that the laboratory not only has QA and QC systems in place, it is also a sign that these systems meet a minimum, external, standard verified by an independent third party. The AIHA EMLAP program, which accredits laboratories for a variety of environmental microbial analyses, requires correct identification of isolates sent throughout the year to the laboratory to demonstrate analytical proficiency. It sets minimum education and experience standards for the analysts. It requires ongoing QA and QC samples, active participation by management in the quality systems and dictates a level of transparency in reporting. It also requires compliance with ISO 17025 standards and passing of an onsite visit by an AIHA site inspector. This particular accreditation should not be viewed lightly and is a serious quality statement by a laboratory.
Unfortunately, laboratories can and do skirt the rules of accrediting bodies and such accreditations do not cover all of the aspects we've discussed here. Unfortunately, the only 'real' way to assess these factors would be to conduct your own thorough site visit of your lab. Then you could see what processes exist and whether they are genuinely followed or merely posted for show. Naturally, most people and industrial hygiene companies do not have the bandwidth, nor the resources to do this. A more practical approach would be to visit the lab and ask to see training records or evidence that quality assurance samples are being processed at the required frequency. If this is also impractical, then consider calling the lab and asking for evidence of recent QA and QC samples, results of performance testing samples for the year, what percentage of their QC samples passed or the rate of revision of data after reporting. These are indications of whether monitoring and controlling quality is genuinely part of their day-to-day activities.
Finally, consider other aspects that are indirectly associated with quality. For example, does the lab routinely suggest sampling as the answer to all ills? Frequently, samples are not required in an investigation. This is an indication of how honest or ethical the lab is and whether they are looking out for your interests, including your need to control lab fees. Does the personnel you deal with know the limits of their knowledge and occasionally say, "I don't know. Let me find somebody who does?" Further, does the lab have experts who have years of experience and training in particular niches whom you can contact for help?
Since this is often a rather dry subject, I'd like to conclude with a humorous pipe specification: "Outer-diameter of all pipes must exceed the inner-diameter. Otherwise, the hole will be on the outside of the pipe."
By Louie Bustillos, EMLab P&K Asbestos Analyst
It has been found that point counting may be a more accurate method to determine the percent of asbestos present in samples that contain low concentrations of asbestos. The most important part to point counting a sample is 'how' the sample is prepared by the laboratory. Certain types of samples need to be manipulated a certain way in order to 'expose' any possible asbestos present in the sample. If during the sample preparation a portion of the sample is dissolved away then gravimetric reduction should be used. An example of this is when samples are treated with an acid, to dissolve away any carbonates in the sample. These types of samples need to be gravimetrically reduced before doing point counting, because the analyst needs to know how much of the sample was dissolved away by the acid treatment. If the dissolved carbonate weight is not taken into account then the asbestos point count result could be reported at a much higher concentration than what the 'true' asbestos concentration was in the original untreated sample. This could possibly lead the lab to report a false positive for the sample. A sample that was originally < 1.0% asbestos could now be reported as > 1.0% asbestos and will be considered as asbestos containing material. This may also happen with samples that are treated in a way that dissolves away organic material from a sample by either treating the sample chemically or by burning the sample in a muffle furnace. Some examples of samples that need to be gravimetrically reduced before point counting are certain types of joint compounds, mastic samples, roof samples and floor tiles. The purpose of gravimetric reduction is to keep track of the weight of the sample during the preparation process.
A common question that clients ask the lab is: "Why does gravimetric point count cost more then regular point count?" This is due to the fact that it takes much longer to prepare samples for gravimetric point count and the supplies to prepare gravimetric point samples cost more then a regular point count.
A regular point count is prepared by making multiple preparations of the sample. A portion of the sample is homogenized and then multiple sub-samples from this homogenization are placed onto multiple slides that have specific refractive index oil. Then a cover slip is placed over each sample mixture. The slides are then placed onto the microscope for analysis. Every slide is moved randomly and each time the analyst stops on a part of the sample, a "point" is counted. The number of points that need to be counted depends on the analytical sensitivity that is needed. The most common point count is the 400 point count.
There are many more steps in the preparation of a sample for gravimetric reduction. The sample needs to be weighed, then burned over a period of time, then treated with an acid and then collected onto a filter in a filtration apparatus. Once it has been filtered then the sample needs to be placed in a drying oven. Only at this time is the sample prepared like a 'normal' PLM sample for point counting. During each step of the preparation process the sample weight is measured and recorded. Using this weight information allows the laboratory to back calculate what the 'true' asbestos concentration was in the sample before it was manipulated for analysis.
This article was originally published on May 2008.