Choosing an Analytical Laboratory

If you need to have an environmental sample analyzed to determine whether or not it meets criteria for some use or disposition, you will want to choose an analytical laboratory that has a demonstrated capability to produce reliable and accurate results. The task of finding such a lab can be difficult for those unfamiliar with all the technicalities of the regulations and analyses.

This guidance was created to aid in this process. Analytical costs can be significant, but with a little background you can make sure you are getting results that will meet your needs without unnecessary testing, with the quality you deserve, at a reasonable cost.

Cost should not be your only consideration in choosing a laboratory. Certain questions should be asked of the analytical lab to determine if they can produce the services you will need to meet the regulatory requirements. Ask whatever questions you must to determine the following:

Is the lab accredited by Department of Ecology?

WAC 173-303 does not require the use of accredited labs for dangerous waste designation analyses. However, WACs 173-216, 173-226 (waste discharge), 173-220 (National Pollutant Discharge Elimination System--NPDES), and 173-340 (toxics cleanup) require the use of accredited labs. Also, Executive Policy 1-22 requires use of accredited labs. In any case, it is prudent to use such labs as the best indicator that the lab has the capability to do a good job. Furthermore, it is unlikely there is a lab in the state that can analyze environmental samples that is NOT accredited by Ecology.

Ecology’s Lab Accreditation Unit maintains lists of accredited labs which are available by mail or e-mail. The list can also be accessed on the internet here.

Not all accredited labs can perform all analyses that might be of interest to you. The lists of accredited labs identifies which categories of testing the accredited lab can do (i.e., general chemistry, trace metals, organics by gas chromatography (GC), organics by GC/Mass spectrometry, radiochemistry, microbiology, and bioassay). A database including each specific analyte for which the labs are accredited is available on the internet here.

If there is any question about whether or not an accredited lab can perform a specific method, you can contact the lab (contacts and phone numbers are part of the "List of Accredited Labs"). Labs accredited for general chemistry AND trace metals AND organics by both GC and GC/Mass spec are very likely to be able to do whatever you need.

If you have trouble finding an accredited lab for a specific method, you can contact the Lab Accreditation Section, either by phone at (360) 895-6145, fax at 895-6180, or e-mail to cosc461@ecy.wa.gov

One final word about lab accreditation. Data users are prone to relying on the fact that a laboratory is accredited as "proof" of the quality of data. This is far from the truth. Just as accreditation of a university shows that it is capable of teaching certain subjects but not that every student who graduates is going to be successful, laboratory accreditation only shows that the lab is CAPABLE of providing accurate, representative, comparable, complete, and defensible data. It is up to you to demand quality control test results and other data (copies of chain of custody forms, for example) to show that a lab’s capability was used effectively for analysis of your samples.

If a lab is not accredited by Ecology, you need not continue using this guide… find a lab that IS accredited.

Does the laboratory perform the test in-house?

To provide full analytical services, many laboratories will accept samples for testing that they will in turn send to another lab for the actual analysis. This is an accepted procedure, and there is nothing inherently wrong with it. But you lose some degree of control. If a laboratory uses outside services, it should be able to provide you with all the information you require about those supporting labs. You should demand that supporting labs meet the same criteria as the primary lab, including the accreditation requirement. If the primary lab does not want to identify their supporting lab(s) for any reason, choose some other lab.

Does the laboratory identify the methods used in its reports, and are they the methods you require them to use?

Environmental regulations are usually very specific about which analytical methods must be followed for samples governed by a given rule. If the laboratory uses a different method or significantly modifies the method, results will not be valid unless you have received permission from the Department of Ecology to use the alternate or modified method(s).

Many tests have several steps involved in obtaining the desired result. For example, some samples require an extraction procedure and sometimes a cleanup procedure before the actual analysis. There are standardized methods for doing these extractions and cleanups, just as there are for the analytical process. The extraction, cleanup, and analytical methods to be used are usually specified by the regulatory agency in rules, such as WAC 173-303, 173-220, 173-216, etc.

It is YOUR responsibility to make sure the lab is using those methods unless, as mentioned above, you have a waiver from Ecology to use an alternate or modified method. You should keep in mind that it is not the LAB that is given permission by Ecology to use an alternate or modified method, even though some might try to convince you that such is the case. For example, if WAC 173-303 says you must have your sample analyzed for organics using EPA Method 8270, and the lab tells you they have permission to use EPA Method 625 (a method used for Clean Water Act analyses) which is essentially the same method, unless YOU have permission from Ecology to use Method 625, tell them they MUST use Method 8270 or you will have to select another lab.

Will the laboratory commit to providing sufficient quality control data associated with your analyses for your use, and use by other users of the data, to verify that the lab met performance criteria required by the method?

As part of any analytical test method, specific quality control procedures are required to make sure results are unbiased and precise for the batch in which your samples were tested. The methods, for example almost always require testing of…

Blanks. A blank is a sample that should not contain the contaminant of concern. It could be pure water, or clean sand, for example. Blanks are run to detect contamination and other analytical problems. Very few blanks have results of "zero" as would be preferred…random error, which cannot be avoided but should be controlled, can cause positive and sometimes negative blanks, but blank results should usually not exceed the method detection limit identified in the method. The lab should provide blank results at no extra cost to you… good laboratory practices (GLP) and the laboratory accreditation program require that blanks be analyzed in each batch.

Standards. A standard is a sample containing a known amount of contaminant. The lab should analyze at least one standard in each batch and keep track of results for that standard over several batches. By calculating the average value of those repeated analyses and comparing it to the true value, the lab can detect bias, an indication of how close they can be expected to come to the true value under ideal conditions. Most methods give a range of what is allowed for this bias. The lab should provide the result to you for the standard it analyzed in the batch in which your sample(s) was run at no extra cost… like the blank, analyzing a standard is required by GLP and lab accreditation. Even though it is not required to do so, a reputable lab with nothing to hide would also provide their average result for repeated analyses of that standard for your use in checking for bias. If the lab does not understand what you mean when you ask for the average result, you might consider seeking another lab. If the lab is willing to tell you their average result for repeated analyses of the standard, they should also give you the standard deviation of those same results. The standard deviation is an indication of precision, or how close the results are to each other, and the expected standard deviation is usually given in the method. The lower the standard deviation, the better the precision.

Duplicates. A duplicate is a sample (usually a real environmental sample, but sometimes a standard) that has been divided into two equal aliquots, each of which is analyzed in exactly the same manner. The results of the duplicate pair are often expressed as "relative percent difference," or %RPD which is the difference between the two results divided by the average result, and expressed as a percent. The %RPD is another indication of the lab’s precision and the method usually specifies what you should expect for that value. The lower the %RPD, the better the precision.

The blank, standard, and duplicate are the quality control test results you should expect to get from the lab at no extra cost. You should be aware that there is no SINGLE test that can be used to prove that the lab is doing accurate work. The lab can analyze the standard once and get the exact true value…but that does not mean precision is good and that they can be expected to get the correct answer…or even close to it…the next batch, or even on the next sample, which might by yours. Only by looking at both precision and bias estimates can you judge whether or not the lab is likely to have done accurate work.

There is one other quality control sample that is often incorrectly used as an indicator of accuracy (remember, there is no ONE test that can indicate accuracy) and that is a matrix spike (MS). The matrix spike is an environmental sample to which a known amount of contaminant is added. The sample is analyzed WITH the spike, and WITHOUT the spike, and the difference between those two results is compared to the amount of contaminant spiked into the sample. That amount divided by the amount added and expressed as a percentage is called the percent recovery, or %R. 100% recovery on a matrix spike would be ideal, but just because a lab has a low percent recovery does not mean the lab did not do a good job. If the lab had a good result on their standard, and a low recovery on the MS, it probably means the matrix (the medium the sample was composed of, such as surface water, or soil) is interfering with the analysis. This information is of interest to the user of the data because it may cause the user to require the lab to use a different extraction, cleanup, and/or analysis to try to overcome this matrix interference. If you want a matrix spike, and perhaps a matrix spike duplicate done on YOUR sample(s), you should expect to pay for it. Before doing that, however, make sure you have a plan for use of the resulting data… it will not do you any good to get results for a matrix spike, finding that recovery is low, and then not doing anything about it… and doing something about it can be expensive.

Another statistic that may be of use to you in choosing a lab is the lab’s Method Detection Limit (MDL). The MDL is a figure derived by experiment in the lab to determine the smallest concentration the lab can reliably detect and be reasonably sure of the concentration. Expected MDLs can be found in the method, and if a lab tells you they cannot meet that published MDL, unless there is a very specific reason (your sample matrix could be a reason), you should choose some other lab that can. A reputable lab should be willing to share data from the current MDL study as evidence that they can actually meet that MDL. Another statistic associated with the MDL is the reporting limit (RL). The RL is a multiple of the MDL, usually five times, sometimes ten times. Results reported between the MDL and the RL should be interpreted as meaning the contaminant is definitely there, but the concentration is questionable (this is not an indication of problems at the lab…it is just an indication that random error has a very significant effect at those low concentrations). Results above the RL mean the contaminant is not only there, but the lab is very confident of the reported result.

Does the laboratory properly dispose if its dangerous waste?

Analytical laboratories fall under the same regulations that you do when it comes to disposal of hazardous waste. In addition to the chemical wastes generated in performing the analyses, any excess amount of your waste sample may also need to be managed as dangerous waste. To avoid possible litigation from mismanagement of your samples by the laboratory, it is incumbent upon you to require the lab to certify that it will dispose of the samples according to existing regulations, OR return the samples to you for disposal.

Does the laboratory report holding times (i.e., date and, if necessary, time the sample was received, and when it was extracted (for organics) and analyzed?

As a requirement for establishing "chain of custody" which is important should your data be challenged in court, the lab should provide a form (some large companies have their own form) for your use showing when and by whom the sample was taken. When the sample is delivered to the lab, a sample receiving person should sign the form and give a copy to your representative showing the date and time the sample was received by the lab. You should insist that the lab’s report back to you show when the sample was extracted (if an organic analysis is being done), and when it was analyzed. This information can be critical in a legal case…even a mediocre lawyer is smart enough to have data thrown out of court if it cannot be shown that the samples were analyzed within holding times, making the entire, expensive sampling and analytical process worthless. Laboratory accreditation makes sure accredited labs are capable of meeting holding times.

In summary, choosing an accredited lab assures you that the lab CAN provide reliable accurate results; it is up to you to make sure they use their capability to do that. And the way you do that is to demand the results of quality control tests done with the batch in which your samples were analyzed. If there are any questions concerning interpretation of the QC results, or in the quality of the analytical work done by a lab, contact Ecology’s Lab Accreditation Section at (360) 895-6149, fax (360) 895-6180, by writing Ecology LAS, P.O. Box 488, Manchester, WA 98353, or e-mailing the Lab Accreditation Unit.

Please send comments or questions to WebMaster