Analysis of the analytical performance of laboratories participating in two major U.S. bulk asbestos proficiency testing programs

Our plain-English summary of Analysis of the analytical performance of laboratories participating in two major U.S. bulk asbestos proficiency testing programs, by B.W. Harvey, W.G. Winstead (Journal of Occupational and Environmental Hygiene, 2008).

The study

• Authors: B.W. Harvey, W.G. Winstead
• Journal: Journal of Occupational and Environmental Hygiene
• Published: 2008;5(7):455-62
• Cited by: 1 article, as counted by NIH iCite
• Read it: PubMed · Publisher

Abstract

Quality assurance and proficiency testing programs have been available for 27 years for U.S. laboratories using polarized light microscopy to analyze bulk building materials for asbestos. Total enrollment in the two principal bulk asbestos proficiency testing programs conducted by RTI International peaked at nearly 900 laboratories in the mid-1990s. Enrollment has stabilized in the last 5 years at approximately 475 laboratories, and rates of return of analysis results currently exceed 95%. More than 115,000 test samples have been sent worldwide by RTI to laboratories in these two programs. A review of more than 109,000 analysis results was undertaken to determine the frequencies of various qualitative error types, the types of bulk building materials most responsible for those errors, and the accuracy of precision of semiquantitative results. This assessment revealed that a small number of bulk building materials cause the majority of difficulties in each program, although error frequencies vary between the two programs. Overall laboratory performance has improved since the programs first began in the late 1980s, with a significant decline in qualitative errors and a much less dramatic decrease in errors associated with semiquantitative estimation of the amount of asbestos present.

In plain English

This was not a study of people's health. The authors, from RTI International, which runs the testing, reviewed archived results from two U.S. proficiency testing programs. Participating laboratories use polarized light microscopy to check bulk building materials for asbestos. These were the NIST bulk NVLAP (35 rounds) and AIHA program (69 rounds), through November 2006. Overall sample classification error rates were 3.4% (NVLAP) and 3.1% (AIHA); identification error rates were 1.8% and 6.8%. A few difficult samples caused most errors. Specific difficult samples with low chrysotile percentages or short chrysotile fibers had false negative rates of 17% to 65%; other positive samples had 2% or less. Polyethylene fibers and wollastonite caused 33% (NVLAP) and 56% (AIHA) of false positives. Samples containing actinolite, anthophyllite or tremolite caused 39% (NVLAP) and 49% (AIHA) of identification errors. Laboratories tended to overestimate asbestos amounts, more so at low percentages. On repeated difficult samples, error rates fell in 8 of 10 cases. The authors noted that test results probably underestimate errors in routine work. The authors concluded that participation, presumably with more analyst training and experience, has improved laboratory capabilities.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Retrospective review of archived laboratory proficiency testing results (not a health study)
• Who was studied: Laboratories enrolled in two programs. The NIST bulk NVLAP began in 1989 and has sent more than 53,000 test samples to laboratories in the U.S. and Canada. The AIHA Bulk Asbestos Proficiency Analytical Testing Program began in 1989 and has sent more than 62,000 test samples to laboratories in the U.S., Canada, Europe and the Far East. Combined enrollment peaked at nearly 900 laboratories and was about 475 at the time of the article.
• Exposure: Not a human exposure study. Laboratories analyzed test samples of bulk building materials. Some contained asbestos (chrysotile, amosite, anthophyllite, tremolite, actinolite). Others contained non-asbestos look-alikes (polyethylene, wollastonite, anhydrite, Libby winchite and richterite). The study counted false negatives, false positives, asbestos identification errors and errors in estimating the amount of asbestos.
• Main results:
  • Overall error rates were as follows. Classification false negatives: 3.0% in NVLAP and 3.0% in AIHA. False positives: 5.0% in NVLAP and 3.3% in AIHA. Overall classification: 3.4% in NVLAP and 3.1% in AIHA. Asbestos identification: 1.8% in NVLAP and 6.8% in AIHA.
  • A small set of specific, difficult samples had false negative rates of 17% to 65%. These samples had low percentages of chrysotile or short chrysotile fibers, and included spray-on coatings, vinyl floor tiles and short-range chrysotile materials. On all other positive samples, the false negative rate was 2% or less (1.8% in NVLAP and 2.0% in AIHA in Table III).
  • Polyethylene fibers and wollastonite cleavage fragments caused 33% of false positives in NVLAP and 56% in AIHA.
  • Samples containing actinolite, anthophyllite and tremolite had identification error rates of 5% to 66% on specific samples. These samples accounted for 39% of all asbestos identification errors in NVLAP and 49% in AIHA.
  • Laboratories tended to overestimate asbestos content in samples where the asbestos was easily detected. Overestimation was twofold to threefold for samples with 5–10% asbestos. It was threefold to eightfold for samples with trace levels to 4% asbestos, and one sample exceeded tenfold.
  • On known difficult samples used more than once, the error rate was lower on later use in 8 of 10 sample/program combinations. The average decrease was 19%, ranging as high as 72%.
  • For samples repeatedly used to test estimates of asbestos amount, 15 of 18 consensus means were lower on subsequent use (t-test p = 0.0005). Variation between laboratories (CV) was higher on subsequent use for 9 of the 12 AIHA samples.
• The authors' conclusion: The authors concluded that participation in bulk asbestos QA and PT programs, presumably together with greater analyst training and experience, has improved overall laboratory analysis capabilities. They found that improvement in semiquantitative analysis was less dramatic than in qualitative analysis.
• Limitations the authors state:
  • Error rates in proficiency testing probably underestimate error rates on routine samples, because laboratories likely give test samples more attention.
  • The programs differ in their laboratories, requirements and sample difficulty, so the authors intentionally avoided direct comparisons between them.
  • Lower error rates over time may reflect real improvement, attrition of weaker laboratories, or both.

Mines named here

Asbestos mines this study is about. Each mine's page says who ran it, when it worked, and what it produced.

• Libby vermiculite mine, Montana

All asbestos mines

Topics

Medical terms in this study, explained in our Lexicon: Polarized Light Microscopy (PLM).

Related studies

• Analytical performance issues. Two different methods, three different variations, four different sources: what bulk asbestos PLM method is your laboratory using? Part 2 — D.B. McGrath, D.R. Van Orden, et al. (2010)
• A comparison of fibre counting across three European national proficiency testing schemes — A.D. Jones, M.C. Arroyo, et al. (2005)
• A proposal for harmonising laboratory performance assessment criteria in national asbestos fibre counting schemes — M.C. Arroyo, J.M. Rojo (2001)
• A PC-based quality control (QC) program for an industrial hygiene laboratory — D.K. Verma, K. Ferguson, et al. (2006)
• Letter: Hazards in laboratories — C.L. Lavelle, A.D. Landman (1976)
• Mortality among sheet metal workers participating in a respiratory screening program — L. Welch, J. Dement, et al. (2015)

Browse more research in Journal of Occupational and Environmental Hygiene, or return to medical research.