Testing the reliability of asbestos analyses from multiple laboratories using blind bulk samples

The sporadic occurrence of contradictory lab results regarding the presence of asbestos in building materials has raised questions about the quality of routine analyses performed by labs that offer their services on the Swiss Asbestos Forum (FACH) website, which is administered by the Swiss National Accident Insurance Fund (Suva). These labs have committed to a careful preparation of "difficult" samples, such as window putty, tile adhesive and other materials containing often low concentrations of asbestos, so as to increase the likelihood of detection. Most of the labs are accredited for asbestos analysis by their respective national accreditation body, and all have regularly passed proficiency testing (PT) schemes such as the Asbestos in Materials Scheme (AIMS) of the Health and Safety Executive (HSE, Buxton, UK) or the Bulk Asbestos Proficiency Analytical Testing (BAPAT) provided by the American Industrial Hygiene Association (AIHA, Falls Church, VA). Sample sets of pre-tested building materials were shipped to these labs from unsuspicious senders at irregular intervals, approximately once per year since 2019. Error rates were high at the beginning of the program, with a maximum of 17% false negative results in the second round, decreasing to less than 5% in subsequent rounds. The rate of false positives was markedly lower, with a maximum of 3% in the first of five rounds. Overall, Swiss labs performed better than non-Swiss labs, and concentrations greater than 1% were correctly detected more often than those of lower concentrations. The method used (mostly polarized light microscopy (PLM) or scanning electron microscopy (SEM)), as well as accreditation and price of the analyses, did not significantly influence analysis outcomes. Approximately half of the laboratories provided an estimate of the asbestos content in the samples, often with significant discrepancies between individual and reference laboratories. The use of SEM analysis has led to significantly higher asbestos content estimates than PLM analysis.