Respirable dust and respirable crystalline silica exposures among workers at stone countertop fabrication shops in Georgia from 2017 through 2023

Abstract

Objectives

This longitudinal study examines the severity of worker exposure to respirable crystalline silica (RCS) and respirable dust and demonstrates the need for increased education and implementation of both appropriate engineering controls and respiratory protection (RP) programs for stone fabricator shops, given the growing global number of accelerated silicosis cases associated with the fabrication of engineered stone (ES) countertops.

Methods

Personal air sampling results and detailed job description notes obtained from 17 industrial hygiene air sampling visits conducted at 11 stone fabrication facilities between 2017 and 2023 in Georgia were used to align similar exposure groups (SEGs) for tasks for workers performing stone fabrication. Bayesian decision analysis was used to determine appropriate RP selection recommendations for the 4 proposed SEGs: SEG1—Support, SEG2—Automated Tool Operator, SEG 3—Small Tool Operator, and SEG 4—Fabrication/Lamination.

Results

The analysis concluded that all employees in stone fabrication shops that process ES should wear a respirator with a minimum assigned protection factor (APF) of 10, regardless of the engineering controls in place. For SEG 4, it is recommended that workers use respirators with an APF between 50 and 1,000. Among the 75 full-shift personal air samples for RCS dust, 41 samples (53%) exceeded the permissible exposure limit of 50 µg/m³.

Conclusions

This is the first study to present the 4 SEG categories with sampling data to support the importance of including all employees (even support workers) in RP programs, exposure monitoring, and medical surveillance.

Recommendation and implications

Employers, occupational health professionals, and inspectors may use these SEG categories and corresponding RP recommendations to determine if employees have received appropriate RP for workers at stone countertop fabrication shops.

What’s Important About This Paper?

This study used respirable crystalline silica exposure data to define 4 similar exposure groups among workers fabricating engineered stone. This study recommends respiratory protection programs for all workers in this industry to supplement engineering controls, owing to the high variability in exposures and observed failures of control technologies.

Introduction

Respirable crystalline silica (RCS), a toxic material, is well-known to cause silicosis, lung cancer, chronic obstructive pulmonary disease, a wide range of autoimmune disorders, and kidney disease (Thomas and Kelley 2010). Although the incidence of silicosis has generally decreased over the last century due to an increase in the usage of control measures, the incidence of cases globally has increased by 64.61% from 1990 to 2019 (Thomas and Kelley 2010; Huang et al. 2024). An estimated 96,366 workers in the United States workers were employed in the stone fabrication industry in 2018, which included the manufacturing and processing of both natural and engineered stone (ES) (sometimes also referred to as artificial stone or quartz) products, primarily as countertops (or benchtops) (Rose et al. 2019). Employees working with ES in this industry, specifically in the post-manufacturing phase known as fabrication, are very often employed by small shops where slabs are customized for installation in kitchens and bathrooms. Employee exposures occur during processing, which includes cutting, drilling, grinding, and polishing of the slabs by stationary (fixed) equipment such as bridge saws or computer numerical control (CNC) machines and hand-held power tools. An analysis of OSHA inspections from the Emphasis Program for RCS in Engineered Stone found that 68 out of 332 samples collected, exceeded the OSHA permissible exposure limit (PEL) of 0.05 mg/m³ as a time-weighted average (TWA) for RCS, and 1,059 citations were issued to this targeted industry (OSHA 2024a). These exposure level findings mirror those documented in the recent literature (Glass et al. 2022; Surasi et al. 2022; DeVaughn et al. 2024; Weller et al. 2024).

An ES slab is markedly different from a natural stone slab as the RCS content of ES products can be ≥ 90% compared to natural marble (<5%) or granite (10% to 45%) (Ramkissoon et al. 2022). ES is comprised of finely ground quartz silica and polymer resins, and due to its durability, cost, and customizability, it has exploded in popularity; between 2010 and 2018, United States imports of ES slabs increased by 800% (Gandhi et al. 2023).

Studies have revealed that ES fabrication employers are failing to adequately adhere to the ancillary provisions required by the OSHA RCS Standard (1910.1053), with air monitoring demonstrating exceedance of the OSHA PEL (Surasi et al. 2022; DeVaughn et al. 2024). In California, 72% of employers inspected under the 2022 Special Emphasis Program (SEP) were issued citations, with common violations including failure to provide adequate respiratory protection (RP), conduct training, implement required medical surveillance, and utilize engineering or work practice controls to reduce exposures to or below the PEL (Surasi et al. 2022). Worker interviews conducted by Cal/OSHA under this SEP in which employees self-reported tasks and use of controls or protective equipment were analyzed by Spiegel et al. (2022). Of the 92 workers who completed this questionnaire, 75% reported wearing RP for more than 30 days/year, which is the threshold under 29 CFR 1910.1053 for employer-mandated medical surveillance. Only 20% of respondents reported completing respiratory fit tests in the prior 12 months of work (as mandated by 29 CFR 1910.134), and only 5% reported being sent for required RCS medical examinations (Spiegel et al. 2022).

OSHA citation statistics illustrate that employers, in general, are not fully and properly applying the OSHA Respiratory Protection Standard (29 CFR 1910.134), a regulation that was the seventh most frequently cited standards in 2023 among all industries (OSHA 2024b). OSHA RP citations jumped from 1,895 (totaling $2.4 million in fines) in 2019 to 2,475 (totaling $4.4 million in fines) in 2023 (Houlroyd 2024; OSHA 2024c, 2024d). This uptick in RP citations indicates that increased education and guidance regarding RP selection and use must be provided in industries with high risk of inhalation hazards, such as stone fabrication shops.

This study evaluated employee exposures to RCS and RD at stone countertop fabrication shops that voluntarily requested air sampling surveys from the Georgia OSHA 21(d) Consultation Program to obtain compliance with the OSHA RCS standard (1910.1053). The OSHA 21(d) Consultation Program is funded by the United States Department of Labor and provides no-cost, confidential compliance assistance to small businesses upon their request. It is designed to improve the safety and health of employees throughout the country by assisting employers in complying with OSHA standards. Companies served by 21(d) Consultation are typically ≤ 250 employees physically onsite and often lack full-time professional safety and health staff. All shops fabricated a combination of ES and natural stone slabs using wet and dry methods. Most countertop fabrication shops are small, and workers often complete various tasks, making it difficult for employers and employees alike to navigate the complexity of exposure groups and appropriate respirator selection. This study aims to (i) analyze RCS air sampling data collected from 11 Georgia stone fabrication shops during 17 exposure assessments completed between 2017 and 2023 and then (ii) use air sampling data results and job descriptions from these exposure assessments to align similar exposure groups (SEGs) for tasks performed during the process of stone fabrication, and (iii) use Bayesian decision analysis (BDA) and the Industrial Hygiene Data Analyst tool (IHDA-AIHA) to project probability employees in the various SEG categories would need different assigned protection factors (APFs) of RP equipment. IHDA-AIHA allows inputting a prior and likelihood estimate to project posterior Bayesian decision charts and estimate appropriate exposure control categories, in this case, to estimate RP equipment selection. The novelty of the present study is to use empirical exposure measurement data to characterize silica exposures and integrate qualitative observational data of observed respirator use/type with personal exposure data to determine the appropriate APF, which provides the ability to evaluate the differences between actual versus appropriate selection. The proposed methodology using BDA with and without informative data to determine the minimally acceptable RP selection for stone countertop fabricators was the first time adopted in this stone countertop industry and follows an approach for controlling RCS exposure similar to Table 1 from the OSHA construction RCS standard 29 CFR 1926.1153(c)(1).

Table 1.

Summary of ES fabrication shops sampled.

Site	NAICS Code	Employees in establishment	Employees monitored	Primary fabrication methods as listed in field notes	Respiratory protection worn as listed in field notes
1	423320	≥25	5 full-shift samples 2 samples were < 3 hours	Polishing, CNC operation, bridge saw operation, support, and mock installation/demolition. A mixture of wet and dry methods.	3 employees conducting dry cutting wore N-95s 4 employees wore no respirators
2	327991	≥25	8	Polishing, cutting/ saw operator, fabrication, and quality control support. Wet methods used for all tasks	Employees monitored wore N-95s
3	327991	≥25	7	Bridge saw operator, polishing, and fabrication with hand tools. Wet methods used for all tasks.	Employees monitored wore N-95s
4	203429	1 to –25	2	All hand tools for polishing and fabricating slabs. Wet methods used for all tasks.	No respiratory protection worn
5	327991	≥25	3	Polishing and fabricating using a mixture of hand tools and CNC machines. A mixture of wet and dry methods.	½ mask elastomeric respirator worn by all employees
6	327991	≥25	24	Polishing, CNC operation, bridge saw operation, and support activities. All tasks conducted in the main building were done wet and dry methods for fabrication/lamination were conducted in a room with water curtain dust extractors.	Progression from no respirators worn by wet polishers to the use of N-95s. Progression from ½ mask elastomeric respirator to full-face respirator use for fabricators/lamination
7	327991	1 to –25	4	Handling slabs and conducting fabricating tasks (including saw cutting and polishing). Wet methods used for all tasks.	No respiratory protection worn
8	327991	1 to –25	5	Polishing, CNC operators, fabrication tasks. Wet methods used for all tasks.	½ mask elastomeric respirators worn by all employees
9	327991	1 to –25	4	Robotic saw operation, polishing, and fabrication tasks. Wet methods used for all tasks.	No respiratory protection worn
10	327991	1 to –25	7	Polishing, CNC operators, fabrication tasks. Wet methods used for all tasks. Dry methods used in adjacent areas by contractors.	N-95s worn by employees on a voluntary basis
11	327991	1 to –25	6	Polishing, CNC operators, fabrication tasks. Wet methods used for all tasks.	Two employees wore N-95s on a voluntary basis, and no respiratory protection worn by other workers

Methods

Respirable dust and respirable crystalline silica sampling and analysis

Data for this study were selected from stone fabrication companies that requested air sampling from the OSHA 21(d) Consultation Program in the state of Georgia between 2017 and 2023. Table 1 summarizes basic information about each of the 11 stone countertop fabricators in Georgia that requested this service. Personal air sampling for RCS was conducted by OSHA 21(d) Consultation Program industrial hygienists on 77 employees.

Representative full-shift air monitoring for RSC and RD was performed in the breathing zone (defined as the zone 6 to 9 inches from the nose and mouth of the employee) of employees performing fabrication tasks. Calculated TWAs represent the time sampled, as opposed to the CAL/OSHA or OSHA 8-h TWA method that assumes zero exposure during any unsampled time (Surasi et al. 2022).

Dorr-Oliver nylon cyclones (Sensidyne, Clearwater, FL, USA) at a sampling flow rate of 1.7 L/min or GK2.69 cyclones (BGI Inc., Waltham, MA, USA) at a flow rate of 4.2 L/min were used over the course of this 7-year sampling period. These cyclones were connected to battery-operated pumps using either the SKC AirChek Essential Pump, Gilian 5000/3500 pumps, or Gilian 10i pumps. Samples were collected on 5 µm pore size 37-mm polyvinyl chloride (PVC) (GLA5000, SKC Inc.) assembled in conductive polypropylene cassettes to comply with International Standard ISO 24095:2021 and minimize particle deposition on sampling cassette walls (Soo et al. 2014a, 2016; ISO 2021). Flow rates were set and verified prior to sampling, and the rates were re-measured after sampling using a BIOS DryCal Meter (BIOS International Corporation, Butler, NJ, USA) to confirm that they did not change significantly (all remained within ± 5%) (Soo et al. 2014b; ISO 2022).

Gravimetric analysis of the RD was performed by weighing the filters before and after sampling using the method National Institute of Occupational Safety and Health (NIOSH) 0600, followed by analysis for RCS, which was performed by X-ray diffraction by an American Industrial Hygiene Association (AIHA) accredited laboratory (Wisconsin Occupational Health Laboratory; WOHL) according to NIOSH Method 7500 and OSHA ID-142 (NIOSH 2003; OSHA 2016). The reporting limit used by WOHL for RCS was 10 µg until 2019 and then was reduced to 5 µg. Prior to 2019, sampling results for RCS only represented quartz silica concentrations; however, starting in 2019, all crystalline silica polymorphs were included in the RCS value reported, including quartz, cristobalite, and tridymite. The reporting limit for RD remained constant at 53 µg for RD for all samples.

Personal air sampling data for this study were extracted from the OSHA Information System database (OIS) to create a dataset of RCS and RD air sampling results from consultation visits at stone fabrication shops; these air sampling results were then matched to confidential client reports to review RP use and detailed job descriptions for each task performed. A separate dataset was created with all company identifiers removed before analysis. Detailed descriptions of the specific slab composition processed during each consultation visit were not available; however, a wide range of types of ES and natural stone slabs were fabricated at each site, with site 1 being the only site that primarily fabricated natural stone slabs.

Data analysis

Personal sampling results were analyzed using JMP software, Version 16 for Windows (SAS Institute, Cary, NC, USA). RCS and RD data were transformed using the logarithm prior to analysis. Geometric means (GM), geometric standard deviation (GSD), and confidence intervals were calculated. The exposure profiles of RD and RCS were based on the sampling strategy recommended by the AIHA Exposure Assessment Strategies Committee (Damiano et al. 1998). The lognormality, the GM, and its corresponding 95% confidence interval for each selected exposure group were computed. Each exposure group’s exposure profile was examined using lognormality using (i) the log-probability plot by the goodness-of-fit (Fillibens test) or (ii) the normality of log-transformed data using the Shapiro–Wilk test.

Workers’ RD and RCS exposure profiles by SEGs

Following the steps for aligning SEGs as outlined by Bullock et al. (2015), industrial hygienists used detailed job descriptions from consultation reports to assign employees into 2 initial SEG categories: (i) automated tool operators (such as the computer numerical controlled [CNC] machine, bridge saw, and robotic arms saws) and (ii) fabricators using hand tools or performing material handling (Bullock et al. 2015; Zwack et al. 2016). These SEGs were refined into 4 SEG categories (Support, Automated Machine Operator, Small Tool Operator, and Fabricators/Laminators). The 4 SEGs incorporate aspects of the 3 exposure groups identified by Seneviratne et al. in their assessment of fabrication shops in Australia (2024) and the employee air sampling results and observation of work processes:

1. Support (SEG1): Employees performing tasks outside the scope of countertop fabrication, such as materials handling, working in shipping and receiving, performing quality control, and housekeeping.

2. Automated Machine Operator (SEG2): Employees operating or programming automated machines that are water-fed such as CNC routers, saws, water jets, robotic arm saws, miter saws, and a variety of automated polishers.

3. Small Tool Operator (SEG3): Employees operating hand tools outfitted with automated water delivery systems observed to be in use, such as angle grinders, polishers, miter saws, and circular saws.

4. Fabricators/Lamination (SEG4): Employees performing fabrication tasks dry or lamination, a finishing process where 2 slabs are joined, or an edge is affixed to a slab using an adhesive that must cure. During this time, cutting, sanding, or polishing is often done dry to protect the curing process.

Bayesian decision analysis

BDA was used to project what APFs would be necessary to protect ES fabricators from exposure to RCS for each of the SEG categories. The process used for the BDA analysis is outlined in Fig. 1. Air sampling results below the reporting limit, considered censored data, were adjusted based on maximum likelihood estimation for the analyses (Hewett and Ganser 2007). Measured workers’ RCS exposure concentrations were used to estimate the probability that workers may be exposed to different maximum use concentrations (MUCs) based on tasks. For an “employer’s choice” prior, the percentages of different types of respirators observed in use during site visits were used to establish the APFs distributions required to protect workers (OSHA 2009). A uniform prior was used in a secondary analysis. The measured personal exposure results of the RCS concentrations were used to establish the likelihood. Together, the above 2 estimations (ie prior and likelihood exposures) were used to further estimate the resultant posterior APF distributions of workers. The MUC for a selected respirator was calculated using the formula: MUC = APF × OEL, with the occupational exposure limit (OEL) for RCS set at 50 µg/m³. In this study, the APF was classified into 5 categories: APF 1—no RP needed (representing exposures < 50 µg/m³), APF 10 required (exposures 50 to < 500 µg/m³), APF 50 required (exposures 500 to < 2,500 µg/m³), APF 1000 required (exposures 2,500 to < 50,000 µg/m³), and APF greater than 1,000 required (any exposure expected to be greater than 50,000 µg/m³), respectively. The RP selection charts created using the BDA analysis are presented in Figs. 2 and 3 and were developed using the IHDataAnalyst software, Version 1.37.08 (Exposure Assessment Solutions Inc., Morgantown, WV, USA).

Fig. 1.

Bayesian decision analysis methods for the development of minimally acceptable respiratory protection selection for stone countertop fabricators.

Fig. 2.

The prior, likelihood, and posterior APF distributions for respiratory protection recommendation among the 4 SEG categories and the entire ES dataset using the “Employer’s Choice” as a prior. Note that censored data (n = 3) were adjusted based on maximum likelihood estimation for analyses using the BDA.

Fig. 3.

The prior, likelihood, and posterior APF distributions for respiratory protection recommendation among the 4 SEG categories and the entire ES dataset using a uniform prior. Note that censored data (n = 3) were adjusted based on maximum likelihood estimation for the analyses using the BDA.

Results

Of the 77 employees monitored, 2 were excluded from the final analysis because the samples were collected for less than 3 h. A total of 75 employees were monitored for a full shift: 21 (28%) were below the OSHA action level (AL) of 25 µg/m³, 15 (20%) were above the AL but below the PEL of 50 µg/m³, and 39 (52%) were at or above the PEL. A total of 10 of the 11 sites had at least 1 employee exposed to RCS above the PEL, with 4 sites having all monitored employees exposed to RCS above the PEL. Table 2 summarizes the TWA exposure ranges for RD and RCS: exposure levels for RCS ranged from below the reporting limit to 5,100 µg/m³. The GM for each site varied between 22.6 and 941 µg/m³ (GSD = 1.39 to 6.06). Respirable dust exposure ranged from below the reporting limit to 8.9 mg/m³ (OSHA PEL 5.0 mg/m³ 8-h TWA), with the GM for each site ranging from 0.162 to 4.89 mg/m³ (GSD = 1.33 to 6.13). Less than 4% of all measurements for RCS were below the reporting limit (n = 3), with the censored data adjusted based on maximum likelihood estimation for analyses. The log-transformed RD and RCS mass concentrations exhibited a strong correlation, with a Pearson correlation coefficient of 0.844 (P < 0.05). In addition, both data sets followed a lognormal distribution, as confirmed by the Shapiro–Wilk test (P < 0.05). For 22 RD samples below the reporting limit, 8 were at or above the AL for RCS.

Table 2.

Results of RCS and RD collected from personal air sampling by site: number of personal dust samples collected by site and visit, RCS and RD exposure ranges and concentrations [geometric mean (GSD)] and the estimated crystalline silica mass fraction [arithmetic mean (standard deviation)] obtained from personal air samples for each company site reports.

Site	Sampling visits conducted	Number of employees monitored	Exposure ranges for RCS (µg/m3)	Exposure ranges for RD (mg/m3)	RCS mass concentration (µg/m3)	RD mass concentration (mg/m3)	Crystalline silica mass fraction (%)
1	1a	5	ND* to 32	ND* to 1.9	22.6 (1.63)	0.768 (3.72)	1.41(0.382)
2	2a	8	36 to 200	0.15 to 0.61	91.7 (1.78)	0.329 (1.54)	28.5(6.55)
3	3a	3	81 to 200	0.46 to 0.99	111 (1.39)	0.536 (1.33)	21.1(4.14)
	3b	4	91 to 150	0.41 to 0.52
4	4a	2	35 and75	0.23 and 0.48	35, 75	230, 460	15.8 (0.77)
5	5a	3	180 to 2100	0.72 to 8.9	4.89 (3.63)	2.03 (3.72)	24.1 (0.75)
6	6a	5	15 to 2000	ND* to 4.7	70.3 (6.06)	0.396 (6.13)	36.1 (17.2)
	6b	8	16 to 230	ND* to 0.36
	6c	6	26 to 5100	ND* to 6.8
	6d	5	18 to 30	ND* to 0.97
7	7a	4	68 to 200	0.29 to 0.49	127 (1.59)	0.350 (1.40)	36.7 (6.02)
8	8a	5	78 to 370	0.2 to 0.88	133 (1.83)	0.385 (1.72)	34.9 (5.38)
9	9a	4	9.6 to 120	ND* to 0.57	36.5 (3.17)	0.246 (2.77)	23.2 (2.93)
10	10a	3	16 to 76	All ND*	26.3 (2.15)	0.162 (1.64)	14.8 (8.32)
	10b	2	26 to 44	0.18 to 0.25
	10c	2	9.2 to 14	ND* to 0.095
11	11a	6	16 to 50	ND* to 0.17	28.5 (1.48)	0.170	20

^*Non-detect (ND) listed for sampling results below the limit of detection for the sampling period.

Table 3 summarizes and compares the 2 SEG classification systems used to assess tasks for fabricators. Regardless of the SEG classification method adopted, the designation of the SEG being fabricator for SEG_initial or fabricator for lamination SEG_final showed the highest RCS exposures (GM = 80.4 vs 279.0 µg/m³) and was statistically significantly different (P < 0.05) when comparing the SEG_initial with the SEG_final. The classification of operator for SEG_initial versus the Automated Machine Operators SEG_final had the lowest GM exposure concentration (GM = 39.2 vs 36.4 µg/m³). Using the RCS profiles in SEG_final for illustration (Table 3), the exceedance fraction of workers being exposed to RCS above the PEL (>50 µg/m³ 8-h TWA PEL) is 51.8% for support workers (SEG 1), 37.9% for automated machine operators (SEG 2), 56.4% for small tool operators (SEG 3), and 80.0% for those performing fabrication or lamination (SEG 4), which often includes periods of dry fabrication. All employees performing tasks for SEG2 and SEG3 used wet methods. Six (50.0%) employees in SEG 4 performed all tasks dry, and the remaining employees in SEG 4 reported using all wet methods.

Table 3.

Personal RD (mg/m³) and RCS data (µg/m³) among different similar exposure groups (SEG).

	RD (mg/m3)					RCS (µg/m3)
	GM	GSD	>PEL Compliance statistics	Lognormal Goodness-of-fit test		GM	GSD	>PEL Compliance statistics	Lognormal Goodness-of-fit test
SEGinitial
Operator (n = 17)a	0.220	2.48	ExcFrace = 0.0003 95%LCL ≤ 0.001 95%UCL = 0.024	Yes (R = 0.978)	Operator (n = 17)b	39.2	2.79	ExcFrac = 0.407 95%LCL = 0.254 95%UCL = 0.579	Yes (R = 0.939)
Fabricator (n = 58)a	0.50	3.02	ExcFrac = 0.0188 95%LCL = 0.006 95%UCL = 0.053	Yes (R = 0.941)	Fabricator (n = 58)b	80.4	3.90	ExcFrac =0.636 95%LCL = 0.549 95%UCL = 0.716	Yes (R = 0.943)
SEGfinal
SEG 1: Support (n = 3)c	0.398	2.31	ExcFrac = 0.0013 95%LCL ≤ 0.001 95%UCL = 0.581	N/A	Support (n = 3)	52.8	3.36	ExcFrac = 0.518 95%LCL = 0.181 95%UCL = 0.839	Yes (R = 0.999)
SEG 2: Automated machine operator (n = 14)c	0.190	2.50	ExcFrac = 0.0002 95%LCL ≤ 0.001 95%UCL = 0.031	Yes (R = 0.962)	Automated machine operator (n = 14) d	36.4	2.78	ExcFrac = 0.379 95%LCL = 0.215 95%UCL = 0.574	Yes (R = 0.975)
SEG 3: Small tool operator (n = 46)c	0.359	1.85	ExcFrac = 0 95%LCL ≤ 0.001 95%UCL ≤ 0.001	Yes (R = 0.987)	Small tool operator (n=46) d	57.8	2.44	ExcFrac =0.564 95%LCL=0.466 95%UCL=0.658	Yes (R = 0.974)
SEG 4: Fabricator for lamination (n = 12)c	0.159	5.11	ExcFrac = 0.240 95%LCL = 0.102 95%UCL = 0.460	Yes (R = 0.956)	Fabricator for lamination (n = 12)	279	7.70	ExcFrac = 0.800 95%LCL = 0.606 95%UCL = 0.917	Yes (R = 0.951)

^aResults lower than the reporting limit (RL) of 6-digit microbalance, were found in 41.2% (n = 7 out of 17) and 25.8% (n = 15 out of 58) of samples from operators’ and fabricators’ samplers, respectively.

^bRCS sample below RL of the NIOSH X-ray diffraction method (NMAM 7500), was found in 11.8% (n = 2 out of 17) and 1.67% (n = 1 out of 60) of samples from operators’ and fabricators’ samplers, respectively.

^cResults lower than the reporting limit (RL) of 6-digit microbalance, were found in 33.3% for SEG 1 (n = 1 out of 3), 42.9% for SEG 2 (n = 6 out of 14), 28.3% for SEG 3 (n = 13 out of 46), and 14.3% for SEG 4 (n = 2 out of 12) of samples from supports’, automated machine operators’, small tool operators’ and fabricators’ samplers, respectively.

^dRCS sample below RL of the NIOSH X-ray diffraction method (NMAM 7500) was found in 14.3% (n = 2 out of 14) and 2.17% (n = 1 out of 46) of samples from automated machine operators’ and small tool operators’ samplers, respectively.

^eExceedance fraction (ExFrac): the exceedance fraction of workers exposed to RCS/RD above the OSHA PEL.

^fLognormal distribution was determined using the log-probability plot by goodness-of-fit (Fillibens test).

Respiratory protection (RP) selection chart based on Bayesian decision analysis (BDA)

Figures 2 and 3 illustrate the prior, likelihood, and posterior APF distributions of RP selection for fabricators contained within the entire dataset and each SEG_final category. Figure 2 utilizes the “employer’s choice” RP selection prior, and Fig. 3 utilizes a uniform prior. The posterior distribution for SEG 4 in Fig. 2 projected the probability of an employee needing to wear an APF 50 was 0.993 compared to the “employer’s choice” prior probability of 0.139, indicating that the employer’s choice prior does not align with the RP APF the actual air sampling results would require. In contrast, the likelihood distributions and the posterior distributions in Fig. 3 indicated higher probabilities for requiring an APF of 50 (0.529) and an APF of 1,000 (0.471). Similar patterns of mismatched perceptions of RP needs are evident for each SEG and the whole dataset. The RP recommendations derived from BDA suggest that the barrier protection offered by filtering facepiece respirators (FFRs) such as N95 FFR or elastomeric half-mask respirators that have an APF 10 may be inadequate for stone countertop fabricators.

SEGs and approaches to the last line of defense: respiratory protection

We observed inconsistent RP use among stone countertop workers. Actual RP usage was as follows: 32 employees (41.5%) wore no RP, 29 employees (37.7%) wore N-95s (APF 10), 14 employees (18.2%) wore ½-mask elastomeric RP (APF 10), and 2 employees (2.6%) wore full-face elastomeric respirators (APF 50). Results of the BDA, using the actual RP usage as the prior, indicate that for every SEG category, employers and employees consistently underestimated (shown as the prior) the minimally acceptable RP compared to the results of the likelihood (which used the air sampling results for RCS). RP should be required for all SEGs, with the minimally acceptable APF being 10 to 50 for SEGs 1 to 3 and an APF of 50 to 1,000 for SEG 4.

Discussion

Only 6 employees (all classified as SEG 4) were observed exclusively cutting dry, with exposures to RCS ranging from 32 to 5,100 µg/m³, which is much higher than results previously published in the literature (Qi and Echt 2016; Zwack et al. 2016; Salamon et al. 2021; Weller et al. 2024; DeVaughn et al. 2024; Seneviratne et al. 2024) and indicates dry fabrication in highly uncontrolled environments leads to exposures averaging more than 43 times the OSHA PEL. Seneviratne et al.’s classification of dry finishing as an exposure group aligns with the exposures seen among this worker population (Seneviratne et al. 2024). For the employees in SEG 4 who reported using wet methods (n = 6), exposures ranged from 22 to 230 µg/m³. For the employees in SEG2 and SEG3 (n = 60) who used wet methods, 48.3% had exposures at or above the PEL (max: 370 µg/m³) and 70.0% had exposures above the AL. These results support that solely using wet methods during stone fabrication activities does not reduce exposures below the PEL (Qi and Echt 2016; Surasi et al. 2022; Weller et al. 2024).

Furthermore, 91% of the sampled fabrication shops had at least 1 employee exposed to RCS at or above the PEL, compared to the 51% of workplaces with a worker exposed above the PEL in the Surasi et al. (2022) study and the 55.6% of workplaces with a worker above the regulatory exposure limit for Australia (0.05 mg/m³ 8-h TWA) by SafeWork NSW by Weller et al. (2024). One potential reason for these higher percentages is that 4 of the 11 companies requested consultation as part of abatement assistance following a federal OSHA inspection where citations were issued for exceeding the RCS PEL. These results confirm that there is widespread overexposure to RCS for small countertop fabrication shops. The one company with levels of RCS below the PEL had a very low crystalline silica mass fraction content (1.4%) compared to the other companies (14.8% to 36.7%), potentially indicating that they were primarily fabricating a higher percentage of natural stone slabs with a lower RCS content than the high-RCS content ES slabs.

Why repeat monitoring matters: high variability and engineering control failures

Multiple visits were conducted for 3 stone fabrication shops: 2 visits for site 3, 4 visits for site 6, and 3 visits for site 10. These repeated visits were initiated either at the request of the employer or at the request of the consultant to verify the implementation of engineering controls. For site 3, even after the implementation of engineering controls, exposure to RCS remained above the OSHA PEL. For site 6, in which the 4 visits were conducted over 5 years, exposure monitoring results varied greatly, depending on the maintenance of the water curtain dust extractor, similar to the one recommended by Qi in his 2024 NIOSH report (Qi 2024). Results from the first visit to site 6 measured exposures as high as 2,000 µg/m³ when dry cutting with no ventilation. The company installed a water curtain dust extractor and established a wet-cutting policy. These changes reduced exposures to 230 µg/m³. However, when consultants returned for visit 3, the water curtain dust extractor was no longer functioning as designed, and the employees had returned to dry cutting. As a result, exposures reached as high as 5,100 µg/m³. This observation highlights the critical importance of ongoing preventative maintenance and verifying the effectiveness of control measures through air sampling. It also underscores why a comprehensive RP program must be implemented and maintained. Indeed, the OSHA standard 1910.1053 considers “failure of engineering controls to be a situation that is reasonably foreseeable,” and therefore, if an engineering control is being used to keep employee exposure below the PEL, and the control method is not sufficient, the employer is required to use them regardless and supplement with RP (OSHA 2023). The variability in air sampling results validates the importance of repeated monitoring (1910.1053(d)(3)), maintaining a comprehensive RP program for when engineering systems fail, and educating all workers about the risks associated with exposure to RCS, including those who may be exposed to precarious work arrangements, such as contract work, day-laborers, no work contract, or working for cash.

Consultants documented instances of employees wearing RP with facial hair, visibly inadequately fitting RP, improper storage of RP, and removal of RP during work tasks due to filter clogging, obstructed vision, or heat stress, similar to observations reported by Salamon et al. (2021). Furthermore, this study highlighted those employees providing support (such as those driving a forklift, supporting other workers with tasks in a “float position,” or conducting quality assurance) are often overlooked in the respiratory selection process. However, our data demonstrate these positions should be required to wear RP without successful exposure controls in such workplaces. The large percentage of workers not wearing RP aligns with Seneviratne et al.’s reporting that if the employer does not require and supervise to ensure proper use of RP, employees are likely to forego RP due to the discomforts of RP or the assumption that if wet methods are used, there is no hazard (2024).

Results from the analysis of RP use highlight the need for better employer and employee education about the importance of consistent and proper RP selection and use, including fit testing prior to performing the work, ensuring workers are clean-shaven for the use of tight-fitting respirators, and training on the proper cleaning and storage of respirators. The implementation of regulatory changes in 2019 in Australia and an increase in education and enforcement to the stone countertop fabrication shops prompted an increase in the proportion of Australian workers reporting the use of the recommended APF respirator from 44.9% to 86.5%, creating a model for OSHA to implement in the United States to improve worker RP selections (Hore-Lacy et al. 2024). Despite these improvements in RP use, silicosis cases from exposure to ES dust continued to rise in Australia, and based on a review of the scientific evidence conducted by the National Dust Disease Taskforce and Safe Work Australia the Australian government banned the use and import of ES (Cavalin et al. 2024). The United States should adopt this RP approach until comparable protections can be put in place to protect American workers and align with the current CAL-OSHA Standard for ES RCS exposure requirement, which requires a full-face, tight-fitting powered air-purifying respirator (APR) or a respirator providing equal or greater protection, which provides an APR of 1,000 (California Department of Industrial Relations 2024).

Study limitations

The limitations of this study include incomplete information, eg the details of the engineering control measures implemented and the exact composition of slabs each employee was fabricating. Realistically, sampling captured real-life exposure variability for this industry in Georgia: depending on product orders, on any given day, the majority of slabs could be ES or, conversely, natural stone or a mixture of both. While consultants documented that both natural and ES slabs were fabricated at each site, future visits could include documenting the specific slab types fabricated on the sampling date. In addition, there is some expectation of temporal variability in this data set; in addition to slab composition changes during this timeframe (trending ES fabrication increase over selected dataset), production or personnel changes, shop modifications, housekeeping measures, implementation, or discontinuation of engineering controls are all factors influencing outcomes. This study was limited to stone fabrication shops in Georgia that requested industrial hygiene monitoring for RCS, which may not be representative of the exposure of the entire United States or of other countries and may reflect selection bias for companies voluntarily requesting this no-cost assistance. Finally, although consultants offered to conduct follow-up monitoring, some companies declined, limiting the ability of the program to verify the effectiveness of the reported installation of engineering controls.

To our knowledge, this is the first study to create and analyze an RCS exposure database across 4 SEG categories for stone countertop fabrication. Most countertop fabrication shops are small, and workers often complete various tasks, making it difficult for employers and employees alike to navigate the complexity of exposure assessment and appropriate respirator selection. Our findings support the importance of including all employees—support staff and those using exclusively water-fed tools—in RP programs, exposure monitoring, and medical surveillance. The current study confirms prior findings that RCS exposures in stone fabrication shops often exceed OELs and contributes to the literature by suggesting that workers using dry methods may require substantially more protective respirators, such as those with an APF of 1,000. Following the hierarchy of controls, eliminating dry operations and requiring the use of wet methods is the logical and preferred approach to reducing RCS exposures; however, achieving this ideal has proved elusive. Weller et al. (2024) assessed Australian fabrication workers’ exposure to RCS following a 5-year awareness and compliance campaign. Even though all assessed locations used wet fabrication methods, Weller et al. (2024) found workers exposed to RCS above the Australian OEL of 0.05 mg/m³ (TWA) and stated that wearing RP remained necessary. Medical reports, occupational exposure assessments, and scientific research have all provided evidence that RCS exposure in shops primarily fabricating ES countertops leads to accelerated silicosis (Kramer et al. 2012; Leso et al. 2019; Ramkissoon et al. 2022, 2023, 2024a, 2024b; Hoy et al. 2023; Weller et al. 2024)

Gandhi et al. (2023) reported that California silicosis cases identified by active surveillance had higher predicted lung function parameters at diagnosis than those identified by passive surveillance and were more likely to have developed milder disease. If employers were at least completing medical surveillance (mandated by 1910.1053 for workers meeting the exposure criteria), we would expect that more pre-symptomatic workers could be identified. In the absence of robust public health surveillance to improve early detection of silicosis, the onus for employee protection is squarely on the employer to prevent the disease through the implementation of a robust RP program, comprehensive and consistent use of functional engineering controls (wet methods and additional ventilation systems), and a regimented housekeeping schedule.

This degree of prevention was not observed at our sampling sites, mirroring current data and trends indicating employers are not successfully implementing appropriate control measures or providing required medical surveillance to affected workers (Surasi et al. 2022; OSHA 2024c). Our results highlight significant deficiencies in employer enforcement of RP use and selection of RP that is adequate to provide sufficient protection. Employers who are unaware of the exposure levels in their workplaces are also likely unaware of the selection criteria for choosing the correct APF respirator.

Our study proposes minimizing employer uncertainty by encouraging the categorization of work activities in stone fabrication shops into 4 SEGs—each with a minimally acceptable level of required RP—to simplify the steps for appropriate RP selection. This approach mirrors the OSHA RCS standard for construction (1926.1163) and its “Table 1,” where common RCS-generating construction tasks are grouped based on widely available exposure monitoring data, providing a decision tree for employers related to engineering controls and required RP. When an employer follows Table 1, the steps they need to take for regulatory compliance and protective measures are clearly prescribed—easing the burden on the employer, especially smaller employers with limited access to occupational safety and health resources.

Conclusion

While historically most RCS exposures arising from processing common quartz-containing materials such as granite slabs or concrete have proved controllable with adequate application of water and/or local exhaust ventilation or use of appropriate RP, the results of this study demonstrate that RCS remains widely uncontrolled at shops of different sizes and sophistication where ES is fabricated. For example, the emergence of additional research into the release of nanoparticles during fabrication, interactions with resins and other volatile chemicals released during the fabrication process, and the presence of other elements used as fillers and colorants highlight the importance of exercising the precautionary principle until more is understood about the confluence of these unknown factors. Our results demonstrate that existing engineering controls provided at the investigated shops are insufficient to protect workers from RCS exposure, and all employees performing similar tasks at comparable shops should wear RP by following applicable OSHA standards until—or unless—sufficient exposure controls—either through slab composition modification or traditional engineered dust collection or suppression techniques—are improved, implemented, and documented with repeatable success.

Funding

The work to collect the data for this article was performed under funding from the US Department of Labor, Occupational Safety and Health Administration as part of the OSHA 21(d) Consultation Program grant. No funding was provided for the analysis of this data or the writing of this article.

Conflict of interest

Jenny Houlroyd has served as an expert witness in silicosis cases unrelated to this research. All other authors declare no conflict of interest.

Data availability

Data used for this analysis is available upon request.