1. Introduction
Radon, a known human lung carcinogen [
1], is a colorless, odorless, and radioactive gas that naturally forms when uranium breaks down in soil, rock, and water [
2]. It can travel through rock fractures and soil pore spaces, entering buildings through cracks in the foundation or vents, where it can accumulate to dangerous levels, particularly in lower levels of buildings [
2]. Health Canada’s 2012 Cross-Canada Survey found that concentrations exceeded the national guideline of 200 Bq/m
3 in approximately 7% of Canadian residential homes [
3].
The pooled analysis of eleven studies with residential radon measurements identified several characteristics that increased indoor radon levels [
4]. They observed that radon exposures were higher in newer homes, bungalows (i.e., one-floor buildings excluding trailer mobile homes), homes with a forced air heating system, homes using geothermal heating, homes with a full basement, and homes with a private well water source [
4]. Newer buildings and rural communities with a higher proximity to groundwater wells were found in other studies to also have higher radon levels [
5,
6]. Potential remediation and prevention methods exist to substantially lower radon levels in buildings [
7]. No information on workplace characteristics was identified.
A study of over 7000 Canadian federal workplaces found that exposure levels ranged from <15 to >2500 Bq/m
3, and exposure averages differed by region [
8]. Approximately 2% of these workplaces exceeded the national guideline of 200 Bq/m
3 [
8]. The authors speculated that the lower exposure levels measured in federal workplaces compared to national residential homes in the same geographical area may reflect differences in building construction and ventilation rates [
8]. They also found that the ratio of the proportion of workplaces to the proportion of homes exceeding the national guideline was higher in Ontario (0.47) than the national average (0.27), which may reflect differences in provincial building codes [
8]. A recent study of twenty-one workplaces in Nova Scotia, Canada, reported that 10% of workplaces (i.e., two small, unventilated workplaces) exceeded 150 Bq/m
3 [
9]. The study recommended that future research focus on small workplaces (i.e., fewer than 25 workers) without mechanical ventilation [
9].
Utilizing the federal building study results, Brobbey et al. estimated that 3.3% of Canadian workers (i.e., 603,000 workers) were exposed to radon at levels exceeding the World Health Organization (WHO)’s reference level of 100 Bq/m
3 [
10]. They reported that the educational services; professional, scientific, and technical services; healthcare and social assistance services; and public administration industries had the highest number of workers exposed to levels above 100 Bq/m
3 [
10]. However, workers in mining, quarrying, and oil and gas extraction had the largest proportion of workers exposed to high radon levels above 800 Bq/m
3 [
10]. Administrative assistants, general office support workers, receptionists, and elementary and kindergarten teachers were occupations with the largest number of workers estimated to be exposed to radon [
10].
While there have been studies modeling radon exposure in Canadian workplaces [
10,
11], there are limited studies measuring radon in Canadian small/medium-sized workplaces [
9]. This is the first large-scale survey of non-federal and non-mining workplaces in Canada. There are also insufficient studies evaluating the differences by building characteristics and effectiveness of remediation methods in workplaces and how they can affect radon levels. We aim to characterize radon exposures in public buildings and small businesses in Ontario, Canada. This will fill the knowledge gap on occupational radon exposure in the province.
2. Materials and Methods
2.1. Selection of Municipalities
Provincial health regions were divided into ‘high-exposed’ and ‘low-exposed’ regions based on the national guideline of 200 Bq/m
3 [
7]. This is the level when remediation efforts are recommended to reduce radon concentrations below 200 Bq/m
3 in Ontario workplaces whenever possible [
12]. Potential background radon levels were estimated based on the federal government’s cross-Canada survey of residential homes [
3]. ‘High-exposed’ regions were defined as regions where more than 10% of the residential homes tested had radon levels above 200 Bq/m
3. ‘Low-exposed’ regions were defined as regions where less than 10% of the homes had radon levels above 200 Bq/m
3.
A total of ten municipalities were selected: six in high-exposed regions and four in low-exposed radon regions. The municipalities were also spread out geographically within Ontario to capture varying geologies. The ten municipalities included in this study were as follows: Windsor, Chatham, Woodstock, Brantford, Guelph, Toronto, Elliot Lake, Sudbury, Kingston, and Ottawa.
Sudbury and Elliot Lake are situated on the Canadian Shield, which comprises 61% of the province of Ontario, including most of the sparsely populated North [
13]. The Canadian Shield comprises Precambrian rock and contains large mineral deposits [
14]. Elliot Lake was of particular interest in this study because of extensive uranium deposits near the city. The remaining eight cities—Windsor, Chatham, Woodstock, Brantford, Guelph, Toronto, Kingston, and Ottawa—are in the southern geographic region of Ontario known as the St. Lawrence Lowlands. This area is composed of sedimentary limestone, shale, and sandstone [
14]. There are fewer uranium deposits than in the Canadian Shield but small pockets do exist in southwestern Ontario, overlapping with the Windsor and Chatham regions [
15].
2.2. Workplace Eligibility and Recruitment
Public buildings and small-to-medium-sized businesses were recruited to voluntarily participate in this study using a variety of communications including email, phone, social media, and in-person communication. The research team began by reaching out to partners in Public Health Ontario, local public health units, Workplace Safety & Prevention Services, and the Public Services Health & Safety Association (PSHSA). This study was described to representatives from each organization, and they shared the information by newsletter when possible. PSHSA pursued an extensive phone-calling and email campaign to workplaces in their mailing list that met the study criteria. The research team also reached out to local chambers of commerce and business improvement areas for each municipality to explain this study so that they could share study details if they wished. Using Google Maps, the team searched for keywords of workplaces known to be eligible for participation (e.g., dental offices, tax filing services) and the street view of the business to confirm that the location was on the ground floor or basement. If the workplace had a website, this was searched for contact information. These workplaces received an email using a generic recruitment template and/or a phone call using a similar script. Social media messages were posted roughly once per month on X (Twitter) during the recruitment periods. Lastly, research staff conducted door-to-door recruitment using a script similar to the email and phone calls in densely populated downtown areas and near already-recruited workplaces to maximize efficiency in recruitment.
To be eligible, the workplace had to have workers performing work tasks on the ground floor or basement. Workplaces had to have normal air circulation. Places like restaurants, grocery stores, and gyms with fans or machinery generating high amounts of air movement were considered low-risk and excluded, as these locations most likely had excellent air exchange. Local, independent, and non-franchised locations were prioritized as these workplaces more often had decisionmakers present to agree to participating and met the criterion of being a small-to-medium-sized business. Workplaces also had to agree to allow research staff access to the workplace to set up radon monitor(s) and place them for the three-month testing period.
2.3. Monitor Deployment and Laboratory Analysis
Radon testing took place between November and June in 2019–2023 following the Canadian National Radon Proficiency Program (C-NRPP) and Health Canada Guidelines [
7]. Radon exposures were measured using passive alpha track radon monitors (Model: AT-101 Alpha Track Detectors) provided by Radiation Safety Services, Inc (RSSI). The radon foils arrive in the RSSI lab as large CR-39 sheets made from the same batch of plastic. To calibrate the foils, they are cut to size, etched with serial numbers, and a sample is analyzed without any radon exposure to determine the background already present. Then, a sample of new foils placed in radon monitors are sent to the Bowser-Morner lab and exposed in a chamber to known radon exposure. Based on the analysis of the foils with the known exposure, an exposure conversion factor is calculated for each batch of CR-39. Radon monitors were placed at a workplace for at least three months by trained staff with certification from the C-NRPP. Monitors were deployed inside sealed, breathable Tyvek bags to prevent monitor tampering. For smaller workplaces (any location with three or fewer rooms that met the testing requirements), only one radon monitor was deployed on a desk or shelf, away from vents and at a minimum of 30 cm from an exterior wall. The radon monitors were also kept a minimum of 10 cm from other objects (i.e., books, picture frames, computer monitors) to allow for normal airflow around the monitor (
Figure 1). For larger workplaces, the number of samples deployed was decided based on the number of rooms, number of occupants, duration of occupancy, and proximity to exterior walls if underground (another method of radon entry) by the research staff. Radon testing was prioritized in basements and only placed in rooms that were occupied for more than four hours a day (i.e., no supply closets, bathrooms, kitchens, etc.). A photo was taken of the radon monitor after placement and this was compared to a photo after the three-month deployment time to confirm that the monitor was not moved.
Due to the COVID-19 pandemic, monitors at several workplaces (n = 117) were left in place for longer than the usual deployment period of 3 months. The longer testing period was still within the manufacturer’s recommendations and the radon measurement was calculated according to the extended testing period.
For quality control, duplicate monitors were deployed for 10% of the samples and transportation blanks were deployed for 5% of the samples to account for radon contamination during shipment. Duplicate monitors were placed side by side, <10 cm apart. The radon monitors were shipped in individually packaged radon-proof mylar foil and stored in a low-radon environment until deployment to prevent radon exposure before placement in workplaces.
Radon monitors were collected and mailed to the Radiation Safety Institute of Canada (RSIC) for anonymization prior to shipment to the RSSI laboratory for analysis. The detector element of the radon monitors, the foils, was made of CR-39 plastic and laser-engraved with a unique serial number. The foils were chemically etched overnight in a potassium hydroxide solution until golden brown and matching with a calibrated foil to reveal alpha particle tracks. These tracks were then counted with a semi-automated system where a microscope counted six fields and the serial number on each foil. Using calibration factors, the average radon level was then calculated from the track density observed by the microscope, exposure conversion factor, and the number of days the detector was exposed. Microscopes at RSSI are checked daily for quality control. The microscopes count the foils multiple times to ensure that the values are consistent. Unused radon monitors from RSSI are also sent to Bowser-Morner, where they are spiked with an unknown radon exposure. The actual radon level and measured value from the radon monitors must be within ±25% difference for laboratory accreditation.
The limit of detection (LOD) for the radon monitor used in the analysis was 4 Bq/m3 for the deployment time frame used (minimum of 3 months). The RSSI laboratory reported the average radon concentration in units of Bq/m3 and the total detector exposure in Bq-h/m3 for each device. Radon monitors with broken seals, damaged filters, or those that were moved from their original testing location were marked as invalid and excluded from the analysis.
Following testing and analyses, workplaces were informed of their radon levels. Workplaces with exposure levels above the Health Canada guideline of 200 Bq/m3 or above the WHO recommended maximum of 100 Bq/m3 were provided with resources on radon mitigation methods with their exposure results.
2.4. Workplace Details, Building Characteristics, and Background Radon Levels
Workplace study participants filled out a ten-minute online questionnaire using SurveyMonkey at the start of testing, which collected information on the workplace details and building characteristics (questionnaire 1). A second twenty-minute online questionnaire was deployed at the end of testing in February-to-June (questionnaire 2). For workplaces that had an extended sampling duration due to the initial 2020 COVID-19 restrictions, a supplemental COVID-19-specific survey was requested that collected information on any changes in traffic and air flow during the lockdown period.
The background radon zone was estimated from the Radon Potential Map of Ontario [
16] (
Figure 2), based on multiple sources including geological, aerial, water, and indoor air analysis. The map provides three levels of radon risk: guarded, moderate, and high. These were renamed as low, medium, and high for easier interpretation. Sample sites were located on the map and assigned their corresponding background radon zone.
2.5. Quality Control and Data Analysis
The percentage difference between duplicate samples was calculated and compared to the acceptable range in the Health Canada guide (see
Table S1: Allowable variances for duplicate) [
7]. If within the acceptable range, the samples were averaged for analysis. If the samples were outside of the accepted range, they were omitted from analysis. Samples that were not detectable (i.e.,
17].
The data were checked for normality using the Shapiro–Wilk test/QQ plot and log-transformed to calculate geometric means (GM) and geometric standard deviations (GSD). Because data were not normal after log transformation, non-parametric statistic tests were performed. Descriptive statistics were calculated including counts and percentages for categorical variables and means, medians, and ranges for continuous variables. Non-parametric statistics were performed to compare exposures between groups. The Wilcoxon–Mann–Whitney test was used to compare differences between two categories, and the Kruskal–Wallis test was used to compare exposure levels between three or more categories. The Spearman correlation was used to evaluate relationships between two continuous variables. Data were analyzed using R version 4.1.2.
3. Results
Overall, 453 workplaces were recruited from ten municipalities (
Figure 3). While questionnaire 1 achieved a high response rate of 98.7%, the rate for questionnaire 2 was lower at 68.2%. Workplaces tended to be located in areas with a high geological radon risk, have windows, be accessible from the outside of the building, not have an elevator, and have forced air heating (
Table 1. To see all variables, refer to
Table S2: Full summary of characteristics of the 453 participating workplaces). Most workplaces were either from retail (30.2%), health services (29.6%), and other service (21.5%) industries.
Of the 751 samples placed in workplaces, 4 monitors were not collected because of workplace withdrawal from this study, 29 were lost, 13 were located in workplaces that closed down, 1 monitor was in a workplace that moved location, 2 monitors were moved to an inappropriate testing location, 14 had broken seals, and 1 had a damaged filter. A total of 687 valid radon samples were obtained from the 453 workplaces. Most (72.4%) workplaces had one monitor, 18.8% had two monitors, and 8.8% had three or more monitors. The mean sampling duration of monitors was 157.9 days (range 90–422 days, IQR 96–214 days). Four samples exceeded the recommended time frame and one sample had a missing retrieval date. These five samples were retained for the analysis because the results still fell within normal among the radon measurements.
The average radon exposure across the 687 samples was 40.2 Bq/m
3, ranging from <4 to a maximum of 566 Bq/m
3 (AM = 40.2 Bq/m
3, GM = 26.9 Bq/m
3, GSD = 2.31). In 2.5% of workplaces, at least one measurement exceeded Health Canada’s limit of 200 Bq/m
3, while 7.2% of workplaces exceeded WHO’s reference level of 100 Bq/m
3. The geometric mean of workplaces was positively and significantly correlated (ρ = 0.81,
p = 0.005) with the geometric mean of residential homes in each municipality using census division data from the 2024 Cross-Canada Survey of Radon from the Evict Radon team [
18] (
Figure 4).
Radon exposures were negatively correlated with workplace size (number of employees: r = −0.31,
p < 0.0001; square footage: r = −0.20,
p < 0.0001) but not with building age (r = −0.01,
p = 0.8). Radon exposures varied significantly by city, background radon zone, and industry, where higher levels were observed in Elliot Lake, in high-background-radon zones, and in the Finance and Public Administration industries (
Table 2). Radon exposures were also significantly different when stratified by various workplace characteristics including the presence of an elevator, window-opening behavior, and business access type (To see all variables, refer to
Table S3: Full statistical summary of radon results for all monitors placed in the OCRC radon study, stratified by various workplace characteristics). Exposures did not significantly differ by monitor location (i.e., basement vs. ground floor). When evaluating the reported presence of different types of openings into the basement, the presence of cracks in the foundation walls, a humidity/air conditioning drain, a gap in the foundation, and cracks in the foundation were found to be associated with differing radon levels.
A sensitivity analysis was performed by restricting to samples no longer than 141 days (excluding the workplaces that had monitors in during the COVID-19 lockdown). Patterns did not change significantly.
Of the 453 workplaces, 7% (n = 32) had at least one radon monitor in the basement and ground floor. In a sub-analysis of paired samples, 91% (n = 29) of workplaces had higher radon exposures in the basement compared to the ground floor (Wilcoxon signed-rank test result
p = 0.000006).
Figure 5 illustrates radon concentrations for each paired workplace sample, stratified by background radon zone category.
4. Discussion
This study aimed to characterize radon exposures in public, small-sized, and medium-sized workplaces across Ontario. Elevated radon levels were observed in some workplaces, with exposure levels varying by city, background radon zone, industry, and some building characteristics. These characteristics included workplace size, presence of building additions, building access type, window-opening behavior, heating fuel type, and the presence of air conditioning, an elevator, and foundation openings.
Our findings are consistent with Health Canada’s study, which found that 2.1% of federal workplaces in Ontario exceeded 200 Bq/m
3 [
19], with similar exposure averages (AM = 40.2 Bq/m
3 vs. 33.2 Bq/m
3, GM = 26.9 Bq/m
3 vs. 20.4 Bq/m
3, respectively). No other studies have reported workplace exposure levels in Ontario. Two studies conducted in Latvia [
20] and Finland [
21] reported similar exposure levels; however, five studies—in Missouri, US, Spain, Italy, Greece, and Nigeria—reported higher levels [
22,
23,
24,
25,
26,
27]. Arithmetic means were almost three-times higher in the Greek study [
26] and over seven-times higher in the Nigerian study of university offices [
27] compared to our findings. Two studies—Japan and Australia—reported lower levels, half and a quarter, respectively, of the arithmetic means of our study [
28,
29]. These differences in workplace radon levels between countries may be due to geological variations, making direct comparisons challenging. Factors such as differences in soil permeation, ventilation methods, or building construction materials and practices could contribute to these inconsistencies.
This study also observed differences in exposure across Ontario, over 1 million square kilometers, approximately twice the size of Spain. Previous work has shown that background risk fluctuates widely in Ontario from high- to low-background-radon zones [
16]. This variability aligns with our findings of differing exposure levels by background radon zones. Similarly, a study by Ruano-Ravina et al. in Spain, which measured workplace radon levels in six regions and five different sectors, also found that exposure levels varied by geographical area, with ‘medium radon-prone areas’ having the highest exposures [
30]. A study by Song et al. of underground subway police officers in South Korea also observed regional variations [
31]. They speculated that the subway station with the highest radon concentrations was attributable to its location near granite rocks and very airtight construction of the offices [
31].
This study observed a moderate correlation between workplace radon levels and residential home levels using Evict Radon’s 2024 Cross-Canada Survey of Radon [
18]. In Reste’s study, Latvian workplaces had lower radon concentrations compared to residential measurements [
20]. Whicker also found lower radon concentrations in US federal buildings compared to residences in the same county [
32]. There are a few feasible explanations for these variations, such as differences in building codes and construction practices by region. Materials used in commercial buildings may vary from those in residential buildings and inner construction materials were found to affect radon levels in Ruano-Ravina’s study [
29]. Workplaces tend to have higher air exchange rates, which can reduce radon accumulation, compared to homes [
32].
This study found significant differences in radon levels between industry groups, with the highest average levels in public administration and finance/insurance/real estate sectors. A Spanish study [
29] found the highest radon concentrations in the health sector, followed by public administration, and other sectors/the private sector. A British study of workplaces in radon-affected areas found the highest geometric mean in miscellaneous, office, and retail workplaces, which is also similar to the findings of this study [
33]. Differences by industry may reflect differences in the types of buildings or usage patterns.
The finding of a higher exposure in smaller workplaces was expected as smaller workplaces may have less-sophisticated ventilation systems. Only one study of US federal buildings reported radon levels by size. Their findings differ, reporting no differences in radon levels by workplace size [
23]. That study, however, covered workplaces in the Department of Energy that were larger in size. The average area of the second-smallest grouped rooms was almost 13,000 square feet, which is much larger than the median workplace size in this study of 2000 square feet (IQR = 1000–4500 square feet) [
23]. The types of workplaces and size disparity may contribute to different findings.
The presence of an elevator in the building was associated with lower radon levels. A study of Bulgarian kindergartens reported similar findings, where they found workplaces with elevators had about half the radon compared to workplaces without elevators (GM = 113 Bq/m
3 vs. GM = 200 Bq/m
3) [
34]. It is possible that the stack effect of warm air rising and then exiting through the higher floors of a building can increase air flow, which could increase the dissipation of radon from the lower levels of a building and contribute to lower radon levels [
7]. Buildings with elevators may also be larger in size and this study found that larger buildings had lower radon levels.
This study found varying radon levels based on window-opening behavior and business access type. A study of Italian workplaces estimated that the ‘frequent entrance/exit of persons or mechanical ventilation’ can create considerable air exchange, where the difference between working hours compared to the whole day are different by 20% on average [
35]. Workplaces that have business access to the outdoors have the highest geometric means; however, this is slightly counterintuitive as it should be increasing air exchange. This difference might be because workplaces that open to the inside of the building may keep the doors open during the winter, since no cold air is coming inside. The buildings that open to the inside of another building are more likely to be larger buildings, which was found in this study to be associated with lower radon levels. No differences were found in radon levels for any renovation in this study except for adding an addition, which likely increases the overall volume of air inside of a building available for exchange.
This study found no significant difference in radon levels from different heating systems but did find a difference between heating fuels. No studies were identified that evaluated this in workplaces; however, two studies reported conflicting results in homes [
36,
37]. Average radon levels in this study were lower in workplaces that had air conditioning. Studies have shown that radon levels are lower when air conditioning systems are on compared to when they are off [
38,
39]. This effect was likely not observed as the testing in this study took place during the winter except for the radon monitors (n = 117) left in place for a year during the COVID-19 lockdown. Overall, keeping a building’s heating, ventilation, and/or air conditioning system on helps to reduce radon levels and may also be a key component in radon reduction [
40,
41].
Buildings with cracks in the foundation and foundation wall, humidity drains, and gaps in the foundation all had lower geometric means compared to not having the opening. This was unexpected because the radon literature commonly suggests that radon enters a building via these cracks, gaps, and drains and that these are then sealed in the process of radon reduction [
3,
8]. However, there were no scientific papers that looked at this in workplaces. The questionnaire was completed by owners or managers of the workplace and it is possible that they were unaware of the condition of the basement, as many workplaces rented (62%) and did not own the building within which their workplace was located. This may have led to misclassification of the basement opening status. Future studies would benefit from conducting walkthroughs and inspections of basements for openings rather than relying on participants. This would also ensure consistency across workplaces.
The radon measurements in basements were not significantly different compared to measurements on the ground floor. However, when looking within participating workplaces with measurements on both floors, the basement measurement was higher than the ground floor over 90% of the time. This is in line with what is known in the literature, as radon moves from the soil into the lowest floors of a building [
3,
8]. Other studies in a hospital and school have similar findings of higher radon readings in basements [
42,
43]. These two studies each only sampled in one workplace and compared findings within the building, which means that all other factors like geography, building construction materials, etc., remained the same and the only different factor was the floor. This was similar to the analysis in our study within the same buildings that compared measurements between floors. The other factors affecting radon levels could play a greater role in determining radon levels and that was why the pattern was not statistically significant when comparing all radon measurements in basements vs. the ground floor. This was the case in Reste’s study, where many workplaces were tested and the radon concentrations by floor were not statistically different [
20].
This study had several limitations. Firstly, this was a voluntary study that was limited to a moderate number of participating workplaces in a few industries (limited to mostly those in health services, other services, and retail, which only covers a small percentage of all industries in Ontario). This number was still substantial given that the COVID-19 pandemic halted recruitment and the long radon monitor placement time resulted in some losses of monitors. Further stratification by background radon risk and various building characteristics, however, made group sizes very small. The voluntary nature of this study also meant that workplaces with known issues may have been more hesitant to participate. Secondly, the response rate for questionnaire 2 was poor compared to questionnaire 1, which limited analysis of certain building characteristics. Thirdly, the interruption of the 2020 COVID-19 shutdown in the Ontario region resulted in drastically different radon monitor deployment periods. Radon is typically tested for in the winter months when doors and windows are less-often opened, resulting in potentially higher radon levels [
44]. Year-round averages for the radon monitors deployed for a longer period of time could be slightly lower and we were unable to correct for these small differences in the results. Lastly, this study did not assess the potential impact of building and furniture materials. Small businesses in Ontario are mostly furnished with wooden or metal furniture and inspection of photos taken of the monitor placement at each workplace did not reveal any granite or other materials that could be a potential source of radon.
This study recruited smaller workplaces in different industries, which provides a greater range of radon measurements in different Ontario workplace types compared to previous results that were only in federal buildings. This study also asked questions about building characteristics that have not been previously studied in workplaces. There were good quality-control results from the duplicates and blanks, which also increases confidence in radon results in workplaces. However, it is important to recognize that these results may not be generalizable to other jurisdictions and may be limited to buildings in Ontario and Canada due to the unique geography and history of Canada. Additional measurement data are needed in untested regions with more thorough questionnaires or workplace inspections to understand if patterns persist in other Ontario regions and identify more building characteristics that are associated with high radon levels.
This study contributes to the limited knowledge of radon exposure in Ontario’s non-mining workplaces, highlighting key geographical and building characteristics that influence radon levels. While most workplaces had radon concentrations below the Health Canada guideline, a small percentage exceeded the recommended limits, demonstrating the need for continued monitoring and potential mitigation in high-risk areas. The variability in radon levels across different workplace types, regions, and building characteristics emphasizes the importance of tailored radon management strategies. Given the geology and building practices across Ontario, the findings may not be generalizable to other jurisdictions. Future studies should focus on expanding workplace radon measurements in other regions, and further exploring the relationship between specific building features and radon concentrations to better inform public health and safety regulations.
