Social factors and behavioural reactions to radon test outcomes underlie differences in radiation exposure dose, independent of household radon level

Radioactive radon gas inhalation causes lung cancer, and public health strategies have responded by promoting testing and exposure reduction by individuals. However, a better understanding of how radon exposure disparities are driven by psychological and social variables is required. Here, we explored how behavioural factors modified residential radon-related radiation doses incurred by 2390 people who performed a radon test. The average time from first awareness to receiving a radon test outcome was 6.8–25.5 months, depending on behaviour and attitudes. 20.5% displayed radon test urgency that reduced irradiation between awareness and outcome to 1.8 mSv from a typical 3.5 mSv, while 14.8% (more likely to be men) displayed delaying behaviours that increased exposure to 8.0 mSv. Of those with low radon, 45.9% indicated no future testing intention, underscoring the importance of original tests to reliably establish risk. Among people finding high radon, 38% mitigated quickly, 29% reported economic impediments, and 33% displayed delaying behaviours. Economic barriers and delaying behaviours resulted in 8.4 mSv/year or 10.3 mSv/year long term excess exposure, respectively, increasing lifetime risk of lung cancer by ~ 30–40%. Excess radiation doses incurred from behaviour were independent of household radon level, highlighting the strong influence of psychological and socioeconomic factors on radon exposure and lung cancer risks.


Results
Study cohort details. A description of the demographic details of this cohort, as well as radon test outcomes and property types were described recently in Ref. 20 . The total study population currently includes Canadians enrolled in the 'Evict Radon' national study, a citizen science-based radon awareness and testing program of interdisciplinary research led by investigators at multiple Canadian universities. Participants performed a long-term alpha track radon test within their primary residential property, and consented to provide de-identified radon data and selected property metrics to researchers. In this analysis, 7481 participants active in the study in 2018 were invited to answer retrospective questions (in relation to radon testing) concerning their radon test decision making, emotional reactions to test outcomes, known and intended course of actions following test outcome, personal demographics, and activity patterns as it related to time spent in the radon-tested property per year. Of these, 2390 people provided responses for age, sex, gender, relationship status, parental status, occupants per property, household income, number of years lived in the tested property, whether they had ever performed a radon mitigation on the property, and their activity patterns. Response distributions for the entire group are indicated throughout this study (denoted by ' ALL') and have also been previously described in Refs. 20,24 . As we obtained gender identity responses from a majority but not all participants, we defaulted to sex for individuals whose gender data was missing (see "Methods" section). Responses regarding sex had options for "other" or "prefer not to say" that were ultimately chosen by 1% of participants. For ease of reading, we will hereafter refer to this metric as 'sex/gender' and will use the terms 'man' , 'woman' and 'everyone else' . Our cohort comprised 42.9% women (mean age 49), 56.1% men (mean age 53), 1% everyone else (mean age 53), an average of 2.56 occupants per property, 73% parents to least one child, with 44% reporting that minors lived in the www.nature.com/scientificreports/ property at the time of surveying. As a whole (and prior to the COVID-19 pandemic), the cohort spent 5549 h/ year inside their primary residence (data was all collected pre-COVID-19 pandemic), with less time for those in full time work or education (5205 h/year), and more for those retired, telecommuting, or on leave (6200 h/year), or unemployed (6726 h/year).
Residential radon exposure periods modified by radon awareness and testing behaviour. We first defined six radiation exposure time periods of interest, across the spectrum of radon awareness to mitigation behaviors (Fig. 1A). These were: t 0 , the time before initial awareness of radon after moving into primary property; t 1 , the time between first gaining awareness of radon to obtaining a radon test; t 2 , the time taken to deploy the test; t 3 , the radon testing period; t 4 , the time between test conclusion and receiving outcomes; t 5 , time  www.nature.com/scientificreports/ after radon levels are known but before action is taken; t 6 , all periods after action is taken (including a decision to do nothing). For t 0 , we stratified people into three groups based on how long it took them to first hear about radon after moving into their primary residence. 26.4% lived in the property for > 20 years before encountering knowledge of radon gas and its health effects, and, as a group (we denoted as 'late awareness'), were older with fewer household occupants compared to the 46.8% who gained awareness within the first decade ('early awareness') or 26.8% who gained awareness within the second decade ('intermediate awareness') ( Fig. 1B). Those gaining radon awareness earlier after moving into a given property were relatively younger, had more occupants per property, were more likely to be women, and have minors living at home (Fig. 1C,D). For people who had just become aware of radon, we had previously noted 24 differences between groups in 'time to action' (t 1 ), best visualized by expressing number of months required to obtain a radon test as a function of the total number of radon awareness information interactions over the period (Fig. 1E). A majority of people (64.7%) required 2-5 interactions with radon awareness material over the course of a year before obtaining a radon test (denoted as 'typical'); this contrasted with the 20.5% who displayed radon testing 'urgency' (obtaining a test in only a few weeks to months after a single encounter), or the 14.8% who exhibited 'delaying' behaviours (requiring > 6 interactions over many years). While there were no major demographic differences between the typical and urgent radon testing behavioural groups, the delaying group were substantially and significantly (p < 0.0001) over-represented by men, and were also somewhat less likely to have minors living in the property (Fig. 1F,G). For t 2 , the majority (92%) deployed their radon test within the first month of obtaining it, with no significant (p > 0.05) demographic differences to report compared to the minority who took longer (Fig. 1H). For t 3 (radon testing period = 107.3 days) and t 4 (time to receive outcomes = 19.2 days), there were also no clearly discernable differences between behaviours or demographic groups to report (Fig. 1I,J). From this analysis, we concluded that the largest differences driven by behaviour are based on time needed to first learn about radon (t 0 ), and then the time taken to obtain a radon test (t 1 ).
Variable alpha particle radiation doses as a function radon awareness and testing behaviour. Based on the distinct behavioural groups outlined above, we next calculated doses of alpha particle radiation to the lungs (in mSv) using our activity pattern data for all individuals and a conversion formula standardized by the ICRP (see "Methods" section, Fig. 2A). Three periods were considered: the un-avoided exposure period (all time prior to radon awareness, t 0 ), the 'passive' excess exposure period (all time after general awareness of radon but before specific household radon levels are known, t 1+2+3+4 ), and the 'active' excess exposure period (all time after personal radon exposure information is known, t 5+6 ). For all participants, irrespective of behavioural group, the geometric mean rate of alpha radiation exposure due to residential radon was 3.9 mSv/ year (min = 0.08; max = 157), demonstrating that household radon levels and activity patterns were random with respect to radon awareness and testing behaviour type (Fig. 2B). However, sizable differences emerged based on behaviour when total mSv radiation exposure was considered. For the total un-avoided excess exposure period (t 0+1 ), 146.6 mSv was experienced by those who became aware of radon later after moving into the property, and this was significantly (p < 0.0001) and substantially (474%) greater than the 30.9-77.7 mSv total dose absorbed by those who gained awareness earlier (Fig. 2C). Within these groups, the only notable demographic difference to report was that un-avoided exposure doses were slightly but significantly (p < 0.05) higher in men versus women for the intermediate awareness group (Fig. 2D). For the total passive excess exposure period (t 1+2+3+4 ), although the rate of radiation exposure from residential radon was identical (3.9 mSv/year) for all behavioural groups (i.e., random), differences in the total time from radon awareness to the receipt of household radon levels were clear. People displaying typical testing behaviour took, on average, 11.9 months from first hearing about radon to obtaining a radon test outcome for their property, resulting in 3.5 mSv excess exposure (Fig. 2E,F). By contrast, those exhibiting testing urgency obtained their household radon level in just 6.8 months (during which they experienced 1.8 mSv), while those who delayed required 25.5 months to receive their radon test outcome, during which they received 8.0 mSv alpha particle radiation to the lungs. Within these groups, men over age 50 experienced higher total passive excess exposure doses versus younger men, or women of any age (Fig. 2G).
Reactions to a long-term residential radon test outcome. We next assessed how individuals were able to recollect and interpret the outcome of their radon test, and how they reacted emotionally to this information. To assess emotional reactions, we asked participants to reflect on their feelings and select an intensity based upon a standard 5-point psychometric Likert scale (as we had done previously 24 ). The majority were able to correctly recollect their responses completed at the time of surveying (Fig. 3A). Perhaps expectedly, recall was the least precise for those whose household radon level was at or below 100 Bq/m 3 , as this level has been determined to be non-impacting to cancer risk 17,18 (i.e., might be considered "unmemorable"). In terms of how this information was interpreted, all those with radon levels ≥ 500 Bq/m 3 considered this level as high or unsafe, while the overwhelming majority (97%) with levels ≤ 150 Bq/m 3 interpreted their level as low or safe (Fig. 3B). People with intermediate household radon outcomes nearer the regional, administrative radon action levels (of 200 Bq/m 3 ) had mixed interpretations regarding safety, with 23.5% of those in homes with 200-499 Bq/m 3 considering this as low or safe, and 25.8% of those in homes with 150-199 Bq/m 3 considering this as high or unsafe. In total, 19.1% of participants interpreted their radon levels as high or unsafe, and this aligned with immediate emotional reactions (Fig. 3C), wherein 16.1% of participants expressed negative emotions (anxiety, fear, anger or disgust), 12.1% declared no strong feelings, and 71.8% experienced a positive emotion (relief or confidence) in a manner that was mostly congruent with their recalled radon result (Fig. 3D). A minority of participants emotional reactions were, at face value, incongruent with their recalled radon level, with 3.2% of people expressing positive emotional reactions to 200-999 Bq/m 3 radon, and 3.1% experiencing anxiety in response to radon levels below 150 Bq/m 3 . To a small extent, this perceived incongruency correlated with age, with those reacting www.nature.com/scientificreports/  www.nature.com/scientificreports/ negatively to lower radon levels being younger or more likely to be in full time work or education, and those being more positive about higher radon being significantly (p < 0.0001) older or less likely to be in full time work or education (Fig. 3E,F).
Intentions and radiation doses experienced by people interpreting their radon as low or safe. For the 80.9% of participants who interpreted their household radon level as low or safe, we queried whether they would consider radon testing at a future date. Almost half (45.9%) of people indicated no future intention to test ('Test Only Once'), with the remainder split between 30.1% with periodic testing intentions ('Would Retest'), and 24% who would only retest after either home renovations, furnace replacement, or in preparation for a real estate transaction ('Retest with Reason') ( Fig. 4A). While there were no observable age effects  www.nature.com/scientificreports/ between women in these groups, men with no future radon testing intention were significantly (p < 0.0001) older than men who would retest (Fig. 4B); this observation correlated with the same group initially requiring a greater number of interactions with radon awareness before obtaining their first radon test-indicative of consistently slower behaviour with regard to radon-related action (Fig. 4C). A possible explanation for why some expressed an intention to retest periodically without specific reason is that they had a modest but significantly (p < 0.0001) higher annual rate of radiation exposure from radon, indicative of comparatively greater household radon and/or time spent per year in the property relative to other groups (Fig. 4D).
To understand the long-term radiation exposure outcomes for those interpreting their household radon levels as low or safe, we next calculated total lung radiation exposures across all time periods in the behaviourally variable paradigm of radon awareness and testing outlined in Fig. 1A. Overall, the entire group of households with 'low or safe' radon had an average of 3.3 mSv/year (min = 0.08; max = 46.9) radiation exposure, with 59.7% being below the 4 mSv/year (≈ 100 Bq/m 3 ) threshold for long term exposure at and above which increased lifetime risk of lung cancer risk is evident 17,18 (Fig. 4D). As no radon reduction action was taken in the context of this 'low/ safe' group (average household radon = 86.5 Bq/m 3 ), time period t 6 was excluded (Fig. 4E). The total un-avoided exposure was 35.3 mSv, based on an average of 10.7 years spent in the property prior to radon awareness, with a further 3.2 mSv during the 11.1-month passive exposure period, totalling to a geometric mean exposure of 40.9 mSv (Fig. 4F). Projecting forward at a mean rate of 3.3 mSv/year (and assuming no other variable changed), the population interpreting their household radon level as low or safe would accrue 61.5 mSv total radiation exposure from their household radon after 5 years, rising to a total of 148.9 mSv after 30 years (Fig. 4G). For context, these levels are only slightly below what would be accrued from the long term rate of irradiation from radon (4 mSv/year) that produces a 16% increase in relative lifetime risk of lung cancer 17,18 . It is important to note that 40.3% of households identifying their radon as 'low or safe' are, in fact, experiencing exposures rates in excess of 4 mSv/year, being as high as 8.0-46.9 mSv/year for most highly-exposed 5%, and receiving total doses of radiation that are predicted to increase lifetime relative risk of lung cancer over 10-30 years by a significant amount (Fig. 4D).

Stated intentions of people interpreting their radon as high or unsafe.
For the 19.1% of all participants who interpreted their household radon level as high or unsafe, we asked people to state their intended course of action (Fig. 5A). A majority (68.3%) expressed the intention to have a radon mitigation system (of any description) installed to reduce radon, with another 10.1% wishing to confer with a professional mitigator to determine how radon might be entering their home, and 12.3% desiring to confirm their radon levels via another test. A small (5.4%) group stated they might alter behaviour to reduce exposure, but otherwise take no other action, and only 3.9% of people declared they would take no action at all. When viewed in the context of radon level, most of those who stated that they would take no action (or only behaviour changes) largely fell into groups with lower radon (< 200 Bq/m 3 ), with nearly everyone with higher radon levels indicating an intention to pursue more tangible radon reduction or retesting measures. This aligned well with stated action thresholds, wherein a majority (83.2%) chose to mitigate at levels in the 100-300 Bq/m 3 range, and a small minority indicating their action threshold at very low (< 50 Bq/m 3 ) levels (4.1% people) or very high (> 500 Bq/m 3 ) levels (3.1% people) (Fig. 5B). To explore issues surrounding the socioeconomics of radon remediation, we next asked at what level of radon and how fast people would invest CAD$2500.00 in a property mitigation (Fig. 5C). This cost was chosen as it corresponded to empirically understood minimums for Canadian radon mitigation at the time of surveying. Somewhat unsurprisingly, both urgency and intentions to invest financial resources in mitigation corresponded with increased radon level, with a clear majority expressing a desire to mitigate radon levels within a week to a year of theoretically learning their property was over the regional action threshold of 200 Bq/m 3 .
Behaviours and decision making of people interpreting their radon as high or unsafe. To understand actual behavioural choices and outcomes, we next captured the actual course of action taken by participants (Fig. 5D). More than a third (38.2%) declared they had completed (or had booked) a professional radon mitigation in response to a radon test outcome they interpreted as high or unsafe; we classified this group as those who 'mitigated quickly' . Another 31.3% indicated the intention to mitigate, but that they had not yet had time; we classified this group as those who exhibited mitigation 'delaying behaviour' . A total of 28.5% of people indicated that, although they desired mitigate their property for radon, they were not able to afford this either at the time of surveying or ever; we classified this group as those with 'economic impediments' . Finally, 1.9% of people completely 'declined action' . When interpreted through the lens of socioeconomics, it was clear that the group reporting economic impediments to radon mitigation had a substantially and significantly (p < 0.0001) lower household (before tax) income versus all other groups (Fig. 5E). This group were also younger, lived with a greater number of household occupants (Fig. 5F), were more likely to have minors living in the property (Fig. 5G), but were otherwise comparable to overall cohort in terms of the balance between men and women (Fig. 5H), or the time needed to initially obtain a radon test (Fig. 5I). Those who were able to obtain a mitigation more quickly encompassed somewhat older people with comparatively higher income levels than those who could not afford this, but were otherwise comparable in terms of the balance between men and women and other variables (Fig. 5E-I). Those who exhibited delaying behaviour encompassed those with the highest overall income levels, were significantly (p < 0.0001) over-represented by men, and overlapped with those individuals who required longer to obtain a radon test (Fig. 5E-I). Similarly, those who declined action altogether were overwhelmingly men, older, had fewer people living in the property and, originally, were those who took the longest time to obtain a radon test (Fig. 5E-I). For context, at the time of surveying, the average regional household income (before tax) was approximately CAD$118,000.00 according to Statistics Canada 31  www.nature.com/scientificreports/ www.nature.com/scientificreports/ CAD$100,000.00 and CAD$200,000.00, the people participating in this study broadly reflected the income levels of the general population, albeit probabilistically somewhat skewed towards middle-income earners (Fig. 5E).

Radiation doses experienced by people interpreting their radon as high or unsafe.
To understand the long-term radiation exposure outcomes for those interpreting their household radon levels as high or unsafe, we next determined the amount of time between radon outcome and mitigation (t 5 ), and the exact reduction in radon level that was achieved. For those that 'mitigated quickly' , an average 156.8 days was needed (after receiving a test outcome) to complete a mitigation that reduced their properties geometric mean radon level from 313.5 to 19.2 Bq/m 3 (Fig. 6A,B). The radon level for those who exhibited radon mitigation delaying behaviour or economic impediments remained high at geometric means of 262.4 and 216.7 Bq/m 3 , respectively. obtaining a radon test outcome to completion of a radon reduction mitigation, for those who interpreted their primary residential radon test outcome as 'high/unsafe' and who obtained a radon mitigation. (B) Residential radon levels for the behavioural groups outlined in Fig. 5D,E, including post-mitigation radon levels verified for those who 'mitigated quickly' (blue dots). (C) Annual rate of lung radiation exposure from radon for the behavioural groups from (B). Rates for the 'mitigated quickly' group were derived from post-radon mitigation levels. (D) A schematic indicating the time periods (t 0 to t 6 ) influencing radon gas exposure in residential properties for which radon levels were interpreted as 'high or unsafe' with three potential long-term outcomes depending on radon mitigation behaviour type. t values = geometric means. (E) Scatter plots of total lung radiation experienced by individuals who identify their radon levels as 'high/unsafe' for indicated exposure periods. Mean values refer to geometric means. (F) Based on current ('now') total exposures (calculated as the sum of exposures from (E)), and variable behavioural responses modifying future exposure (i.e., radon reduction or not), a projection of future total doses of radiation to lung from residential radon over a 5-30-year period. (G) Line graph of total lung radiation exposure data from (E) relative to long term exposure of those interpreting their household radon as 'low/safe' . www.nature.com/scientificreports/ Overall, and prior to any mitigation, the entire group of households with 'high or unsafe' radon had an average of 9.3 mSv/year (min = 0.6; max = 157.8) radiation exposure, with 91.8% being above the 4 mSv/year (≈ 100 Bq/ m 3 ) threshold at which increased lifetime risk of lung cancer risk is evident 17,18 (Fig. 6C). Those who were able to mitigate quickly reduced their annual exposure to 0.75 mSv/year. It is worth emphasizing that this level of exposure is well below the 4 mSv/year threshold of risk, and was substantially (440%) lower than the 3.3 mSv/ year experienced by those identifying their household radon as 'low or safe' (Fig. 4D). We next calculated total lung radiation exposures across all time periods in the behaviourally variable paradigm of radon awareness, testing and observed mitigation outcomes outlined in Fig. 6D. The total un-avoided exposure was 79.3 mSv, based on an average of 8.5 years spent in the property prior to radon awareness, with a further 8.6 mSv during the 11.1-month passive exposure period (Fig. 6E). For those who mitigated quickly, 4.4 mSv was experienced during the period between test outcome and mitigation, and their long-term exposure at 0.75 mSv/year resulted in a projected 5-year total radiation exposure of 128.5 mSv, rising to just 159.9 mSv after 30 years (Fig. 6F). For individuals that exhibited mitigation delaying behaviour or who declined action entirely, their long-term exposure was, on average, 10.3 mSv/year and resulted in a projected 5-year total radiation exposure of 176.6 mSv, rising to 447.3 mSv after 30 years. People unable to afford radon mitigation continued to be exposed to 8.4 mSv/year over the long term, which resulted in a projected 5-year total radiation exposure of 127.7 mSv, rising to 349.2 mSv after 30 years. When examined collectively, the behaviourally driven excess radiation exposure was largest for those with high radon and who delayed/declined action, being > 300 mSv greater compared to those who mitigated quickly or had lower radon (Fig. 6G). Those with economic impediments precluding mitigating their homes for radon experienced a > 250 mSv total excess exposure. Of note, while people who ultimately found that their homes contained 'high/unsafe' radon had already absorbed a far greater un-avoided radiation dose to their lungs at the time of first radon awareness (relative to those with low radon), this difference is predicted to be erased for those able to opt for mitigation, becoming indistinguishable from the low radon exposure group over time.

Radiation exposure from radon as a function of integrated behaviours over time. Our find-
ings so far indicate that there are three major behavioural types influencing the time periods that substantially modify absorbed radiation exposure due to residential radon. In this paradigm of behaviourally modified radon exposure (Fig. 7A), the major considerations are (i) how long it takes someone to first become aware of radon after moving into a property, (ii) the time needed to obtain a radon test and outcome, and (iii) whether or not (and how long it takes) an individual to reduce their radon exposure following a high or unsafe radon outcome. We next considered these variables holistically, and in relation to household radon levels, to determine if these integrated behaviours operated independently of the innate risks of radon exposure produced as a function of geology and the built environment. Strikingly, there was no significant difference in the radon level of the primary residence for the groups split apart by time to first radon awareness and radon testing behaviour, either relative to one another, or the geometric mean for the cohort as a whole (Fig. 7B). However, despite household radon levels being comparable, these groups experienced significantly (p < 0.01 to 0.0001) different total lung radiation exposures (Fig. 7B) based on their distributed behaviour types (Fig. 7C). By further factoring how individuals in these behavioural groups interpreted their household radon test outcome, further differences in total radiation exposure (at the time of radon test outcome) were evident, even for those with lower levels, and especially for those with high radon (Fig. 7D).
Finally, to estimate the added longer-term impact of radon reduction behaviours, we projected absorbed radiation doses forward over two decades for those who did or did not reduce high/unsafe radon, relative to the low radon group (Fig. 7E). Doing this, the importance of pursuing or being able to afford radon mitigation following a high radon test outcome was emphasized, as those who mitigated were able to ameliorate their exposures in a manner that offset previous behavioural effects. For example, an individual who became radon aware late (> 20 years after moving into their property), who then also delayed radon testing (taking > 2 years to get a test following awareness), and then who experienced a high radon test outcome but was able to mitigate their property for high radon quickly was, after a further 20 years, able to reduce their overall exposure to near that of a person who became radon aware earlier in life, who tested urgently, and then experienced high radon but did not (or were not able to) mitigate (317 versus 256 mSv, respectively). Collectively, these results confirm that social, economic, and behavioural differences between people are significant variables that modify radiation exposure from radon beyond the intrinsic exposure risks that are linked to and understood from household radon level information.

Discussion
A key takeaway message from this work is that behavioural and social variables (particularly socioeconomic factors) have a strong influence on realized doses of radiation absorbed by the lungs as a consequence of radon in the residential built environment. These effects also operate in addition to how radon exposures are shaped by region (geology, climate, etc.), and the built environment (house age, type, etc.). This concept is important to consider in the context of population attributable risk of lung cancer from radon exposure, which has largely only been considered as a function of a property or regional radon level (and not group or individual level behaviours). As indicated earlier, 100 Bq/m 3 radon exposure equates to ~ 4 mSv/year and carries the minimum, statistically significant increase (of 16%) in relative lifetime risk of lung cancer over the long term, and this value has been used in many regions as a reference point to derive administrative action levels (e.g., in Canada and the EU, the action level was set at twice where an elevated risk of cancer is first evident, i.e., 200 Bq/m 3 radon). Risk estimations in this context are harmonized across populations irrespective of age, sex and gender, income, activity pattern, etc., and so are less suited for nuanced estimations of lung cancer risk at an individual level. To www.nature.com/scientificreports/ do this in a way that does not over or under-estimate risk, we suggest that mSv measures of radiation exposure be used, as these account for activity patterns and, in light of the information we present in this study, can also include and better reflect the clearly important impact of behaviours. Personalized estimations of lung cancer risk from radon are increasingly important in a practical sense, as low-dose computed tomography is a clinically effective lung cancer screening tool being realized in many health systems, including Canada 2,32-34 . Two of the largest trials to date demonstrate a 20-24% reduction in lung cancer mortality using this approach 35,36 , with eligibility for screening derived from a ≥ 1.5% risk of lung cancer within a 6-year period. However, this eligibility criteria currently only applies to a proven history of heavy tobacco smoking (≥ 20-30 pack years). Indeed, light-, remote ex-, and never-smokers (and others with significant exposures to lung carcinogens) are currently excluded from lung cancer screening on the basis of insufficient evidence for www.nature.com/scientificreports/ benefit vs risk to support their inclusion based on health economics-this is an unaddressed health equity issue in all lung cancer screening programs. As it is anticipated that these programs will have a substantial positive impact on lung cancer outcomes, it is important to try and widen inclusivity while not jeopardizing their cost-benefit value. For radon to be used as a future lung cancer risk factor that confers screening eligibility, it will be very important to derive risk that holistically accounts for residential property radon level, overall human activity patterns, and the impact of behaviour on past, current and projected doses of alpha particle radiation to the lungs. One striking observation we made was that, for those who interpreted their properties radon levels as 'low or safe' , almost half declared no future intention to radon test again. In our study, long term (> 90 day) alpha track radon tests that were quality controlled by blanks, duplicates and spikes were used to ensure highly reliable readings with very little probability of false negative or positive outcome (verified and discussed in detail in Ref. 11 ). However, in the more diverse context of global radon testing, it is important to consider that short term tests are available and even prevalent in some regions (e.g., the USA), and carry a proven 11 much greater risk of false negative outcomes (i.e., the test appears low, while actual exposure is high). As half of those receiving a low reading will not test again, it follows that the long term consequences of any false negative radon test outcome on health will be amplified, potentially across a lifetime. Thus, we suggest that radon test reliability is an especially important consideration that all those offering radon tests to the public need to be concerned about. In this case, sensitivity, precision and accuracy should be prioritized over speed.
Some of the larger effects we documented in the behaviourally driven paradigm of radon exposure (Fig. 7A) occurred in a sex/gendered-, income-, and/or age-biased manner. It is important to recognize that our cohort encompassed people who served in the role of primary radon test decision maker for an entire household, in which an average of 2.56 people lived full time, including 44% of households with minors. Therefore, this is a scenario where the health-seeking behaviour of an individual can have wider implications on a group of people with potentially no or only indirect input into the decision. We consistently found that men (as the primary radon test decision maker) were more likely to become radon aware later in life and, once aware, require far more time and interactions with information before performing a test. Men were also more likely to delay or decline action following what they interpreted as a high or unsafe radon outcome, independent of their household income.
Hence, households where men are directing radon test decision making are, as a whole, likely to be exposed to alpha radiation from radon for longer, incurring greater risks over the long term. This effect was, in part, skewed by age, with men who were under age 50 (at the time of surveying) displaying improved health-seeking behavior (or, at least, distinct structural determinants of health). This has important implications for the targeting of radon testing and mitigation measures, and demonstrates that attempting to engage with younger people or women of any age in risk communication strategies could show a benefit to both deciding to mitigate and doing so more quickly. This phenomenon has precedent in health-seeking behaviour differences by sex and gender more broadly, in that it has been demonstrated across the behavioural and healthcare experiences of men and women 37 . One acknowledged caveat to our work is that we were not able to obtain gender identity information from all participants and, in those cases, deferred to reported sex. It will be important in future work to obtain bona fide gender identity data from a wider population. Similarly, only 1% of our cohort identified with a gender minority, and so we were unable to analyse this group for most outcomes reported here. Future work will be needed to determine effects for the diversity of gender minorities, as well as a function of other demographic factors such as ethnicity and acculturation.
In addition to behaviour, our work demonstrates that a substantial excess exposure to alpha radiation from radon corresponds with the inability to afford radon mitigation, despite the desire to do so. As this situation was more likely to be experienced by younger people with the most people (and minors) living in the household, this is very undesirable from a cancer prevention and health justice perspective. One possible solution to this would be to subsidize radon mitigation costs for homeowners under a certain income level, or those with younger children in the property. Landlords (whether individuals or corporations) might also be encouraged or mandated to ensure rental properties do not contain high radon levels, conferring protection to a wider population of people who are not homeowners. Although undoubtedly complex to enact or enforce, costs would be offset by an increase in successful radon mitigation outcomes following a test interpreted as high or unsafe. In this cohort, mitigation rates were 38.2%, but could rise to 66.7% if those who expressed the desire to mitigate but could not afford to were enabled to do so. We acknowledge that enabling people to mitigate for radon may not purely be an economic issue, and that there are likely other structural determinants of health at play that have yet to be clearly identified. For example, our study was unable to address inequities for Indigenous North American populations who may experience income instability coupled with remoteness (potentially reduced internet access), and lack of tailored messaging and resources to address radon mitigation. This is a particular gap that must be addressed in nations with populations experiencing marginalization. The best solution to this entire problem would be to enact new building codes so that new properties would be required to be 'radon-safe' , and require mitigation strategies in all existing properties at a policy/regulatory level, which would be government/industry subsidized for people unable to afford or carry this out.
Finally, we found that those in properties high for radon, and who also were able to mitigate those houses, had, overall, the lowest possible rates of radon exposure of all groups moving forward. Strikingly, those in high radon-mitigated households were able to achieve 440% reductions in annual average alpha particle radiation exposures, to levels considered low even compared to those in houses where radon tests were interpreted as 'low or safe' , at only 0.75 mSv/year, a level that is not believed to have any measurable impact on lung cancer risk 14,17,18,22 . Conversely, an alarming percentage of those who interpreted their radon test as 'low or safe' in fact experienced exposure rates far in excess of the 4 mSv/year threshold where increased relative risk of lung cancer has been established, reaching > 40 mSv/year in extreme cases. Our work suggests that this, in part, is driven by regional radon action thresholds (i.e., 200 Bq/m 3 ) that are used by individuals to interpret relative safety, and is further evidence that use of administrative thresholds that exceed levels where a statistically significant increase www.nature.com/scientificreports/ in lung cancer risk is evident (i.e., 100 Bq/m 3 ) may be driving radon exposures the are, in fact, causing lung cancers across populations 13 . We speculate that lowering administrative action thresholds to 100 Bq/m 3 would improve overall health-seeking behaviours and decision-making, encourage greater uptake on radon mitigation, and reduce the future burden of radon-induced lung cancers.  38 , and in accordance with all regional guidelines and regulations.

Statement of approvals.
Participant eligibility and enrollment. Enrollment was based on convenience recruitment for all wanting to join, with all adult homeowners and renters in any residential building type being eligible. No data from any constituent part of this cohort were from known or pre-selected lung cancer cases. Commercial offices or hospitality service buildings were not considered. Records of informed consent were obtained in all cases. Participants were permitted to withdraw at any time.

Radon awareness communication.
Communication methods for study recruitment included dissemination of radon awareness information and provision of opportunities for radon testing through print media, public seminars, online (website and social media) and traditional mass media via organic (unpaid) TV, magazine, and radio exposure in an untargeted manner. Each of these points of communication was further boosted in an organic (unpaid) manner by health agencies, non-profit groups, and some private organizations. Radon awareness content included basic, scientifically informed facts on radon, its health effects in the context of lung cancer and highlighted both historic facts, figures, and emerging regional statistics and have been outlined in detail in Ref. 24 .

Radon testing.
Radon tests were quality controlled by Canadian National Radon Proficiency Program (C-NRPP) certified professionals and distributed centrally by researchers. Care was taken to educate participants in the correct test deployment methods through print, video and direct communication with C-NRPP-certified professionals who adhere to Health Canada's guidelines for long-term (90+ day) radon testing. Participants were advised to place devices on the lowest level of the building occupied for approximately 4 or more hours per day. As participants positioned kits in areas and/or floors of the property where occupants spent the majority of their time, we did not apply any correction factor(s). Participants were asked to deploy radon tests within residences for a minimum of 3 months. Readings based on less than 60 full days were excluded from analysis and, for this cohort, the average period of radon testing was 107.8 days. Radon tests were RadTrak2 closed passive etched track detectors made from CR-39 plastic film inside antistatic holders enclosed in electrically conductive housing with filtered openings to permit gas diffusion, intended for long term (> 90 day) use with a typical linear range of 15 to 25,000 Bq/m 3 . To be read, CR-39 films are etched in 5.5 N NaOH at 70 °C for 15.5 min and scored using TrackEtch software at ISO17025 certified Radonova laboratories (Sweden, EU). Precision and accuracy controls have been reported previously in Ref. 11 .
Surveying. From 2015 to 2020, Canadians purchased alpha track 90+ day radon detectors that they then deployed, returned for analysis, and later received their specific radon reading in a confidential manner. Radon outcomes for this cohort were reported most recently in Ref. 20 . Non-profit study kits were $51.99. Following consent and placement of a radon test, all participants active within the study in October of 2018 were invited to complete demographic surveys and questionnaires on radon awareness and testing experiences using the Qualtrics survey platforms. In November of 2020, all active participants were invited to complete further demographic surveys that included gender identity questions. Of all participants, we obtained both sex and gender identity data from 68.5, with 25% providing only sex data, and 6.5% declining to indicate either. As only 1% of participants chose a gender identity option that was not 'woman' or 'man' , we were unable to analyse distinct gender or sex minority groups in greater detail, but rather identified people in this group as 'everyone else' for purely practical purposes in this study. A complete list of survey questions is outlined in Supplemental Information, Sect. I.

Radiation exposure calculations.
To convert Bq/m 3 indoor air radon levels to human mSv radiation exposures (to lungs), the ICRP provides the following formula 39 : Values for the amount of time spent in the primary residence per year for a typical adult ("time in residence per year") were calculated from individually reported residential occupancy data from 1128 participants, and cross-referenced with data from the National Human Activity Pattern Study (NHAPS) 21 . The NHAPS, which included responses from both Canadian and American respondents, estimated that 68.7% of life was spent inside a residence for the average adult. Of the total 8760 h per year, this equates to 6018 h/year and represents an average of all different employment statuses. Participants reported their data by season (Winter, Spring, Summer, Fall), with weekend/holiday versus workdays accounted for within the questionnaires described in detail in Ref. 20 . All response-derived "time in residence per year" outcomes were linked to individually reported employment statuses and used to extrapolate the same values for all remaining participants (2390) for which 6 www.nature.com/scientificreports/ employment data was collected. Please note that for the purposes of participant responses in this survey, being enrolled in full or part time formal education (at a university, college, technical school, etc.) was considered to be a type of employment and grouped with those responses.

Statistical analysis.
Statistical analyses were carried out using Excel and GraphPad Prism 9.1.1 (225) (www. graph pad. com). One-way ANOVAs were carried out to test differences between groups (e.g., year of construction, occupant age, mSv, etc.), with Bonferroni-Holm corrected post-hoc testing carried out to characterize group differences for pairwise comparisons if the ANOVA reached significance. Mann-Whitney pairwise nonparametric t-tests were used to assess the significance of scatter plot data.

Data availability
The de-identified raw data sets generated by the current study are available to academic researchers following reasonable requests to Dr. Goodarzi and will require a legally binding data transfer agreement. Data may not be used for private, commercial, or for-profit purposes for any reason.