Awareness and Perceptions of the Risks of Exposure to Indoor Radon: A Population-Based Approach to Evaluate a Radon Awareness and Testing Campaign in England and Wales
Abstract
The current study aimed to evaluate the locally directed radon roll-out program that was conducted between 2001 and 2005 in England and Wales to increase radon awareness and testing rates. A representative sample of 1,578 residents aged 16 and older were interviewed who lived in radon-affected areas of 15 local authorities in England and Wales that were eligible for participation in the program. The study systematically sampled across participating and nonparticipating local authorities, “actionable” and “nonactionable” radon-affected areas, and geographic regions with different campaign histories (Wales, Southwest England, and the rest of England). As a multistage sampling strategy was used, the data were analyzed from a multilevel perspective. This study found that participants living in participating local authorities had higher levels of awareness and were more likely to have tested their home for radon than participants living in nonparticipating local authorities. Similar results were found for participants living in “actionable” areas as compared to those living in “nonactionable” radon-affected areas. The study further found that radon awareness and testing rates were the highest in Southwest England and the lowest in Wales. This study suggests that the radon roll-out program has been effective in raising awareness and testing rates, and that ongoing domestic radon campaigns in Southwest England may have raised radon awareness and testing in these areas, showing important reinforcement effects of multiple risk communication campaigns.
1. INTRODUCTION
Radon is a naturally occurring radioactive gas found in rocks and soils. Long-term exposure to indoor radon is known as a significant hazard to human health, and is recognized as the second largest cause of lung cancer in the United Kingdom after cigarette smoking.(1) People are most at risk in well-insulated houses in radon-affected areas. Radon gas may leak through cracks and pores in the foundation into the building where it may build up to high levels. When the decay products of radon are inhaled they may become lodged in the lung, which may contribute to the development of lung cancer.(2) Estimates of lung cancer cases that could be attributed to exposure to indoor radon vary substantially. Dixon(3) estimated that radon exposure accounts for 3–5% of lung cancer fatalities in the United Kingdom, while Darby et al.(1) suggested that residential indoor radon could account for about 9% of fatal lung cancers. Darby et al.(5) suggested that in the United Kingdom between 2,000 and 3,000 fatal lung cancer cases are attributable to exposure to indoor radon. However, a later estimate based on a European-wide epidemiological study in which data from 13 case-control studies were combined suggests that it is more likely to be between 1,000 and 2,000 annual fatalities.(4) More recently, Gray et al.(2) concluded that each year around 1,100 deaths from lung cancer (about 3% of all deaths from lung cancer) are related to radon in the home. Furthermore, direct evidence has emerged showing that most lung cancer cases occur among smokers through synergetic effects between radon and smoking.(1,2,6,7) Also ex-smokers are at higher risk of developing radon-induced lung cancer. A recent report of the independent advisory group on ionizing radiation concluded that the risks for recent ex-smokers would be somewhat lower than for current smokers, while for long-term ex-smokers the risks would be closer to those of lifelong nonsmokers.(8) Health economic analyses show that basic radon prevention measures would be highly cost effective and a worthwhile public health intervention to reduce the number of lung cancers, alongside existing policies to reduce smoking.(2) In this study the effectiveness of a locally directed radon awareness campaign that was conducted between 2000 and 2005 in England will be evaluated through a population-based study.
1.1. Radon Awareness and Testing Campaigns in the United Kingdom
As indoor radon poses a significant health risk to the general public, radon-affected areas are generally well defined, and remedial work is relatively straightforward and cost effective, the U.K. government has since the mid-1980s been running a number of campaigns with varying levels of success. The initial radon awareness and testing campaigns mainly targeted Southwest England, with activities later expanding to radon-affected areas in other parts of England. A similar initiative was undertaken in Wales in the late 1990s. However, while in England households were targeted in areas with at least a 5% probability of being at or above the Action Level of 200 Becquerel per cubic meter (Bq/m3), in Wales this threshold was set at 10%. The typical approach of these early initiatives was to offer free home tests to households in identified risk areas, and to give written advice for those who were at or above the action level on how to take appropriate remedial action.(9) Until 1998 invitations were sent out centrally via the former National Radiological Protection Board (NRPB) on behalf of the government. However, while many took up the offer to test their homes for radon, remediation rates of high testers remained low,(10) even in areas where people's awareness of the health risks of radon was reasonably high.(11-13)
Recognizing that the centralized approach of the early campaigns may have contributed to the relatively small number of households in England taking remedial action, a more locally directed approach was adopted. This new approach was based on the idea that householders are more likely to take up offers of testing and to carry out remedial work if they can approach someone locally to discuss any questions they might have about the issue. The new locally directed approach was piloted in 1999 in partnership with three local authorities, which acted as the public face of the campaign.(14) The results of the pilot program were impressive both in terms of the numbers of people who sought advice from the local authority and the numbers that were apparently committed to carrying out remedial action.(15) The evaluation of the pilot remediation program concluded that the total of remediated properties was raised by up to 100% of previously achieved numbers.(16) As a result it was decided to roll out the locally directed approach to other local authorities in England.
In 2000, the government officially launched the Radon Roll-Out Program with the help of the then NRPB.2 This new locally directed program had the aim to raise public awareness of adverse health effects of indoor radon, increase the number of residents in affected areas testing their homes for radon, as well as to encourage residents with high radon levels to take remedial action. While the program was sponsored by central government departments, and supported by a number of specialist agencies, such as the Health Protection Agency (HPA) and the Building Research Establishment (BRE), local authorities acted as the public face of the initiative. Local authorities were eligible to participate in the radon roll-out program if they contain “actionable” radon areas, i.e., areas where indoor radon concentrations are predicted to be at or above the action level of 200 Bq/m3 in more than 5% of homes. Participating local authorities had to appoint a member of staff to organize the activities of the program at the local level and to be approachable for members of the public. Further key features of the program included that householders were pro-actively contacted, and high testers were actively supported in the remediation process through individual advice on available remediation options at so-called radon road shows and free house visits by local authority representatives.(14, 16) The locally directed approach of the program means that the general procedure of the radon roll-out program was standardized, while specific elements varied according to local preferences and circumstances. Overall, more than 30 English local authorities had participated in the radon roll-out program by 2005 when the HPA took over the management in England. The HPA continues to organize similar activities with local authorities that have not previously been involved in the program. Although the program targeted only local authorities in England, in 2004 a three-year project was conducted in Flintshire to test the locally directed approach in Wales. The latter project was supported and coordinated by the Welsh Assembly Government in collaboration with Flintshire Council with support and advice from the BRE and the HPA.(17)
1.2. Radon-Related Awareness, Perceptions, and Behavior
Various studies have been conducted exploring radon-related awareness, perceptions, and behavior of those who are likely to be exposed to indoor radon. Many of these studies have shown that despite high levels of awareness, people do not feel at risk, and/or find it necessary to test their home for radon.(18,19) For example, Ferng and Lawson(20) found that very few residents of a radon-affected area recognized radon as a major problem, even if they reported high levels of knowledge of radon. Duckworth et al.(19) found that most respondents of an area that was believed to have high levels of radon were familiar with the term radon, but did not view it as an immediate health hazard. Similarly, Weinstein et al.(21) found that, despite high levels of radon awareness, most homeowners in a radon-affected area in New Jersey believed that they were less at risk than their neighbors. These studies suggest that attempts to encourage home radon monitoring and mitigation may be hampered by optimistic biases in public perceptions of the risks of indoor radon. This may be due to people being unaware of the specific health effects of exposure to indoor radon. Wang et al. found that while 82% of New York State residents had heard of radon, only 21% knew that it was a naturally occurring radioactive gas and can cause lung cancer.(22) However, residents living in counties with high radon levels were more likely to have heard or to be “knowledgeably aware” of radon than those living in a control area with lower levels of radon.(22) They saw these results as indirect evidence that the New York State public awareness programs may have increased public awareness and promoted residential testing. In a more recent questionnaire survey, in which 1,528 individuals were targeted according to their spatial relationship with the risks of radon and overhead powerlines, Poortinga et al. found that beliefs about the two risk cases mainly differed according to the level of “exposure.”(23) Whereas people living in high risk areas were more concerned about the risks of indoor radon than those living in low risk areas, they also thought that the risks of radon more acceptable and expressed higher levels of trust in authorities to tell them if their health is at risk. Similar to Wang et al., Poortinga et al. interpreted these findings as reflecting the positive effects of successive radon campaigns in high risk radon areas, which may have raised awareness and concern, and at the same time may have helped to increase trust by showing that the government takes the health risks of indoor radon gas seriously.(22,23) Overall, there is strong cross-sectional evidence that prompts of a free test kit together with targeted media campaigns may encourage people to test their home for radon.(24)
1.3. The Effectiveness of Radon Awareness and Testing Campaigns
The evidence from cross-sectional population-based surveys regarding the effects of radon awareness and testing campaigns contrasts with other research that has highlighted the difficulties of public health strategies to communicate the risks of radon to the general public, and to translate these higher levels of awareness and knowledge into people testing and remediating their homes for radon.(12,22,25) Field experimental studies have only found limited evidence that the provision of risk information may influence how people think and behave with regard to radon. Weinstein et al. found that residents living in a radon-affected area in New Jersey who received an intervention designed to emphasize the likelihood of finding high radon levels in their area were more likely to acknowledge the possibility of high radon levels in their homes as compared to those who received a minimal-treatment intervention that only suggested that testing was a “good idea.”(26) Weinstein et al. found that interventions that differed in their presentation of the magnitude of the threat (i.e., likelihood and severity of the risks) and the difficulty of reducing radon levels influenced people’s risk perceptions.(27) However, the different interventions did not affect the likelihood of ordering a radon test. Weinstein et al. suggested that the ineffectiveness of these interventions may be due to a mismatch between the provided information and the stage of the person with regard to the health protective action.(28) As suggested by their precaution adoption process (PAP) model, people have to go through a number of consecutive stages to reach the ultimate desired stage of remediating a home to lower radon levels; and that the barriers impeding progress toward protective action change from stage to stage.(21, 28) In line with the PAP model, Weinstein et al. found that interventions stressing the likelihood of exposure encouraged homeowners to decide to test their homes for radon (i.e., moving from the undecided to the decided-to-act stage), while a low-effort “how-to-test intervention” was found to be relatively more effective in getting homeowners to order a test.(29)
Overall, the results indicate that where simply providing risk information is sufficient to promote testing and remediation, campaigns that contain stage-relevant information about how to test may be more effective.(30) However, it is important to consider that the experiments described here do not evaluate the actual impacts of radon initiatives. Furthermore, the population-based research discussed previously at most provides indirect evidence of the potential positive effects of radon awareness and testing campaigns for people in high radon areas where such initiatives have taken place.(22,23) Hitherto, no studies have been conducted that systematically evaluated the effects of campaigns on radon-related awareness and behavior.
2. THIS STUDY
The current study was set up to evaluate the radon roll-out program through a population-based study of people's awareness and perceptions of the risks of radon in their homes, and associated health-protection behaviors. It systematically sampled across local authorities and regions (i.e., Wales, Southwest England, and the rest of England) with different histories of radon testing initiatives, and across areas with different geographies of radon occurrence. The aims of the study were to examine whether (1) people's radon-related awareness, perceptions, and behavior vary according to the likelihood of exposure to radon, and (2) a locally directed radon awareness and testing campaign (i.e., the radon roll-out program) has had an effect on people's radon-related awareness, perceptions, and behavior. It was expected that, in accordance with previous studies,(22,23) people living in high risk radon areas are more aware, perceive higher risks, and are more likely to have tested for radon. It was also expected that people living in local authorities that have participated in the locally directed radon awareness and testing campaign are more aware of radon and are more likely to have tested for radon. Furthermore, as Southwest England has been targeted in successive campaigns since the 1980s and very few activities have been conducted in Wales,(9) it is believed that awareness and testing rates will be the highest and lowest in these regions, respectively.
3. METHOD
3.1. Sample and Procedure
Data for this study were collected between October 7 and December 1, 2008. The market research company Ipsos MORI interviewed a representative quota sample of 1,578 residents aged 16 and older living in radon-affected areas of 15 local authorities in England and Wales that were principally eligible for participation in the program (see Table I). The face-to-face interviews were carried out using computer assisted personal interviewing (CAPI). CAPI was used to allow a quick turnaround of data and complex routing of questions. On average, the interviews lasted 28 minutes.
| Local Authority | Interviews | Roll-Out Involvement1 | Location | Population Density2 | Maximum Exposure3 |
|---|---|---|---|---|---|
| North Devon | 107 | No | South West England | 81 | 10–30 |
| Caradon | 103 | No | South West England | 120 | >30 |
| Kerrier | 107 | Yes | South West England | 198 | >30 |
| Restormel | 106 | No | South West England | 216 | >30 |
| South Hams | 109 | Yes | South West England | 92 | >30 |
| Torbay | 102 | Yes | South West England | 2,081 | >30 |
| West Somerset | 102 | No | South West England | 50 | 10–30 |
| Lincoln | 101 | No | Rest of England | 2,408 | 5–10 |
| Northampton | 103 | No | Rest of England | 2,404 | 10–30 |
| Kettering | 106 | Yes | Rest of England | 355 | 10–30 |
| Wellingborough | 103 | Yes | Rest of England | 453 | 10–30 |
| Derbyshire Dales | 106 | Yes | Rest of England | 88 | >30 |
| Flintshire | 100 | Yes | Wales | 341 | 10–30 |
| Vale of Glamorgan | 110 | No | Wales | 363 | 10–30 |
| Wrexham | 113 | No | Wales | 257 | 10–30 |
- Note: (1) Involved in the radon roll-out program; (2) persons per km2 (Office for National Statistics, 2002); (3) proportion of properties that are expected to be at or above the action level of 200 Bq/m3.
The study employed a multistage sampling strategy in order to be able to use multilevel modeling for the analyses. First, 15 local authorities were selected on the basis of their radon affectedness, involvement in the radon roll-out program, geographic location, and population density (see Table I). The primary sampling frames of the study were local authorities that contain “actionable” areas (i.e., areas where at least 5% of the houses are predicted to be at or above the action level of 200 Bq/m3), as these local authorities were eligible for participation in the radon roll-out program. The local authorities were identified using the Radon Atlas of England and Wales.(31) The 2002 Radon Atlas of England and Wales indicates the likelihood of houses being at or above the action level on a 1 km square grid in Southwest England and a 5 km square grid in the rest of England and Wales.3 To be able to study the potential effects of the radon roll-out program, the survey included local authorities that had participated in the radon roll-out program as well as local authorities that had not. Furthermore, to ensure a good geographical spread, local authorities were selected on the basis of their location in the Southwest of England, the rest of England, and Wales. Both urban and rural local authorities were included, with population densities ranging from 50 to 2,408 people per square kilometer. Table I displays characteristics of the 15 local authorities that were selected for the study. In each local authority a minimum of 100 interviews were conducted. Second, 214 sampling points were selected randomly from “twinned” output areas that were identified as being radon affected (i.e., area with a probability of 1% or more of a home having radon above the action level of 200 Bq/m3) using the Radon Atlas of England and Wales.(31) The current study only used information from the 2002 Radon Atlas of England and Wales. Output Areas (OAs) are the smallest building blocks of the Census, with an average of 125 resident households; each sampling point consisted of two ‘”twinned” OAs that were either bordering or very close to each other. Third, quotas were set for each twinned output area for gender, age, and working status in order to reflect the demographic profile of the sampling point. At least seven interviews were conducted in each selected sampling point.
3.2. Measures
3.2.1. Dependent Variables
In this study a number of questions were asked to assess radon-related awareness, risk perceptions including concern, and behavior, respectively.
- Awareness
:
General awareness of radon (Radon Awareness) was measured with the item “Had you heard of radon before this interview?” to which people could respond with “yes,”“no,” and “don't know.” As there were only a limited number of “don't know” responses (n= 10), these were removed from the analysis. Awareness of living in a radon-affected area (Area Awareness) was measured with the item “As far as you know, do you think you live in an area affected by radon?” This question could also be answered with “yes,”“no,” and “don't know.” However, as far more people responded with “don't know” (see Table II), this response was included as a separate category in the analyses.
- Radon Concern
:
The degree to which people are concerned about radon was measured with the question “To what extent, if at all, are you concerned about the health risks to you personally associated with radon?” The answer options “very concerned” and “fairly concerned” were taken to reflect “concern,” while “not very concerned” and “not at all concerned” were taken to represent “no concern.” The item also included a “don't know” option.
- Behavior
:
Radon Testing was measured by asking participants to select a statement that “best describes your thoughts before this interview about testing your home for radon.”Table II summarizes both the options and responses given to this question. The answer options reflected different stages of the precaution adoption process (PAP) model, which has been specifically developed to investigate the context of indoor radon testing and remediation.(27) The options “I have already completed a radon test,”“I have a test for radon in progress,” and “I have bought a test for radon” were combined to reflect participants who had taken the decision to test their home for radon, and compared to all other answer options. Due to the small number of participants who responded with “don't know” (n= 61), this category was omitted from the analysis.
| Variable | Response Categories | % |
|---|---|---|
| Sociodemographics | ||
| Gender | Male | 49 |
| Female | 51 | |
| Age | 16–24 | 13 |
| 25–34 | 14 | |
| 35–44 | 18 | |
| 45–54 | 16 | |
| 55–64 | 15 | |
| 65 and older | 23 | |
| Social grade | AB | 24 |
| C1 | 28 | |
| C2 | 24 | |
| DE | 25 | |
| Homeowner | Yes | 73 |
| No (renting) | 26 | |
| Other | 1 | |
| Smoker | Yes | 25 |
| No | 74 | |
| Refused | * | |
| Awareness | ||
| Had you heard of radon before this interview? | Yes | 72 |
| No | 28 | |
| Don't know | 1 | |
| As far as you know, do you think you live in an area affected by radon? | Yes | 33 |
| No | 40 | |
| Don't know | 28 | |
| Radon concern | ||
| To what extent, if at all, are you concerned about the health risks to you personally associated with [radon]? | ||
| Very concerned | 5 | |
| Fairly concerned | 19 | |
| Not very concerned | 30 | |
| Not at all concerned | 23 | |
| Don't know | 24 | |
| Radon Testing | ||
| Which of the following statements best describes your thoughts before this interview about testing your home for radon? | I have never thought about testing my home for radon | 60 |
| I am undecided whether or not to test my home for radon | 10 | |
| I have decided I don't want my home tested for radon | 4 | |
| I have decided I do want my home tested for radon | 3 | |
| I have already completed a test for radon | 19 | |
| I have a test for radon in progress | * | |
| I have bought a test for radon | * | |
| Don't know | 4 | |
- Note: (*) Represents a value of less than half a percent, but not zero.
People were given the opportunity to respond with “don't know” to all questions in order to prevent the generation of so-called pseudo or nonattitudes.(32) Not providing a “don't know” option would force respondents to express an opinion where they may not necessarily have one.
3.2.2. Independent Variables
Independent variables were included at three distinct levels of analyses, i.e., the individual, sampling point, and local authority level, respectively.
At the individual level, the sociodemographic variables of gender, age, social grade, and homeownership (“tenure”) were included. The social grade and home ownership variables were included as indicators of socioeconomic status of the participants. The social grade variable was based on the occupation of the head of the household and included the categories AB (upper middle class and middle class), C1 (lower middle class), C2 (skilled working class), and DE (working class and those at the lowest level of subsistence). As in previous research the presence of children in the household did not appear to affect radon-related awareness, perceptions, and behavior,(23) this variable was not included in the current analyses. Smoking status was included because more direct evidence has become available that smokers are the most at risk from radon.(1, 2, 6) Since the commencement of the study it has become more widely recognized that the risks to ex-smokers are also significant.(8) However, this has not been considered in the current study.
At the sampling-point level we included a variable indicating the likelihood of exposure to radon. The data set was constructed in such a way that only the population living in radon-affected areas within the 15 selected local authorities was targeted. A dummy variable indicated whether people were living in an “actionable” radon-affected area (>5% likelihood of exposure at or above the action level) as compared to living in a nonactionable radon-affected area (1–5% likelihood of exposure at or above the action level).
At the local authority level we included the geographic location (“Region”) of the local authority (i.e., Southwest England, rest of England, or Wales) and whether or not the local authority had participated in the radon roll-out program (“Roll-out involvement”).
3.2.3. Analyses
The data were analyzed using multilevel modeling techniques.(33,34) A multistage sampling strategy created a hierarchical data set of individuals clustered within sampling points (OAs) in 15 selected local authorities (see Table I). This approach was taken with the specific aim to take into account contextual geographical factors in radon-related awareness, concern, and behavior.(23, 35) Simple three-level binomial and multinomial models were constructed with individuals at level 1, the sampling points at level 2, and local authorities at level 3. The binomial models were used for the radon awareness and radon testing variables, while multinomial models were used for the area awareness and radon concern variables (see the Section 3.2.1). Binomial and multinomial analyses were used as many respondents used the “don't know” option to answer the questions. It was therefore not possible to use ordinary (multilevel) regression analyses without a considerable loss of respondents and, as a result, statistical power. It was not possible to create scales using multiple items for the same reason. Respondents were provided with a “don't know” response option in order to prevent the generation of so-called pseudo or nonattitudes.(32)
The multilevel models were constructed within the MLwiN software package, with the model parameters estimated using Monte Carlo Markov chains (MCMC) methods.(36) This article will only report the fixed part of the multilevel models. The parameters of individual-level variables can be interpreted as normal regression coefficients of a single level regression analysis, while the parameters of the variables at the sampling point and local authority level represent the effects that are common to individuals living in the same sample point or local authority, respectively. The sociodemographic variables are included at the individual level in order to control for potential compositional differences. Radon awareness was included in the models as an independent variable to control for important awareness differences between the different sampling points. It appeared that this variable was strongly related to the other radon perceptions and behavior measures. Perhaps not surprisingly, people who had not heard of radon before the interview were far more likely to respond with “don't know” to the other questions.
4. RESULTS
Table II presents the characteristics of the sample, which was representative of the population living in radon-affected areas within the 15 selected local authorities. The table shows that a clear majority (72%) had heard of radon before the interview, 28% had not, while 1% responded with “don't know.” Although only people living in radon-affected areas were approached, just 33% thought that they live in an area affected by radon, 40% thought that they did not, and 28% did not know whether or not they live in an area affected by radon. About one in four (24%) of the respondents indicated that they are fairly or very concerned about radon, while more than half of the respondents (53%) indicated that they are not very or not at all concerned. About one in four (24%) used the “don't know” option to answer the question. Table II also shows that a clear majority (60%) has never thought about testing their home for radon, while 19% had already completed a test for radon. One in 10 (10%) indicated to be undecided about whether or not to test their home for radon, 4% had decided that they did not want to test their home for radon, and 3% had decided that they did want to test their home for radon. Overall, 20% of the respondents had already completed a test for radon, have a test for radon in progress, or have bought a test for radon.
Table III shows that there is considerable socio-demographic variation in radon awareness. Men were nearly twice as likely to have heard of radon before the interview, as were homeowners. In addition, clear age effects were found, with those aged between 55 and 64 being more than 20 times as likely to have heard of radon, and those aged 65 and over nearly 15 times as likely as the reference group of 16–24 year olds. A clear social grade gradient was found, with lower social grades being less likely to have heard of radon before the interview. While evidence has become available that smokers are mainly at risk from exposure to indoor radon, this study shows that smokers are less likely to have heard of radon than nonsmokers. With regard to the contextual variables, people living in actionable areas were 70% more likely to have heard of radon before the interview than those living in nonactionable areas, while at the same time those living in the seven local authorities that had participated in the radon roll-out program were nearly twice as likely to have heard of radon than those living in the eight local authorities that had not participated in the radon roll-out program. Overall, people in Southwest England were more than twice as likely to have heard of radon as those living in “the rest of England,” while people in Wales were 62% less likely to have heard of radon.
| Radon Awareness (No) | Area Awareness (No) | Radon Concern (Not Very/At All) | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Yes | Yes | Don't Know | Fairly/Very | Don’t Know | |||||||||||
| OR | 95%CI | p | OR | 95%CI | p | OR | 95%CI | p | OR | 95%CI | p | OR | 95%CI | p | |
| Actionable (no) | |||||||||||||||
| Yes | 1.70 | 1.19–2.43 | ** | 2.07 | 1.46–2.95 | *** | 1.48 | 1.03–2.12 | * | 1.27 | 0.96–1.68 | n.s. | 1.10 | 0.70–1.75 | n.s. |
| Roll-out involvement (no) | |||||||||||||||
| Yes | 1.95 | 1.36–2.79 | *** | 2.35 | 1.66–3.33 | *** | 0.82 | 0.58–1.16 | n.s. | 1.16 | 0.87–1.55 | n.s. | 0.37 | 0.22–0.60 | *** |
| Region (rest of England) | |||||||||||||||
| Southwest England | 2.05 | 1.37–3.08 | *** | 0.64 | 0.43–0.94 | * | 0.82 | 0.56–1.21 | n.s. | 1.06 | 0.77–1.45 | n.s. | 1.30 | 0.76–2.23 | n.s. |
| Wales | 0.38 | 0.24–0.58 | *** | 0.53 | 0.31–0.90 | * | 1.07 | 0.67–1.70 | n.s. | 1.49 | 0.98–2.27 | n.s. | 2.89 | 1.56–5.34 | *** |
| Gender (female) | |||||||||||||||
| Male | 1.99 | 1.47–2.70 | *** | 1.08 | 0.80–1.45 | n.s. | 0.69 | 0.51–0.93 | * | 0.65 | 0.50–0.85 | ** | 0.59 | 0.40–0.88 | ** |
| Age (16–24) | |||||||||||||||
| 25–34 | 3.21 | 1.84–5.59 | * | 2.56 | 0.97–6.76 | n.s. | 0.77 | 0.44–1.34 | n.s. | 0.94 | 0.42–2.09 | n.s. | 0.90 | 0.44–1.83 | n.s. |
| 35–44 | 1.28 | 0.72–2.29 | n.s. | 3.64 | 1.47–9.01 | ** | 0.76 | 0.44–1.31 | n.s. | 1.47 | 0.70–3.05 | n.s. | 0.72 | 0.36–1.46 | n.s. |
| 45–54 | 1.35 | 0.72–2.52 | n.s. | 4.45 | 1.83–10.87 | ** | 0.75 | 0.43–1.32 | n.s. | 1.59 | 0.77–3.27 | n.s. | 0.62 | 0.29–1.31 | n.s. |
| 55–64 | 20.39 | 10.78–38.55 | *** | 4.78 | 1.94–11.81 | *** | 0.57 | 0.31–1.04 | n.s. | 1.71 | 0.82–3.54 | n.s. | 0.58 | 0.26–1.29 | n.s. |
| 65+ | 14.63 | 8.17–26.18 | *** | 2.72 | 1.11–6.69 | * | 0.51 | 0.30–0.89 | * | 1.17 | 0.57–2.41 | n.s. | 1.10 | 0.56–2.17 | n.s. |
| Social grade (AB) | |||||||||||||||
| C1 | 0.49 | 0.31–0.77 | ** | 0.88 | 0.59–1.30 | n.s. | 0.85 | 0.55–1.32 | n.s. | 0.96 | 0.66–1.39 | n.s. | 1.60 | 0.89–2.87 | n.s. |
| C2 | 0.31 | 0.19–0.49 | *** | 0.59 | 0.38-0.90 | * | 1.01 | 0.65–1.57 | n.s. | 1.21 | 0.83–1.76 | n.s. | 1.47 | 0.78–2.75 | n.s. |
| DE | 0.23 | 0.14–0.38 | *** | 0.62 | 0.39–0.99 | * | 0.97 | 0.61–1.54 | n.s. | 1.24 | 0.81–1.89 | n.s. | 1.58 | 0.82–3.04 | n.s. |
| Tenure (nonhomeowner) | |||||||||||||||
| Homeowner | 2.02 | 1.45–2.82 | *** | 1.21 | 0.81–1.82 | n.s. | 0.52 | 0.33–0.81 | ** | 1.09 | 0.76–1.57 | n.s. | 0.92 | 0.56–1.52 | n.s. |
| Smoking status (nonsmoker) | |||||||||||||||
| Smoker | 0.65 | 0.46–0.90 | * | 0.71 | 0.48–1.05 | n.s. | 0.78 | 0.54–1.12 | n.s. | 0.87 | 0.62–1.22 | n.s. | 1.10 | 0.70–1.73 | n.s. |
| Radon awareness (no) | |||||||||||||||
| Yes | — | — | — | 6.54 | 3.80–11.26 | *** | 0.58 | 0.40–0.82 | *** | 1.88 | 1.09–3.26 | * | 0.03 | 0.02–0.05 | *** |
- Note: Reference categories are given in parentheses; *p < 0.05, **p < 0.01, ***p < 0.001, n.s. = nonsignificant; sample sizes (n) for the groups were: radon awareness: Yes 1,132, No 436; area awareness: Yes 513, No 627, don't know 438; radon concern: fairly/very 375, not very/at all 826, don't know 377.
Overall, older participants were more likely to think that they live in an area affected by radon (see Table III), and participants with a lower socio-economic background were less likely to think that they live in a radon-affected area. Furthermore, men, those aged 65 and over, and homeowners were less likely to use the “don't know” option for answering this question. A strong awareness effect were found with those having heard of radon before the interview being more than six times more likely to think that they live in a radon-affected area, while being 42% less likely to respond with “don't know.” People living in actionable areas were more than twice as likely to think that they live in an area affected by radon than those living in nonactionable areas. Similarly, people living in local authorities that had participated in the radon roll-out program were more than two times as likely to think that they live in a radon-affected area as those living in nonparticipating local authorities. People living in “the rest of England” and Wales were 36% and 47% less likely to think that they live in a radon-affected area, respectively, as compared to people living in Southwest England.
Very few individual level effects were found for concern about radon (see Table III). Overall, men were less likely to report to be Fairly or Very concerned about radon, and to respond with “don't know” (35% and 41%, respectively), while people who had heard of radon before the interview were 42% more likely to report to be concerned about radon and 97% less likely to respond with “don't know.” With regard to the contextual variables, people living in local authorities that had participated in the radon roll-out program were 63% less likely to respond with “don't know” to the radon concern question. No other contextual effects were found.
Table IV shows the results for the radon testing variable. Overall, older respondents were more likely to have already completed a test for radon, have a test for radon in progress, or have bought a test for radon, with the age group of 55–64 being the most likely to have tested for radon. Respondents who were classified as a social grade DE were 50% less likely to have tested for radon. Although it is possible that this result is confounded by other factors, e.g., lower social grades are more likely to live in rented accommodation, housing tenure is controlled for in the analyses. People who had heard of radon were 8 times more likely to have tested for radon as compared to those who had not heard of radon before the interview. With regard to the contextual variables, people who lived in an actionable radon-affected area were 197% more likely to have tested for radon than those living in nonactionable radon-affected areas, while, at the same time, people living in local authorities that had participated in the radon roll-out program were 167% more likely to have done so. People living in Wales were 76% less likely to have already completed a test for radon, have a test for radon in progress, or have bought a test for radon.
| Radon Testing (No) Yes | |||
|---|---|---|---|
| OR | 95%CI | p | |
| Actionable (no) | |||
| Yes | 2.97 | 1.97–4.46 | *** |
| Radon roll-out involvement (no) | |||
| Yes | 2.67 | 1.83–3.89 | *** |
| Region (rest of England) | |||
| Southwest England | 1.14 | 0.76–1.72 | n.s. |
| Wales | 0.24 | 0.11–0.49 | *** |
| Gender (female) | |||
| Male | 1.10 | 0.80–1.52 | n.s. |
| Age (16–24) | |||
| 25–34 | 1.45 | 0.46–4.55 | n.s. |
| 35–44 | 2.17 | 0.75–6.30 | n.s. |
| 45–54 | 3.60 | 1.27–10.24 | * |
| 55–64 | 6.13 | 2.13–17.68 | *** |
| 65+ | 4.64 | 1.63–13.18 | ** |
| Social grade (AB) | |||
| C1 | 0.70 | 0.46–1.07 | n.s. |
| C2 | 0.64 | 0.41–1.00 | n.s. |
| DE | 0.50 | 0.30–0.83 | ** |
| Tenure (nonhomeowner) | |||
| Homeowner | 2.26 | 1.36–3.73 | *** |
| Smoking status (nonsmoker) | |||
| Smoker | 0.93 | 0.61–1.41 | n.s. |
| Radon awareness (no) | |||
| Yes | 8.00 | 3.48–18.39 | *** |
- Note: Reference categories are given in parentheses; *p < 0.05, **p < 0.01, ***p < 0.001, n.s. = nonsignificant; sample sizes (n) for the groups were: radon testing: Yes 319, No 1,198.
5. DISCUSSION
In this study a systematic population-based survey was conducted in order to examine whether radon-related awareness, perceptions, and behavior vary according to the geography of radon occurrence in the area as well as the history of radon awareness and testing initiatives. Its main aim was to study whether a locally directed radon campaign (i.e., the radon roll-out program) has helped to increase radon awareness and testing. A tightly controlled sampling design ensured that people living in “actionable” and “nonactionable” radon-affected areas were included.
This study found that awareness of radon is generally high in radon-affected areas. A clear majority indicated that they had heard of radon before the interview. However, only about one in three thought that they lived in an area affected by radon, even if only residents of radon-affected areas were targeted in this survey. In line with previous studies,(18,20,22,19,23)) these relatively high levels of radon awareness did not translate into higher levels of concern.
The radon roll-out program appears to have been effective in raising radon awareness and testing. As expected, residents of participating local authorities had higher levels of radon awareness and were more than twice as likely to have tested their homes for radon as residents of nonparticipating local authorities. Only limited differences were found for radon-related concern. However, residents of participating local authorities were less likely to respond with “don't know” to the concern question. This suggests that, while the radon roll-out program has had no effect on people's radon-related risk perception per se (i.e., participation does not appear to have increased or decreased concern), the roll-out program has provided people essential information about whether or not they should be concerned by assessing actual radon levels.
In regard to the research question whether people's awareness and perceptions of radon vary according to exposure, the study found that participants from radon-actionable areas were more likely to be aware of indoor radon and also to have tested for radon. These findings are not necessarily surprising, as residents of high radon “actionable” areas are likely to be better informed as a result of ongoing governmental efforts to increase domestic indoor testing rates since the 1980s (for an overview, see Ref. 9). The strong effects found here suggest that these efforts have been successful in raising awareness and testing rates in these areas. However, in contrast to the expectations, no systematic variation was found for concern about radon. These results are only partly in line with earlier findings of Poortinga et al.,(23) who reported that people living in areas with a high radon potential showed higher levels of both radon awareness and concern. This could mean that the effects presented by Poortinga et al.(23) are attributable to potential selection biases that may have occurred in that study. That is, they suggested the possibility that while on average people living in radon-affected areas may not necessarily be more concerned, the issues are more salient to them and as a result they were more inclined to respond.(37)
Geographic location was also found to be an important factor for people's radon-related awareness, perceptions, and behavior. This study found large regional differences, with residents of Wales showing the lowest level of awareness, responding most often with “don't know” to the concern question, and being the least likely to have tested their homes for radon. Overall, these results reveal a clear gap in radon-related awareness and behavior between England and Wales. This is most likely due to the history of awareness raising campaigns conducted in different parts of the United Kingdom.(9) While in Southwest England public awareness campaigns have been conducted since the 1980s, in Wales only a limited number of radon awareness programs have been conducted in the late 1990s. Also, while programs in England targeted radon-affected areas with a 5% likelihood of homes being affected by radon, in Wales this threshold was set at 10%. Furthermore, the radon roll-out program has been conducted in one Welsh local authority area only.(17) The results show that particularly in Wales more effort needs to be made to inform residents about radon in order to enable them to make better informed decisions about testing and remediating their homes for radon.
This study found consistent gender effects across the different measures, showing that men were generally more aware of but less concerned about indoor radon. These findings reflect the general pattern found in other studies that show that men tend to perceive lower health risks of radon than women and express higher levels of concern,(18-20,23) and appear to reflect general gender differences in risk perception.(38-41) The study found clear age effects for radon awareness and testing, with older participants generally being more aware and to have tested for radon. Previous findings that older participants perceived radon to be less hazardous than younger people do(18-20) were, however, not confirmed in this study. There were also clear socioeconomic differences in radon-related awareness, perceptions, and behavior. Participants with a lower socioeconomic background were generally less likely to be aware of radon and to have their home tested for radon. Similarly, nonhomeowners were generally less aware of radon and less likely to have tested for radon than homeowners. These socioeconomic differences in awareness may partly reflect that social renters often do not organize the testing themselves and/or receive the results, or that the testing may have predated the renter's occupancy.4 Furthermore, renters generally have fewer resources tied up in the home than owner-occupiers. An owner-occupier may be therefore more concerned about anything that might appear to affect the value of the home and as a result may be more inclined to test for radon. The findings with regard to the concern and risk perceptions variables were less consistent.
While the research suggests that the roll-out program has raised radon awareness and testing in radon-affected areas, our findings also indicate that there is more scope for further campaigns, as few people—in particular in nonactionable radon-affected areas—are aware of the potential health risks of radon. The study shows that in particular Wales should be targeted in future campaigns, given the low levels of awareness in that region. Furthermore, the study suggests that the program has been less successful in reaching residents from lower socio-economic backgrounds who do not own their homes. Special efforts therefore need to be made to include these groups in future initiatives. Here it may be more effective to target landlords to conduct the tests as part of their responsibility to provide safe housing rather than the renters themselves.
An important methodological conclusion is that multilevel modeling techniques can be used in risk management and communication research to systematically examine the importance of different local policy and communication initiatives for risk-related attitudes and behaviors.(23,35) By taking the geographical context of radon occurrence in the area as well as the history of radon awareness and testing campaigns, this study has been able to provide a more systematic and contextualized view of radon-related awareness, perceptions, and behavior. It should be mentioned that single item measurements were taken, which may have been vulnerable to measurement error. However, additional analyses (not reported here) for perceived vulnerability and perceived severity of the risks of indoor radon produced very similar results to the ones for radon concern, validating the findings of the study. Furthermore, the cross-sectional nature of the study means that no causal claims can be made from the findings. A longitudinal study following a domestic radon initiative would provide more clarity about the dynamics of radon-related attitudes and behaviors. Such research examining the impacts of a radon intervention would also shed more light on which groups of residents are more likely to respond to an invitation to test their home for radon, and which psychological factors might be involved.
Footnotes
ACKNOWLEDGMENTS
This study was funded by the Department of Health Radiation Protection Research Program (RRX 121). We would like to thank Ipsos MORI for collecting the data for the population-based survey and their advice at various stages of the study. We would also like to thank the Health Protection Agency for its help throughout the project. The contents of this report remain the sole responsibility of its authors and do not represent the policy of the Department of Health or any other government department.
