Attitudes towards lung cancer screening in socioeconomically deprived and heavy smoking communities: informing screening communication

Abstract Background While discussion continues over the future implementation of lung cancer screening, low participation from higher risk groups could limit the effectiveness of any national screening programme. Objectives To compare smokers’ beliefs about lung cancer screening with those of former and never smokers within a low socioeconomic status (SES) sample, to explore the views of lower SES smokers and ex‐smokers in‐depth, and to provide insights into effective engagement strategies. Design, setting and participants Using proactive, community‐based recruitment methods, we surveyed 175 individuals from socioeconomically deprived communities with high smoking prevalence and subsequently interviewed 21 smokers and ex‐smokers. Participants were approached in community settings or responded to a mail‐out from their housing association. Results Interviewees were supportive of screening in principle, but many were doubtful about its ability to deliver long‐term survival benefit for their generation of “heavy smokers.” Lung cancer was perceived as an uncontrollable disease, and the survey data showed that fatalism, worry and perceived risk of lung cancer were particularly high among smokers compared with non‐smokers. Perceived blame and stigma around lung cancer as a self‐inflicted smokers’ disease were implicated by interviewees as important social deterrents of screening participation. The belief that lungs are not a treatable organ appeared to be a common lay explanation for poor survival and undermined the potential value of screening. Conclusions Attitudes towards screening among this high‐risk group are complex. Invitation strategies need to be carefully devised to achieve equitable participation in screening.


| INTRODUCTION
In 2011, there were over 35 000 deaths from lung cancer in the UK, representing 22% of total cancer mortality. 1 Early stage at diagnosis is the strongest predictor of survival 2 but only 15% of patients are diagnosed at Stage 1, 3 and the diagnosis is often incidental. 4 One promising strategy for improving outcomes is to screen for early stage disease in high-risk smokers and ex-smokers using low-dose computed tomography (LDCT). The National Lung Screening Trial (NLST) found a 20% relative risk reduction in lung cancer mortality following three annual LDCT screens compared with chest X-ray, 5

and the United States Preventive
Service Task Force (USPSTF) recommend screening in conjunction with smoking cessation counselling. 6 However, a favourable harm-benefit screening ratio depends on engaging those at high risk with screening. 7,8 Individual risk prediction is becoming increasingly sophisticated and the science of identifying eligible screening candidates ever more precise. 9 Smoking is the key risk factor for lung cancer, accounting for 86% of diagnoses made within the United Kingdom; 10 with other risk factors including occupational exposures, 11 respiratory disease 12 and a family history of lung cancer. 13 Rates of smoking in the UK are highest in socioeconomically deprived areas, where lung cancer incidence and survival are worse. 14 Lower socioeconomic status (SES) may amplify risk due to younger age of starting smoking, and greater tobacco smoke inhalation, nicotine dependence, second-hand smoke exposure and difficulty quitting. [15][16][17][18] Other risk factors, such as a history of pneumonia, also increase with measures of deprivation. 19 In the UK Lung Screening Trial (UKLS), the number of participants registering a high score on the Liverpool Lung Project (LLP) risk prediction model ranged from 8% to 18% from the least to most socioeconomically deprived quintiles. 20 Smokers from deprived communities are therefore an important group to engage with screening, but participation in trials has been skewed towards former smokers and the better educated. 20,21 Compared with an eligible population cohort (a subsample of respondents to the US Census Department's Tobacco Use Supplement), NLST participants were younger, had more years of education and were more likely to be former (than current) smokers. 22 Surveys carried out away from clinical settings in the United States have begun to identify attitudes that might discourage smokers' participation. One survey found smokers were less likely to perceive survival benefit from lung cancer screening or believe they would undergo surgery for a screen-detected cancer. 23 In an ethnically diverse sample, fatalism, concern about radiation and anxiety about having a CT scan predicted lower screening intentions. 24 Three studies [25][26][27] specifically investigated the attitudes of those declining screening offered in the trial context, although the response rate in this group (of all those invited) is usually <10%. Non-participants who responded to a follow-up survey (n=97) in the Dutch-Belgian trial NELSON were found to perceive screening as too much effort or unnecessary due to a lack of respiratory symptoms. 25 In the UK Lung Screening Trial (UKLS), a survey of non-participants (n=748) found a small proportion of smokers to report emotional barriers to participation, but practical barriers were most commonly cited. 26 A qualitative study of patients (n=24) declining participation in a UK screening trial (Lung-SEARCH) for COPD patients identified four types of attitudes explaining reluctance to participate: worry, fatalism, avoidance and believing oneself too old to benefit. 27 There are no data from the United Kingdom on beliefs about lung cancer screening among individuals from socioeconomically deprived communities with high smoking prevalence, a group less likely to engage in research carried out through traditional channels and for whom proactive recruitment methods may be better suited.
The effectiveness of any lung cancer screening programme depends, in part, upon uptake and any inequalities in participation ultimately have the potential to exacerbate inequalities in lung cancer survival.
Understanding why screening participation is low among this highrisk group would contribute to the development of evidence-based engagement strategies and invitation materials to ensure the reach of any future screening programme is equitable. Therefore, this study investigated attitudes towards lung cancer screening within lower SES communities with the aim of (i) understanding how attitudes might differ by smoking status, (ii) exploring the attitudes of smokers and exsmokers in-depth and (iii) identifying factors that could be targeted in screening communication strategies.

| Mixed methods design
Study one used a quantitative survey design to compare smokers' beliefs about lung cancer screening with those of former and never smokers, recruited from lower SES communities. Study two interviewed a subset of survey participants to explore the views of smokers and former smokers in-depth, to investigate factors that might be important for screening participation. Ethics approval was granted by the UCL research ethics committee (reference: 5210/001).

| Recruitment methods
Proactive, community-based strategies were used to recruit individuals from lower SES communities, with the aim of involving individuals in research who might be less likely to participate in research carried out using traditional recruitment methods, central to understanding nonparticipation. Individual people leading activities, organizations, venues or development work in socioeconomically deprived areas were identified from online searches, council listings and literature displayed in community centres. Selections were made on the basis that they (i) had networks with relevant local groups; (ii) provided free support and outreach services to those living in deprived conditions; (iii) worked within a manual workplace where smoking rates are higher; (iv) provided social housing; and (v) were not providing a health service. A researcher (SLQ) met with interested community leaders to seek advice on recruitment approaches and to access their networks to identify other recruitment opportunities. The recruitment approach was two-pronged for both studies: (i) participants were approached directly in community settings by SLQ and (ii) two housing associations mailed the survey to tenants.

| 565
Recruitment settings ranged from organized community-based services (e.g. free drop-in advice service for housing issues) to locations with a high footfall (e.g. bus station and markets).

| Participants
Adults aged ≥40 years were eligible. Adults younger than the screening-eligible age (aged 55-80 years 6 ) were included because they are approaching eligibility and are representative of the generation who would comprise the first screening cohort in the United Kingdom, pending recommendation of screening. Never smokers were included for comparative purposes, because former smokers were once current smokers and may share characteristics associated with smoking uptake. The inclusion of never smokers also intended to provide insight into the wider social context within which screening would be considered.

Participants were recruited from lower SES communities in
Central and South-East London. As an example, in one of the South-East London boroughs, population statistics indicate that 86% of residents are classified within the two most deprived Index of Multiple Deprivation (IMD) quintiles. 28 Smoking prevalence is highest within these quintiles: 33% of men and 26% of women smoke in the most deprived IMD quintile, and 26% of men and 20% of women smoke in the second most deprived IMD quintile. 29

| Measures
A survey was constructed using items adapted from existing measures of cancer beliefs (for lung and other types 24,[30][31][32][33][34], and original items were developed based on the findings of an existing qualitative study. 27 The inclusion of items was therefore driven by the available literature rather than a specific theoretical model. The survey began with a brief description of screening, which explained that the test aimed to find lung cancer at an early stage when there was a better chance of cure, and that it used a "type of X-ray called a CT scan." Participants were then asked to rate their agreement (strongly disagree/disagree/agree/strongly agree/don't know) with each item. Items covered the following topics in relation to lung cancer or lung cancer screening: smoking, worry and avoidance, perceived benefit, lung cancer outcomes, perceived risk, practical barriers and the need for screening without symptoms.
Smoking data were also collected, including self-reported smoking status, age started daily smoking and maximum number of cigarettes smoked daily (to calculate pack-year history), number of quit attempts lasting ≥3 months, age of quitting and quit confidence (on a scale of 0 to 10). Demographic characteristics included sex, age, marital status, ethnicity, education, postcode (to calculate IMD rank 35 ) and employment status.
The survey was paper-based and designed to be self-completed.
SLQ assisted participants with completing the survey if they had difficulty or preferred it to be read aloud.

| Analyses
Univariate chi-square analyses and Fisher's exact tests were carried out to explore associations between smoking status and item agreement. Multivariate analyses were not carried out because there were too few cases to provide adequate statistical power. For the same reason, we did not lower the statistical significance threshold to adjust for multiple testing, but report the respective significance thresholds.

| Participants
After each participant had completed the survey, the researcher screened their responses to purposively sample a subgroup of interviewees who had indicated they were current or former smokers, and from lower SES backgrounds (as indicated by an area-based measure of deprivation or their education level). Arrangements for interview were made immediately with participants who indicated they were willing to be interviewed, to minimize attrition. Therefore, recruitment for interviewing and the survey began at the same time and interviews ceased when no new information was gleaned from the data.

| Measures
Semi-structured telephone interviews were carried out, during which participants were asked open-ended questions on three main topics: (i) benefits/barriers to screening (e.g. "Can you talk me through your reasons for going/not going [to screening]?"), (ii) the preferred approach for screening invitations (e.g. "Often, letters from the doctor point out that the person is a smoker. How do you feel about that?") and (iii) attitudes towards provision of smoking cessation support (e.g. "Smokers may also be asked if they would like advice or help with stopping smoking at their appointment. How do you feel about that?"). At the start of the interview, the same brief explanation of screening was given as in Study one, but participants were also asked to imagine they had received a screening invitation from their GP.

| Analyses
Transcripts were analysed using an inductive approach to thematic analysis with NVivo qualitative data analysis software (QSR International Pty Ltd. Version 10, 2012). Familiarization with the data began during the interviews and continued through repeated reading of the transcripts. Data were first coded openly, with minimal interpretation. Themes and sub-themes were then interpreted and developed into a framework. Inclusion of themes depended on their frequency and contribution to the research question, that is factors that may be important when communicating a screening offer.
A second researcher (LAVM) read ten randomly selected transcripts independently, to cross-check the framework, and any disagreements were resolved through discussion and revisiting the transcripts.

| Sample characteristics
Most participants were approached directly in the community context by SLQ (n=14 interviewees; n=100 questionnaire respondents). The remaining number responded to a questionnaire mail-out from two housing associations (n=7 interviewees; n=75 respondents). Around three-quarters of individuals approached in person completed the survey and the majority of eligible participants (>90%) subsequently contacted for interview took part. The response rate for mailed surveys was very low (12%). Post-hoc analyses showed that participants recruited in-person were more frequently younger and current smokers (P<.01) and had a higher level of education (P<.05) than those who responded to the mail-outs. There were no differences by deprivation, sex, ethnicity or marital status.
In total, 175 participants completed the survey. Occasional smokers (n=7) and those reporting a diagnosis of lung cancer (n=5) were excluded from analyses. Of the remaining 163 participants, 28% identified as current smokers, 44% were ex-smokers and 29% had never smoked. Of these, nine smokers and 12 ex-smokers (N=21) were interviewed (see Table 1).
Men and women were evenly represented. Samples had a similar mean age (interviews: 62 years; survey: 64 years) and were mostly white (91% and 79%), unmarried or widowed (86% and 66%) and retired (both 52%). The majority were from lower SES backgrounds, as indexed by an area-based measure of deprivation (most deprived IMD quintile: 76% and 51%), and education (no formal qualifications/left school age ≤15 years: 71% and 55%). The higher proportion of lower SES interviewees compared with survey respondents was intentional and part of the purposive sampling frame. Many participants reported that they had experience of lung cancer through a "close other" (48% and 37%).

| Univariate analysis of survey data
The key findings are described here, but all univariate results are shown in Table 2 and Table S1.

| Beliefs about smoking and lung cancer screening
Current smokers were most likely to agree with some of the beliefs about smoking which conflict with participating in or benefiting from screening. Twenty per cent (n=9) believed they had "smoked too long to benefit" (vs 4%, n=3 of former smokers; P<.05) and almost a third (30%, n=13) agreed that "if the CT scan is negative you can continue to smoke without worrying about lung cancer" (vs 6%, n=4 of former smokers and 4%, n=2 of never smokers; P<.001; see Table 2).

| Perceived risk of lung cancer
More current smokers perceived their risk of getting lung cancer as high over "the next few years" (48%, n=21) than former (10%, n=7) and never smokers (7%, n=3; P<.001), and over a third (36%, n=16) agreed their lifetime risk "is higher than other smokers." Forty-four per cent (n=20) felt "I will get lung cancer sometime during my life" (vs 10%, n=7 and 17%, n=8, respectively, P<.001). Eighteen per cent (n=8) of smokers thought they "may already have lung cancer" but the proportion agreeing was not significantly different to former or never smokers (see Table 2).
However, smokers were no more likely to agree worry would deter them from screening or that they were "very scared of getting lung cancer", items endorsed across smoking groups (see Table 2).

| Lung cancer outcomes
Fewer than half (n=71) of the sample agreed "people with lung cancer can expect to continue with their normal activities" and over a fifth (22%, n=35) thought "treatment is worse than the lung cancer itself." Smokers were most pessimistic about survival, with almost half agreeing lung cancer is "a death sentence" (48%, n=21 vs 13%, n=9 of former smokers and 11%, n=5 of never smokers, P<.001; see Table 2).
However, there was no significant association with smoking status for agreeing that "lung cancer can often be cured."

| Online supplementary results
Additional results on beliefs about perceived benefits of screening, avoidance, symptoms, and practical barriers are provided in Table   S1 in the interests of space. Briefly, there were almost no significant associations with smoking status. Overall, endorsement of perceived benefits was high (>64%), avoidant beliefs were held by around onefifth (18%-25%) and few endorsed practical barriers.  (14) 25.5 (12) .713* *Chi-square analyses, **Fisher's exact test. ***n=122.

| Thematic analysis of interview data
Themes comprise two categories: the first encompasses general attitudes to lung cancer and the second concerns screening-specific attitudes. Short illustrative quotes are included in the text and longer quotes in Tables 3 and 4. Reference codes represent participant number (P), sex (F/M), smoking status (X/S = ex-smoker/smoker) and age.

| Smoking: history, stigma and identity
Participants often explained or justified their smoking history, with many expressing "regret" and "guilt." Current smokers in particular emphasized the difficulties of stopping, feeling "trapped" by addiction, and their attempts at cutting down. Some pointed out that when their generation began smoking, it was seen as "glamorous" and the risks were not well publicized.
Smoking was commonly described as a "stigmatized" and "taboo" behaviour, for which smokers had been "singled out" and "picked on." Current smokers especially felt health professionals blamed them for their ill health and treated them unfairly, worrying that "you ain't gon-  Table 3).

| Uncontrollable disease: risk, survival and treatment
Lung cancer was described as a "killer" disease, which is "aggressive" and a "painful way to go" (P9, M, X, 50) by both current and former smokers. The vital nature of lungs was implicated in poor survival with some questioning how treatment is possible given that "you can't live without your lungs" (P9, M, X, 50). There was concern that treatment is ineffective and detrimental to quality of life. Some participants explained they would decline treatment, perceiving screening as something which "can never give you a new pair of lungs" (P1, M, X, 54).
Regardless of smoking status, many were concerned that there was little they could do to reduce their high risk of lung cancer, because of their smoking history. Some worried they may already have it: "you'll always be worried if you've been a heavy smoker" (P13, F, X, 66). To compound this lack of control, risk was perceived as unpredictable.
Interviewees cited the fact that a non-smoker can get lung cancer even if "they've never put a fag in their mouth" (P5, M, S, 56), and that "a lot of people smoke and still live to a ripe old age" (P16, F, X, 67). Suggestions of other risk factors further reduced perceived control, including genetics, pollution, asbestos, poor housing, workplace exposures, stress and "cancer-grown foods" (P9, M, X, 50). Not smoking was therefore not necessarily perceived as protective and many attributed their risk to chance. A minority underplayed smoking's role in lung cancer risk as something that "doesn't help" (P7, F, X, 57), or "a load of old toffee" (P5, M, S, 56) (see Table 3).

| Life circumstances: poor health, life experiences and addiction
Poor health and comorbidities were commonplace among participants' family and friends, which included other cancers, smoking-related diseases (especially COPD) and chronic diseases (e.g. diabetes). Vicarious experiences of others' suffering with lung cancer were commonly referenced in relation to screening attitudes. Difficult life circumstances beyond health were also mentioned and implicated by some in a pessimistic outlook on life. Some interviewees alluded to their struggles with other addictions (see Table 3).

| Support for screening: early detection, reassurance and preparation before death
Participants were positive in principle about screening, and several indicated that they would participate; often because they perceived benefit in knowing whether they have lung cancer. Early detection was most commonly cited with regard to "less major treatments" (P1, M, X, 54), "better chance of cure" (P14, M, X, 73) and leading "a normal life for longer" (P9, M, X, 50), but notably more frequently by former smokers than current smokers. Furthermore, the language used to convey early detection principles was typically cautious about the ability to survive lung cancer in the long-term. Some appeared to pay "lip service" to the idea without real conviction, doubting the benefits: "whether that's true, I'm not too sure" (P7, F, X, 57). Instead, some participants perceived the benefit of early detection to be "halting the growth" (P23, M, X, 66) until a cure became available, indicating a lack of awareness that early stage treatment can be curative.
Some participants welcomed the potential for reassurance that that they did not have lung cancer and motivation to stop smoking if given a "clean slate." Others saw screening as an opportunity to prepare family and personal affairs before they died. The importance of family was also apparent in the reasoning that "everyone wants to be around for their families" (P13, F, X, 66). One smoker began the interview resolutely against screening, but thought differently when considering his grandchildren (see Table 4).

| Fear: diagnosis, hospitals, getting sick and death
Fear of diagnosis was commonly suggested as a deterrent to screening due to the expectation of a positive result, particularly for current smokers. For some, the decision to attend was interpreted as a decision to "find out" (P4, F, X, 58) about their lung cancer. This fear also appeared to originate from concerns around the whole care pathway, including getting sick, hospitals as a slippery slope from which the sick do not return and ultimately death. (see Table 4).

| Avoidance: rather not know and wait until sick
Not wanting to know was described as a reason to avoid screening, especially for current smokers, although most interviewees also said they would participate. Attending lung cancer screening was likened to "facing the music" (P18, M, S, 47), and even discussing the offer with a friend was perceived as difficult, with one participant avoiding even the word cancer, referring instead to "the big C" (P12, M, S, 68). Avoidance was commonly explored through the third person, perhaps due to social desirability, with participants describing instances of friends or family avoiding seeking medical help for lung cancer symptoms until they were unwell (see Table 4).

| Too late: smoking damage and older age
The perception that irreparable smoking damage had already been done was common and related to the belief that it is too late to prevent lung cancer or change the course of their respiratory health.
Smokers in particular discussed how the "damage is already done" (P3, M, S, 63) and appeared not to see screening as an opportunity to regain control over their risk of lung cancer mortality because "once you've smoked more than 10 years you think if I've got it, there's nothing I can do" (P4, F, X, 58). One ex-smoker described the process of seeing smoking damage inside the lungs as "like going to a mortuary" (P1, M, X, 54).
Tellingly, while most thought that screening should be available to all, some believed it would be of greatest benefit to younger adults, due to their shorter smoking history and ability to "withstand any treatment" (P23, M, X, 66). The benefit of screening for older smokers "I worked in a car plant… everything was covered in powder so you were breathing that… eight hours a day, so that is a concern as well." P9, M, X, 50 "they think, "I'll pack up smoking and live another ten years. T A B L E 3 (continued) (Continues) was described as relatively low, which is contradictory to the screening eligibility criteria. Related to this, some participants attributed their risk of lung cancer, and life expectancy more generally, to chance (see Table 4).

| Screening: approach and support
Participants emphasized the need to provide a non-judgemental screening service and suggested phrasing the invitation in general terms (i.e. in relation to an age group not smoking status), to normalize the offer. Others warned mentioning smoking cessation would be off-putting to smokers and unnecessary, because this advice would be expected.
Some thought there should be interpersonal support preceding and during the appointment including "a family friend that could support them" (P17, F, S, 61), "a chat beforehand just to basically find out what… the procedure entails" (P7, F, X, 57) and "a phone number they could ring" (P4, F, X, 58). Others noted that the invitation wording should be considered carefully so as not to imply lung cancer is specifically suspected (see Table 4).

| DISCUSSION
To our knowledge, this is the first study to explore attitudes towards lung cancer screening among smokers and ex-smokers from socioeconomically deprived communities using proactive recruitment methods to engage a "hard-to-reach" sample. The majority were superficially supportive of screening, but interviewee accounts were often contradictory, acknowledging early detection could be beneficial, but alongside negative views of treatment and survival, especially for older smokers. The findings gave insight into the salience of fear, avoidance, fatalism and stigma around lung cancer screening, particularly for smokers (as opposed to former and never smokers) which could deter participation.
T A B L E 4 Quotes illustrative of the screening-specific attitude themes (Interview Study two) Participant Support screening: early detection, preparation, and reassurance "a cure could be round the corner, and if you're diagnosed early enough and being able to be sort of er put on hold… a cure might appear" Screening: approach and support "if they think someone is going to be there wagging the finger at them, that's it, it's out the window…"If you point at people and say, "Look, you've been picked out and we think you ought to be screened," that will frighten them to death" The level of fear of lung cancer is consistent with studies describing higher cancer fear among low SES groups. 36 Fear discussed during the interviews most commonly concerned an expected diagnosis at screening due to acknowledgement that a significant smoking history put them at high risk. Survey data showed that this fear was substantially more common among smokers. Almost three-quarters agreed that they often worry about their chance of getting lung cancer, compared with a quarter of former and never smokers, an interestingly similar proportion given that former smokers are at greater risk than never smokers. Furthermore, a preference "not to know" was implicated in smokers' avoidance of screening by interviewees. Cancer fear has been shown to promote and deter screening uptake in other programmes, 37 but worry about what the doctor might find is a commonly reported reason by smokers for delaying symptomatic presentation. 38 Therefore, emphasizing risk in screening communication could be counterproductive and unnecessary. Attempts to target risk perceptions using written information prior to screening have been unfruitful, with some evidence this strategy polarizes risk perceptions. 39 Also, most smokers perceive their risk of lung cancer as higher than never smokers, 40 with over a third we surveyed agreeing that their risk was higher than other smokers, and close to one-fifth thinking they may already have lung cancer, more than former and never smokers. In a US study, 41 smokers' awareness of their increased risk was not associated with screening interest, and in the UKLS trial, higher affective risk perceptions were associated with non-participation. 26  ing has been observed previously, 27 and fatalism towards cancer has been associated with avoidance of early detection 42 and later stage at diagnosis. 43 Unfortunately, a poor prognosis for lung cancer is the most common outcome due largely to late diagnosis, 3 and it will be difficult to overlie this with the claim for benefits of early detection. A lay explanation for early stage treatment may help, one that addresses concerns and challenges the perception that the removal of a whole lung is necessary for treatment to be effective. This could be achieved by emphasizing the targeted nature of early treatment by showing the lungs subdivided into lobes.
Qualitative findings also gave insight into the roles that perceived stigma and tobacco dependence may play. Regret, guilt and feelings of entrapment by smoking were commonly expressed. The stigma attached to lung cancer as a self-inflicted, smokers' disease was especially apparent, which has been reported previously, 44 and implicated as a deterrent to screening by a US study of long-term smokers. 45 Addiction is a disorder of motivation 46  are also needed that explore the association between these beliefs, screening intentions and ultimately screening behaviour. A randomized controlled trial 51 is underway testing targeted invitation materials designed to improve uptake, which were informed by the results of this study.
In conclusion, attitudes towards lung cancer screening among lower SES smokers and former smokers are complex, but smokers appear to be the most negative when compared with former and never smokers. Many participants appeared to be superficially positive about screening and the benefits of early detection, but more detailed discussions identified feelings of loss of control of respiratory health and risk of lung cancer, fear and fatalism about diagnosis, blame and the belief that they would be unlikely to benefit personally. Addiction, difficult life circumstances and negative perceptions of treatment appeared to exacerbate this pessimism and lack of control. The approach of lung cancer screening is distinct in targeting individuals based on a highly stigmatized and widely publicized risk behaviour, and the expectation of a diagnosis among smokers appears to complicate decision-making. To achieve an equitable and effective lung cancer screening programme, invitation strategies need to be evidence-based and designed in consultation with low SES smokers to ensure they are sensitive to these complex attitudes.

ACKNOWLEDGEMENTS
We would like to thank all our study participants and our communityacademic partners for their help with recruitment, including but not

SOURCES OF FUNDING
The study was funded by a National Awareness and Early Diagnosis

ROLE OF THE FUNDER
Funders have had no role in the study design, the collection, analysis, interpretation of data, the writing of the article or the decision to submit it for publication.