Abstract
Introduction
Malignant melanoma is one of the fastest increasing cancers in the United Kingdom, with incidence rates quadrupling in the past 30 years. 1 It is the leading cause of death from skin cancer 2 and is mainly caused by ultraviolet light exposure. 3 One route of exposure is through sunbed use 3 and it has been estimated that each year in the United Kingdom, 100 deaths from melanoma can be attributed to this. 4 A meta-analysis concluded that use before the age of 35 years increases the risk of melanoma by 75%. 5
Despite the clear evidence of harm, children have been reported to use sunbeds. 6 -8 Consequently, the Sunbeds Regulation Act 2010 was implemented in April 2011 to ban use in under-18s. 9 However, the legislation only bans provision in commercial settings. 9 It is possible that under-18s may still have access in domestic settings, as was the case before the ban. 8 Therefore, there is now a need to explore underage use beyond the remit of the legislation.
In addition, the ban may not directly influence children’s attitudes and knowledge of sunbed risks. Shortly after the ban, a survey found that sunbeds remained popular among 13- to 24-year-olds and some had incorrect knowledge such as believing sunbeds can prevent sunburn. 10 Knowledge of risks may not translate to behavioral changes. In a Cancer Research UK qualitative study, all underage users were aware of the risks but rationalized that they were not the heaviest users. 11 This qualitative study has generated useful insights but further quantitative data are needed. 11
To date, no study has explored the impact of the ban on behavior since its introduction. The Sandwell Primary Care Trust commissioned this study to investigate underage sunbed use in Sandwell after the ban. As the impact on use cannot be directly measured because of the lack of comparative data specific to Sandwell before the ban, this study will provide a baseline prevalence for future comparisons to assess the long-term impact. The study aims to determine the following among under-18s:
the prevalence of sunbed use after the ban,
awareness of the ban,
tanning attitudes and knowledge of sunbed risks, and
factors associated with use/intention.
With better understanding of the current use, attitudes, and knowledge of young sunbed users, public health authorities can develop targeted strategies in reducing underage use alongside the commercial ban.
Methods
This was a cross-sectional survey of adolescents in schools with anonymous, self-completed, written questionnaires in class.
The study was set in Sandwell, a metropolitan borough in the West Midlands, United Kingdom. It is the 12th most socioeconomically deprived English local authority area 12 with a high proportion of ethnic minorities (23.3%). 13 All 22 state secondary schools in Sandwell 14 were sent an invitation letter.
A convenience sample was recruited; inclusion criteria were adolescents aged 15 to 17 years available on the day, and exclusion criteria were adolescents beyond this age group and those who declined participation. Adolescents aged 15 to 17 years were surveyed as they have been shown to have the highest sunbed use in previous research (11.2%). 8 Assuming this figure would halve after the ban, a sample size of 352 was required to estimate the prevalence of use with ±2% accuracy at the 90% confidence level. Data were collected from February 22, 2012 to March 30, 2012.
The study received ethical approval from the University of Birmingham BMedSc Internal Ethics Review Committee. Schools agreed for adolescents to be surveyed and participants provided written consent.
Questionnaire
A questionnaire was designed with reference to previous UK studies. 7,8,11 Although not previously validated, it was piloted in 84 school adolescents after which changes were made to improve clarity.
The main outcome was adolescents’ sunbed use after the introduction of the ban. This was defined as having used one within the past 12 months. Participant demographics collected were age, gender, ethnicity, self-defined skin type (single item incorporating skin color, degree of tanning, and burning), 15 and current program of study. Other measures include pattern and reasons for use (5 items), tanning attitudes (7 items), knowledge of sunbed health risks (6 items) and skin cancer risk factors (7 items), 16 social factors (family and friends’ sunbed use), and ban awareness.
Attitudes and knowledge of sunbed risks were measured by the percentage of participants who indicated “agree/disagree/neutral/don’t know” for related statements. A summary score for skin cancer risk factors knowledge was calculated by awarding 1 point for each risk factor identified (0-6 points possible).
The questionnaire was self-completed in class in a 15-minute session. This was conducted under exam conditions to avoid discussion amongst participants. Names were not enquired and students were reminded that they will not be identified.
Analyses
Participants were categorized into users, potential users, and nonusers. Potential users were those who have not used, but indicated intention to do so in future. Because of low numbers, users were combined with potential users into “users/potential users” for subsequent analyses, ethnicities were recategorized into “white” and “other,” and current programs of study were recategorized into “GCSE” (General Certificate of Secondary Education) and “other.”
Prevalence of use was compared between demographic subgroups using 95% confidence intervals (CIs). Attitudes and knowledge were compared between users/potential users and nonusers using the χ2 test. As the total scores for skin cancer risk factors knowledge were not normally distributed, comparisons were made using the Mann–Whitney U test.
Univariate and multivariable analyses were conducted to identify factors associated with use/intention. Binary logistic regression was used to calculate the odds ratios (ORs) and 95% CIs for individual variables. Subsequently, all variables were entered in the hierarchical regression model. These include demographics, attitudes, knowledge, social factors, and ban awareness. As age and gender are known to be associated with sunbed use, 7,8,17 they were entered in the first block; the remaining variables were entered in the next block via the backward stepwise variables selection method.
Analyses were performed using Statistic Package for the Social Sciences, version 19. Descriptive statistics were calculated out of the total responses for each item. P < .05 was considered statistically significant. Where multiple items were tested, the Bonferroni-corrected α levels were used.
Results
Participant Demographics
Of the 22 Sandwell secondary schools, 5 (23%) participated. All school types were represented (academy, college, further education college, referral unit).
Overall, 428 adolescents were available on the days of surveys and 407 responded (95.1%). Nineteen participants with missing data for age were excluded from the analyses. Gender was equally distributed (50.9% females). The majority were aged 15 years (60.3%) and the proportion of white ethnicity (67.2%) was lower than the Sandwell population (76.7%). 13 Other participant demographics are summarized in Table 1.
Participant Demographics and Prevalence of Self-Replied Sunbed Use a in 388 Adolescents in Sandwell Schools by Gender, Age, Ethnicity, Education, School, and Skin Type.
Abbreviations: CI, confidence interval; GCSE, General Certificate of Secondary Education; ESOL, English for Speakers of Other Languages; n, total number of responses for each item (proportions are rounded and may not add up to 100, denominators are total responses for each item).
Nine participants did not answer the question on whether they have used sunbeds.
Includes mixed ethnicities, Chinese, and other ethnicities.
Includes Entry Level, Information Technology Users Qualification (ITQ) and Business and Technology Education Council Qualification (BTEC).
Surveys were conducted by the researcher in Schools A, B, and C, and by teachers in Schools D and E.
Sunbed Use and Intention to Use
A small proportion of participants had used sunbeds after the ban (n = 20, 5.3%, 95% CI = 3.4-8.0). One fifth indicated that although they have not used, they may do so in the future (n = 78, 20.6%, 95% CI = 16.8-24.9). Table 1 presents the prevalence by demographic subgroups.
Use was higher in “Hair and Beauty” students (26.8%) and English for Speakers of Other Languages (ESOL) students (30.8%) and both groups were from School C. Excluding this school, the adjusted prevalence was 1.7% (95% CI = 0.7-3.9) and there were no users in the 17-year-old age-group.
Of the 20 users, most had used sunbeds in business premises (n = 16), whereas a quarter had used them in domestic settings (n = 5; multiple options were allowed). Five used once a week or more, 5 used 1 to 3 times a month, 6 used once or a few times annually, and 4 reported one-off use. The median for reported time spent on a sunbed was 9.0 minutes (interquartile range = 6.5-15.0), with 6 spending longer than the recommended 10 minutes per session. 18
The most frequently quoted reasons for use/intention (n = 94) were to look tanned (67.0%), to look good for social events (34.0%), to treat acne/spots/skin conditions (23.4%), and to prevent sunburn on holidays (19.1%; multiple options were allowed).
Tanning Attitudes
Tanning attitudes differed significantly between users/potential users and nonusers (Table 2). More users/potential users associated a tan with positive attributes, with the majority associating it with looking healthy, confidence, and attractiveness.
Adolescents’ Tanning Attitudes, Knowledge of Sunbed Risks and Skin Cancer Risk Factors, and Ban Awareness by Sunbed Use Status in Sandwell Schools a .
Abbreviations: CI, confidence interval; IQR, interquartile range.
Where figures do not add up to 100, these indicate the proportions of students choosing the “don’t know” or “neutral” options. Denominators are total responses for each item. Although combined for the purpose of comparisons, users and potential users are presented separately in the table for clarity and to demonstrate trends. “√” denotes the desirable answer.
The Bonferroni-corrected α level is P < .007.
The Bonferroni-corrected α level is P < .008.
Mann–Whitney U test to compare the median scores between users/potential users and nonusers.
χ2 test to compare the proportions of participants who were aware/unaware of the ban between users/potential users and nonusers.
Knowledge of Sunbed Health Risks
Overall, there appears to be good knowledge of risks (Table 2). However, users/potential users had less awareness of the risks compared with nonusers. Only 67.7% of users/potential users were aware that sunbeds can cause skin cancer (vs 83.3% of nonusers), 39.1% believed that using once a week is safe. Just more than half of all participants were aware that sunbeds can cause premature skin aging (56.5%).
Awareness of the Skin Cancer Risk Factors and the Under-18s Ban
The median score for knowledge of skin cancer risk factors was 3 (interquartile range = 2-5; 0-6 points possible). Most were aware that sensitive skin that burns easily is a risk factor (82.7%), whereas red hair (36.6%) and freckles (44.0%) were the least recognized. Worryingly, only 48.2% were aware that fair skin is a risk factor. Less than half of all participants were aware of the under-18s ban (48.2%, 95% CI = 43.2-53.3) and this was similar among users (44.4%). Awareness of the skin cancer risk factors and the ban did not differ significantly between users/potential users and nonusers (Table 2).
Factors Associated With Sunbed Use/Intention to Use
Table 3 presents the univariate and multivariable analyses.
Univariate and Multivariable Analyses of Factors Associated With Sunbed Use/Intention to Use in Adolescents in Sandwell Schools.
Abbreviations: OR, odds ratio; CI, confidence interval; GCSE, General Certificate of Secondary Education; ESOL, English for Speakers of Other Languages; “√” denotes the desirable answer.
Includes Asian, black, and other ethnicities.
Includes ESOL, Entry Level, Information Technology Users Qualification (ITQ), and Business and Technology Education Council Qualification (BTEC).
The Bonferroni-corrected α level is P < .007.
The Bonferroni-corrected α level is P < .008.
For the hierarchical regression model, variables entered in the first block were age and gender. Variables entered in the next block include ethnicity, program of study, school, skin type, tanning attitudes, knowledge of sunbed risks, total score for skin cancer risk factors knowledge, ban awareness, and family and friends’ sunbed use. Nagelkerke R 2 = .572, percentage of correct classification = 84.3%; 308 participants were included in the analysis.
Only significant variables (P < .05) are shown in the table.
In the univariate analyses, factors significantly associated with use/intention include being female, studying Hair and Beauty or ESOL, associating a tan with positive attributes, and having family or friends who are sunbed users. Belonging to an ethnic minority and knowledge of risks were negatively associated with use/intention.
When all variables were adjusted for, 3 independent factors were found to be significant: female gender, family or friends’ sunbed use. The odds for use/intention were doubled among females (OR = 2.57, 95% CI = 1.12-5.91), increased 3-fold with family members’ use (OR = 3.53, 95% CI = 1.52-8.19), and 4-fold with friends’ use (OR = 4.54, 95% CI = 1.94-10.63).
As all Hair and Beauty and ESOL students were from School C, this may have contributed to the high prevalence of sunbed use in this school (Table 1). However, when programs of study and schools were entered in the multivariable model, they were not found to be significant factors and neither were there concerns for collinearity.
Discussion
Main Findings of This Study
Some adolescents still appeared to be using sunbeds despite the implementation of the under-18s ban in April 2011 but the adjusted prevalence was low. However, sunbed use was considerably higher in certain programs of study. Less than half of the participants were aware of the ban and a quarter of underage use occurred in domestic settings that are beyond the remit of the commercial ban. Furthermore, 1 in 5 adolescents expressed intention to use sunbeds in the future. Use/intention to use was increased 3- to 4-fold if family members or friends use sunbeds.
What Is Already Known on This Topic
Before the ban, the English National Study reported that 11.2% of 15- to 17-year-olds had used sunbeds, 14.1% had the intention to use in the future, and about a quarter of underage use occurred at home (23.0%). 8 A systematic review on correlates of sunbed use concluded that this was more likely in female, late adolescence, white ethnic group, and with parental use. 19 Knowledge of the potential skin cancer risks was high among users. 11,17 However, certain misconceptions existed such as believing that sunbeds can prevent sunburn and using once a week is safe. 10
What This Study Adds
This is the first study to investigate the impact and awareness of the ban among under-18s since its introduction. Although direct comparisons cannot be made, the sunbed use prevalence in this study was lower than the national figure before the ban. 8 In contrast, intention for use in the future was higher 8 presumably as adolescents may plan to use sunbeds once they have passed the age limit. The extent of underage use in domestic settings was similar. 8 Supporting previous findings, users were more likely to be female 7,8,19 and have family or friends who use sunbeds. 7,19,20
Notwithstanding the legislation, adolescents in Sandwell were still using sunbeds. A recent inspection using the underage secret shopper scheme in Wales and Sandwell found that 32% of tanning salons visited did not check the customers’ age. 21,22 This raises the question of how well the ban has been implemented. The low awareness of the ban among adolescents indicates that current publicity is not reaching this age-group effectively and targeted strategies need to be explored further.
This study also supports the need for parental education as a considerable proportion of participants reported sunbed use at home. Furthermore, family members’ sunbed use was associated with underage use. It also highlights the gap in knowledge of risks associated with sunbed use amongst adolescents that may be addressed via school-based education, which may also help reduce peer influence. Only half of the participants knew that sunbed use can accelerate skin aging. As motives for sunbed use were driven by appearance considerations, this is a potentially useful health message.
As sunbed use was significantly higher among Hair and Beauty students, campaigns should target students engaged in related courses, especially as they may pass on this message when they start working in the beauty industry.
Limitations of This Study
The study relied on participants’ self-reported sunbed use. However, the self-administration ensured anonymity and aimed to reduce social desirability bias.
As a convenience sample was used, the resulting disproportionate age and ethnicity distribution may mean difficulty in generalizing the findings. However, the attitudes and knowledge of participants in this study may be shared by adolescents of the wider population as related findings were similar to previous studies. 7,10,11
Although data on social grade were not collected, this was assumed to be similar throughout the sample as Sandwell is a uniformly deprived area. 12 The proportion of white ethnic group was lower than that of the Sandwell population (67% vs 77%). 13 As this group was known to have higher sunbed use, 19 the prevalence may have been underestimated.
In schools that may have participated because they have more interest in health education, students may be more health aware and be less likely to use sunbeds. In contrast, schools may have participated as sunbed use was known to be high among their students. The resulting responder bias may therefore have affected the prevalence either way.
Despite these limitations, it is clear that adolescents still have access to sunbeds, and this calls for stricter enforcement and greater publicity of the ban alongside education programs targeting parents and adolescents in schools.
Footnotes
Acknowledgements
The authors would like to thank Helen Levett and Shane Middleton (Sandwell Metropolitan Borough Council Environmental Health Officers) for providing the list of secondary school contacts, Dr Sayeed Haque and Mr Roger Holder for advice on statistical analyses, Julie Shore for administrative support, head teachers, and students of participating schools.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the University of Birmingham and Siang Lee was supported by a Wolfson bursary awarded by the Royal College of Physicians.
Author Biographies