Factors Associated With Sun Protection Compliance
Factors Associated With Sun Protection Compliance
We performed a nationwide cross-sectional study to determine factors associated with compliance to sun protection measures in a population of patients who consulted a dermatologist, regardless of the way they were previously educated or informed regarding sun protection behaviour. From June to August 2010, an anonymous self-administered multiple-choice questionnaire was distributed to patients to whom a dermatologist had prescribed sunscreen. As this study did not involve manipulation of the subject or the subject's environment, institutional review board submission was waived.
Investigators were asked to recruit 10 consecutive patients for study inclusion. The questionnaire, designed by a multidisciplinary team including dermatologists, statisticians and quality-of-life and health economics experts, explored the following dimensions: reason for sunscreen prescription, demographic and clinical characteristics, compliance with current recommendations concerning sun protection, knowledge about correct sun protection behaviour and UV-associated risks, and level of exposure to UV radiation. In total 3470 questionnaires were distributed throughout France by 347 office-based dermatologists during patient consultations. The dermatologist recorded the reason for prescribing sunscreen and asked the patient to complete the second part of the questionnaire independently (see Supporting Information http://onlinelibrary.wiley.com/store/10.1111/bjd.12966/asset/supinfo/bjd12966-sup-0001-DataS1.docx?v=1&s=756ccee3c3ba0d8f83f305d32f6554b4422e8e0a). Patients mailed completed questionnaires to the research team. For each item, missing data were excluded from the analysis.
In order to quantify patient answers and normalize the population, four scores were established prior to the study: (i) compliance with sun protection measures (sun protection behaviour); (ii) knowledge about accurate sun protection recommendations (sun protection knowledge); (iii) level of UV exposure (UV exposure); and (iv) knowledge about UV-associated risks (UV knowledge).
Sun Protection Behaviour and Sun Protection Knowledge Scores. Calculation of the 'sun protection behaviour score' was based on a 22-item multiple-choice questionnaire: 'Indicate which measures you put into practice' to protect from UV. Items assessing recommended sun protection measures scored 1 point when declared as practised by the patient.
Calculation of the 'sun protection knowledge score' was based on the same 22-item multiple-choice questionnaire as above, but with the heading: 'Which among the following measures will protect you from UV?' Similar to the 'sun protection behaviour score', items representing recommended sun protection measures scored 1 point when selected correctly by the patient.
For the sun protection behaviour and sun protection knowledge scores, items corresponding to measures for which the effectiveness is unproven (e.g. using self-tanning products) scored 0 points, and items concerning measures proven to be dangerous (e.g. using tanning beds and not applying any special measures) were worth −0·5 points when mentioned by the patient. The maximum possible score was 15 points.
Ultraviolet Exposure Score. Calculation of the 'UV exposure score' was based on a selection of 15 leisure and daily life activities, some of which are associated with UV exposure and some not. Participants were asked to list commonly practised activities. Items assessing genuine UV exposure, such as outdoor activities or occupational use of devices generating UV radiation, scored 1 point when declared as practised by the patient. Items assessing false UV exposure (e.g. indoor activities) or situations in which UV exposure is not proven were worth 0 points. The maximum possible score was 12 points. A patient was considered as being subject to significant UV exposure when the score was > 3 points.
Ultraviolet Knowledge Score. Calculation of the 'UV knowledge score' was based on 25 questions (10 questions on knowledge about UV and 15 questions on situations of UV exposure), evaluating knowledge about UV-associated risks in the following dimensions: nature of the risk, situations in which the patient is exposed to the risk, and health-related and aesthetic consequences of UV exposure. Each correct answer scored 1 point. Items assessing a situation in which UV exposure is not proven scored 0 points. The maximum possible score was 22 points.
In order to determine factors associated with better-reported compliance with sun protection recommendations, the population was divided into three groups based on the distribution of the sun protection behaviour score. Group A comprised patients with the lowest score (from the minimum observed score to the 25th percentile), group B included patients scoring from the 25th to the 75th percentile, and group C included patients with scores above the 75th percentile. Based on this distribution, a range for the sun protection behaviour score was defined for each group: 0·5–3 (group A), 3–9 (group B) and 9–15 (group C).
Between-group quantitative variables were compared using Student's t-test (two groups) or anova (more than two groups). If the conditions required for these tests were not met, nonparametric Wilcoxon and Kruskal–Wallis tests were performed. Qualitative variables were compared using a χ or Fisher's exact test. A two-sided significance level was fixed at 5%. The Tukey method was used to establish significance for pairwise comparisons of mean scores. Statistical analyses were performed using SAS software version 9.2 (SAS Institute Inc., Cary, NC, U.S.A.).
For convenience and ease of interpreting the scores, all evaluations (sun protection behaviour, sun protection knowledge, UV exposure and UV knowledge) were grouped by the distribution of the values (according to 25th and 75th percentiles). These three groups may be interpreted as having low, medium and high scores of the variable under consideration. Firstly, each variable was evaluated independently in a bivariate analysis to identify factors associated with sun protection behaviour. Secondly, all significant variables at a 10% level were entered into a proportional odds model, as sun protection behaviour consisted of three ordinal classes. Unfortunately, the assumption of proportional odds was violated (P < 0·0001). Consequently, we performed a polytomous logistic regression; variables were retained in a stepwise manner in order to determine variables that were independently associated with sun protection behaviour at a probability threshold of 0·05. Odds ratios (ORs) and the corresponding confidence intervals (CIs) were generated, with the middle score used as the reference: low vs. middle score, high vs. middle score.
Patients and Methods
Study Group
We performed a nationwide cross-sectional study to determine factors associated with compliance to sun protection measures in a population of patients who consulted a dermatologist, regardless of the way they were previously educated or informed regarding sun protection behaviour. From June to August 2010, an anonymous self-administered multiple-choice questionnaire was distributed to patients to whom a dermatologist had prescribed sunscreen. As this study did not involve manipulation of the subject or the subject's environment, institutional review board submission was waived.
Investigators were asked to recruit 10 consecutive patients for study inclusion. The questionnaire, designed by a multidisciplinary team including dermatologists, statisticians and quality-of-life and health economics experts, explored the following dimensions: reason for sunscreen prescription, demographic and clinical characteristics, compliance with current recommendations concerning sun protection, knowledge about correct sun protection behaviour and UV-associated risks, and level of exposure to UV radiation. In total 3470 questionnaires were distributed throughout France by 347 office-based dermatologists during patient consultations. The dermatologist recorded the reason for prescribing sunscreen and asked the patient to complete the second part of the questionnaire independently (see Supporting Information http://onlinelibrary.wiley.com/store/10.1111/bjd.12966/asset/supinfo/bjd12966-sup-0001-DataS1.docx?v=1&s=756ccee3c3ba0d8f83f305d32f6554b4422e8e0a). Patients mailed completed questionnaires to the research team. For each item, missing data were excluded from the analysis.
Scoring
In order to quantify patient answers and normalize the population, four scores were established prior to the study: (i) compliance with sun protection measures (sun protection behaviour); (ii) knowledge about accurate sun protection recommendations (sun protection knowledge); (iii) level of UV exposure (UV exposure); and (iv) knowledge about UV-associated risks (UV knowledge).
Sun Protection Behaviour and Sun Protection Knowledge Scores. Calculation of the 'sun protection behaviour score' was based on a 22-item multiple-choice questionnaire: 'Indicate which measures you put into practice' to protect from UV. Items assessing recommended sun protection measures scored 1 point when declared as practised by the patient.
Calculation of the 'sun protection knowledge score' was based on the same 22-item multiple-choice questionnaire as above, but with the heading: 'Which among the following measures will protect you from UV?' Similar to the 'sun protection behaviour score', items representing recommended sun protection measures scored 1 point when selected correctly by the patient.
For the sun protection behaviour and sun protection knowledge scores, items corresponding to measures for which the effectiveness is unproven (e.g. using self-tanning products) scored 0 points, and items concerning measures proven to be dangerous (e.g. using tanning beds and not applying any special measures) were worth −0·5 points when mentioned by the patient. The maximum possible score was 15 points.
Ultraviolet Exposure Score. Calculation of the 'UV exposure score' was based on a selection of 15 leisure and daily life activities, some of which are associated with UV exposure and some not. Participants were asked to list commonly practised activities. Items assessing genuine UV exposure, such as outdoor activities or occupational use of devices generating UV radiation, scored 1 point when declared as practised by the patient. Items assessing false UV exposure (e.g. indoor activities) or situations in which UV exposure is not proven were worth 0 points. The maximum possible score was 12 points. A patient was considered as being subject to significant UV exposure when the score was > 3 points.
Ultraviolet Knowledge Score. Calculation of the 'UV knowledge score' was based on 25 questions (10 questions on knowledge about UV and 15 questions on situations of UV exposure), evaluating knowledge about UV-associated risks in the following dimensions: nature of the risk, situations in which the patient is exposed to the risk, and health-related and aesthetic consequences of UV exposure. Each correct answer scored 1 point. Items assessing a situation in which UV exposure is not proven scored 0 points. The maximum possible score was 22 points.
Composition of the Groups
In order to determine factors associated with better-reported compliance with sun protection recommendations, the population was divided into three groups based on the distribution of the sun protection behaviour score. Group A comprised patients with the lowest score (from the minimum observed score to the 25th percentile), group B included patients scoring from the 25th to the 75th percentile, and group C included patients with scores above the 75th percentile. Based on this distribution, a range for the sun protection behaviour score was defined for each group: 0·5–3 (group A), 3–9 (group B) and 9–15 (group C).
Statistical Analysis
Between-group quantitative variables were compared using Student's t-test (two groups) or anova (more than two groups). If the conditions required for these tests were not met, nonparametric Wilcoxon and Kruskal–Wallis tests were performed. Qualitative variables were compared using a χ or Fisher's exact test. A two-sided significance level was fixed at 5%. The Tukey method was used to establish significance for pairwise comparisons of mean scores. Statistical analyses were performed using SAS software version 9.2 (SAS Institute Inc., Cary, NC, U.S.A.).
For convenience and ease of interpreting the scores, all evaluations (sun protection behaviour, sun protection knowledge, UV exposure and UV knowledge) were grouped by the distribution of the values (according to 25th and 75th percentiles). These three groups may be interpreted as having low, medium and high scores of the variable under consideration. Firstly, each variable was evaluated independently in a bivariate analysis to identify factors associated with sun protection behaviour. Secondly, all significant variables at a 10% level were entered into a proportional odds model, as sun protection behaviour consisted of three ordinal classes. Unfortunately, the assumption of proportional odds was violated (P < 0·0001). Consequently, we performed a polytomous logistic regression; variables were retained in a stepwise manner in order to determine variables that were independently associated with sun protection behaviour at a probability threshold of 0·05. Odds ratios (ORs) and the corresponding confidence intervals (CIs) were generated, with the middle score used as the reference: low vs. middle score, high vs. middle score.
Source...