Sun protection in outdoor workers – Development and validation of standardized questionnaires for behavior and knowledge

Outdoor workers are at increased risk of developing non‐melanoma skin cancer. We aimed to address the lack of validated German‐language measurement instruments for outdoor workers’ sun safety behavior and knowledge by compiling and validating two questionnaires.

that sun exposure is reduced through technical and organizational measures 9 such as setting up sun sails or shifting outdoor working hours. 10In addition, employees need to apply individual protective measures.Targeted health education and training of persons at risk was shown to promote knowledge about sun protection, risk perception, motivation and, finally, the implementation of individual protective measures.However, the evaluation of the benefit of health education interventions and especially the association between protective behavior and skin cancer knowledge requires reliable measurement of protective behaviors adopted by outdoor workers (OWs), which are scarce up to now. 11revious studies have already investigated sun safety knowledge and behavior in German-speaking OWs or the general population.However, information on the validity of the applied questionnaires was often not provided, [12][13][14][15][16] or limited to an evaluation of internal consistency, [17][18][19] pre-survey, 20 cognitive interviews, 21,22 or expert review. 23nly two methodological studies investigated sun safety questionnaires in the German language.First, the Sun Exposure and Protection Index (SEPI) 24 has been validated but focusses on recreational sun safety.Second, the Skin Cancer and Sun Knowledge (SCSK) scale has recently been pilottested with a sample of students. 25However, questionnaires need to be validated among their target population.In the context of German-speaking OWs, no validation efforts have been reported to date.It is hence unknown which instruments are suitable for the investigation of sun protection in this vulnerable population.
This study therefore aims to compile and validate instruments for measuring sun protection behavior and knowledge in German OWs.In addition to their value in research, such questionnaires are also needed for developing, conducting and evaluating tailored health education programs.

Questionnaires
The questionnaires were composed in an iterative topdown process, including a literature search, consensus with an interdisciplinary panel, and pre-test (Figure 1, Online Supplementary Material S1).
Sun protection behavior (OccuSun): A set of core questionnaire items, developed and validated through cognitive interviews by an American workgroup of skin cancer prevention investigators formed the basis for this instrument. 26he workgroup encouraged researchers to investigate the reliability and validity of these measures in various research contexts.They also suggested to make minor adaptations based on the target population. 26Our expert panel agreed to adapt the questionnaire for OWs and include elements from further instruments [27][28][29][30] (Online Supplementary Material S2, Online Supplementary Figure S1).
Sun safety knowledge: We selected the Skin Cancer and Sun Knowledge (SCSK) scale, which comprises 25 items (Online Supplementary Material S3, Online Supplementary Figure S2).The English SCSK had already been validated for parts of the general population (undergraduate students, age 18-26 years). 31The authors recommend to investigate the validity of the SCSK across a range of participant groups. 31A Turkish version showed acceptable psychometric properties among nursing students. 32Recently, the SCSK has been translated and pilot tested with a sample of students of vocational education teaching. 25

Participants and procedure
The validation study was conducted between winter and autumn 2020 at the study sites in Dresden and Erlangen (Figure 2).The study protocol was approved by the ethic committees at the TU Dresden (EK113032016) and the Friedrich-Alexander-University (FAU) Erlangen-Nuremberg (4_20 BC).Participants were recruited in Dresden and Erlangen, Germany, by contacting outdoor employers.Adult workers, usually working outdoors for at least 2 hours per day between 11 a.m. and 4 p.m., were included.Due to restrictions during the COVID-19 pandemic, they were allowed to complete the questionnaires either at their workplace or at home without supervision.Participants who did not complete their questionnaires in time were reminded weekly by email, telephone, or e-mail unless they withdrew from the study.

Measures of reliability and validity
The details of the validity and reliability measures are provided in Supplementary Online Material S2-S3.An overall sum score of sun safety knowledge was computed.For sun protection behavior, we computed total and body region scores (Online supplementary Table S1).We evaluated the criterion validity of the assessed sun protection in comparison to self-reported sunburn frequency and increases in skin pigmentation, 33 objectively measured with a Minolta Chromameter CR200 at T1 and T3 (Figure 2).The construct validity was assessed based on two self-reported outcomes: First, participants completed a one-week sun safety diary (T2).We calculated the proportion of sunny workdays on which each protective measure was applied.Second, we compared a control item (frequency of not applying protective measures) with a corresponding protection score.

Statistical analysis
We explored mean differences of the responses at T3 and T4 with 2-sided t-tests if normally distributed, or Mann-Whitney U tests if not.The test-retest reliability was assessed with Pearson's r for normally distributed metric scales, with Spearman's r s for ordinal scales and not normally distributed metric scales and with Cohen's kappa for nominal scales.Test-retest agreement was assessed with Lin's concordance correlation coefficient (CCC), as recommended by Berchtold. 34The CCC is nearly identical to certain intraclass correlation coefficients. 34,35e criterion validity for sun protection behavior was assessed as the correlations with the increase in skin pigmentation and the reported number of sunburns (1-sided tests).The construct validity was assessed as the correlations with the sun safety diary and the control item (1-sided tests).
When a protective behavior score or independent item did not reach the threshold of 0.7 (rounded) for the retest correlation or Lin's CCC or Cohen's kappa, or differed significantly between test and retest, we considered post-hoc analyses to explore potential causes for the deviation and identify possible solutions for the final version of the questionnaire.

Line of work
Gardening/landscape gardening 32 (47)   Construction/mining 18 ( Abbr.: SD, standard deviation The statistical analyses were performed with IBM SPSS 27/28 and MS Excel 2016.Missing values were deleted pairwise.

Online supplementary Figures S1 and S2 show the two final German instruments (OccuSun, SCSK). In Dresden and
Erlangen, 68 OWs aged 20 to 61 years completed the validation study (Table 1).One participant was lost to follow-up after T1.

Test-retest reliability and agreement
The total OccuSun score's test-retest correlation was r = 0.93, but we observed retest discrepancies concerning five region scores and five independent items (Table 2).We performed analyses for individual score items (Online supplementary Table S2) and post-hoc analyses to explore potential causes for deviations (Online Supplementary Material S4).

Criterion validity
At T1, the skin pigmentation in the cheek was between ITA 16 • (tanned) and 60 • (very light); at T3, it ranged from ITA -4 • (brown) to 57 • (very light).The change in pigmentation correlated significantly with the face exposition/protection score at T3 (Table 3), i.e. longer sun exposure and less protection were associated with a larger increase in pigmentation.The correlation with the same score, but gathered at T4, was weak.The number of work-related sunburns (participant-reported for the summer of 2020 retrospectively at T3 and T4, respectively) was not correlated with the whole-body exposition/protection score (Table 3).

DISCUSSION
This study presents reliable tools for assessing sun protection behaviors and knowledge among OWs for research and health education purposes.The OccuSun assesses protective behavior retrospectively for the preceding summer, with better criterion validity for surveys in early autumn than in mid-autumn.Our findings suggest that the OccuSun questionnaire provides a very reliable total score for protective behavior and that the German SCSK total score for knowledge has good reliability.
Both questionnaires can be applied in research to evaluate the effects of skin cancer prevention interventions.In established preventive educational programs, they can be  useful to assess the participants' baseline protective knowledge and behaviors so that the intervention can take into account their specific needs.

Sun protection behavior (OccuSun)
The OccuSun questionnaire includes ten region scores of occupationally sun exposed body regions with their corresponding potential protective measures.While the total score's reliability was excellent (r > 0.9), the reliability of the region scores and independent items was mostly good (r s > 0.7) and sometimes moderate (r s > 0.5) (Table 2).A comparison of these properties with the questionnaire developed by Glanz et al., 26 on which the OccuSun is partly based, is not possible because the former instrument does not address body regions separately, provides no scores, and was not evaluated by a retest.To our knowledge, only one validation study of a German-language questionnaire for sun safety behavior has been published to date, the SEPI 24 (part 1: behavior).The study reports the score's internal consistency (α = 0.70) and correlations

Correlation of number of work-related sunburns with the whole-body exposition/protection score at T4
All with corresponding items from another questionnaire 17,20 (0.48 ≤ r s ≤ 0.87).Moreover, like the OccuSun, the SEPI has acceptable reliability (test-retest stability of the 8 items: Cohen's weighted kappa = 0.52-0.75).Even though it was developed for the general population, was validated in university students, and includes recreational behavior, the SEPI may be preferred for OWs in situations that require a brief assessment tool for individual risk behavior, e.g. during skin cancer screenings.
In the absence of a validated, target group specific standard questionnaire, previous studies of sun protection measures in German-speaking OWs 13,15,16,21,23,36 have applied their own self-developed instruments, mostly without addressing the measurements' validity.Two of these studies 21,23 reported using items from questionnaires that were previously validated (in other languages and populations) but did not provide the final questionnaires nor details concerning their validity or reliability.
In order to keep the OccuSun reasonably short, we did not include specific questions about ears, lips, or feet.While feet are usually not exposed to the sun at work, ears and lips require protective measures, although they were not part of the questionnaire.Our results indicated that some items required clarification (see Online Supplementary Material S4 for details).First, the OccuSun now refers only to clearly defined severe sunburns.Less severe sunburns can also increase the skin cancer risk, but the results of this validation study suggest a recall bias.Severe sunburns are remembered more reliably. 37Second, we initially asked participants if they were allowed to wear sunglasses, or to freely choose which kind of headgear they wore at work.Both questions were deleted because their retest reliability was low and the information is not required by default.When researchers or coaches have a special interest in promoting headgear or sunglasses, they can discuss the applicability of these measures at work during the training sessions or address it as inclusion criterion for trials.Third, we clarified the questions about having lunchbreak in the shade and working in the shade so that they now clearly relate to the OWs' choice.Forth, for small-and wide-brimmed headgear, we added illustrations and the word "circumferential" to clarify that "brim" does not refer to baseball caps.Finally, the item about using sunscreen for the back and front of the neck was split because some participants may have understood "or" instead of "and".We found no plausible explanations for the small, statistically significant retest changes concerning the legs score (p < 0.05), but given the number of region-score mean-difference tests we performed, this may be a random result (Online Supplementary Material S4).
The OccuSun includes ten region scores of varying reliability; therefore, we recommend that future hypothesis testing be restricted to the total score and that subscores be evaluated descriptively.Exceptions may include a small number of pre-specified region scores that are relevant to the respective research question.

Sun protection knowledge (SCSK)
To our knowledge, this is the first questionnaire for sun protection knowledge validated in German OWs.The retest reliability of the total SCSK score was acceptable in our sample (r s = 0.76) but somewhat lower than for the original SCSK 31 (r = 0.83).Since the single item responses were generally not reliable, we strongly recommend against item-specific hypothesis testing.Data concerning the items' reliability was not available in the original publication or the German pilot study. 25,31We found it necessary to deviate from the original in two details, which should be noted when comparing SCSK scores across studies in the future.As a first and content-wise negligible change, the correct answer to item 13 is "true" in order to avoid a double negative in the translation.Second, an expert argued that density was more important for textiles' sun protection than their color.There is evidence that both factors are substantial. 38As a compromise, we left the original wording in item 20, but added to it.Since the SCSK does not include OW specific items, it is suitable for the general population.However, it should be validated in this context as well.
Some previous studies with German-speaking participants already assessed sun safety knowledge.However, they did not report on validity or reliability, 12,13,[17][18][19]36,[39][40][41] addressed validity only by a pre-test with cogni-tive interviews 22 or written feedback, 20 applied open questions 42,43 or were developed for children. 44,45 ThGerman version of the SCSK scale has already been applied in an ongoing study at the Center for Preventive Medicine and Digital Health (CPD), Heidelberg University (Medical Faculty Mannheim), by Sophie Leer and colleagues.This German-wide survey among 1,200 outdoor sports trainers assesses the trainers' knowledge of risks associated with UV radiation, their health literacy concerning skin cancer and their daily preventive measures.Another study at the Department of Dermatology at Heidelberg University Hospital, conducted by Anke Lonsdorf and colleagues, is currently applying the German SCSK to assess sun protection knowledge among the high-risk population of chronically immune-suppressed patients who received an organ transplantation.

Strengths and limitations
Our study has several strengths in terms of the sample.We included a wide age range (20-61 years) and several occupations.Construction workers, the largest group of OWs in Germany with 47%, 1 were only slightly underrepresented in our sample (38%), although in our experience this group is difficult to recruit.Only one participant was lost to follow-up.A strength of the OccuSun is that it builds on previous instruments that we have adapted for OWs.][48][49] Moreover, both drafts underwent a consensus process with an interdisciplinary panel of experts as well as a pre-test with indoor and outdoor workers, thereby ensuring face validity, clarity, and applicability.A particular strength of the study design is the inclusion of an objective measure of sun exposure/protection: the increase in skin pigmentation.
A limitation of the study is the rather small sample size.Another limitation is that we needed to refine some items of the OccuSun post hoc, so that the reliability of the final instrument has not yet been fully assessed.To complete the validation of both questionnaires, a further study is planned to assess their sensitivity to change under real-life conditions in a larger sample of outdoor workers.This will require the inclusion of additional German study sites.

CONCLUSIONS
Our results suggest that the instruments are suitable for measuring total scores of sun protection knowledge (SCSK) and behavior (OccuSun) in German-speaking outdoor workers.Having established face validity, retest reliability and, for the OccuSun, criterion and construct validity, the questionnaires are ready for subsequent validation concerning their sensitivity to change.The uptake of both tools in studies concerning the prevention of skin cancer will be promoted with the assistance of appropriate German associations such as the Skin Cancer Council Germany (Nationale Versorgungskonferenz Hautkrebs; NVKH), the Working Group for Occupational and Environmental Dermatology (Arbeitsgemeinschaft für Berufs-und Umweltdermatologie e.V.; ABD) and the Working Group on Dermatological Prevention (Arbeitsgemeinschaft Dermatologische Prävention e.V.; ADP).

A C K N O W L E D G E M E N T
We want to thank all experts who contributed to the consensus and all outdoor workers who participated in the validation study.We thank Dr. Frederik Haarig for his help in the development of the study protocol.We thank the authors of the SCSK and the sun protection questionnaires for their supportive responses and the opportunity to build upon their work.This validation study was conducted with funds from the Skin Cancer Council Germany (Nationale Versorgungskonferenz Hautkrebs; NVKH), who has also agreed to promote the implementation of both questionnaires, e.g. through their newsletter.
Open access funding enabled and organized by Projekt DEAL.

F
I G U R E 1 Compilation, consensus, and pre-test of the instruments.F I G U R E 2 Design of the validation study.

TA B L E 2
Test-retest changes, reliability, and agreement of OccuSun sun protection scores and independent items.
TA B L E 1 Comparison of OccuSun sun protection scores and items with sun safety diary.