In this prospective cohort study, we recruited 322 adults (≥18 years of age) from 9 community events in Denver, Colorado from May 2018 through March 2019, primarily in the winter and summer seasons. Participants were recruited from diverse community and health promotion events ranging from cancer benefits, campus wellness fairs, to consulate events. Attendees of the event were introduced to the UV photoaged facial imager, given a brief description of the study, and offered an opportunity to participate. Informed consent was obtained by all participants, and the study was approved by the Colorado Multiple Institutional Review Board.

Data were collected at two time points: (1) baseline, immediately before the photoaged image, and (2) follow-up, immediately after the photoaged image, typically within 30 minutes of each other. All participants completed the baseline questionnaire that assessed demographic information, sun exposure history (both in childhood and in the past year), sun protection behaviors, perceptions of tanning, and perceived risk of skin cancer.

After the baseline questionnaire was completed, participants had a UV photoaged facial image taken and shown to them (Figure 1). The investigators consisted of a medical resident, medical students, and research coordinators who interpreted the images, answered participant questions, and provided sun protection education. Participants were then asked to complete a postimaging follow-up questionnaire that assessed perceptions of tanning, perceived risk of skin cancer, and intent-to-change sun protection behaviors after seeing the UV photoaged facial image.

The Canfield Reveal Imager (UV photoaged facial imager) is a camera that uses standard white light and cross-polarized flash photography to record surface and subsurface skin conditions, capturing two images in quick succession. The crossed-polarizing filter reduces skin surface reflections and allows visualization of skin changes and damages (eg, brown pigmentation, wrinkles, and lines) and provides immediate visual feedback to the individual, demonstrating the harm caused by chronic sun exposure. The UV photoaged facial image can be used as a form of fear appeals educational intervention for skin cancer. The two images are juxtaposed on the screen for visualization and the education of participants (Figure 1).

Study team members were trained on by the principal investigator on interpreting the UV photos and delivering education to the participants. Participants were shown reference photos from the UV Reveal Camera of individuals with varying levels of sun damage visualized prior to seeing their own UV photos. This prior knowledge provided context for participants to self-assess the amount of sun damage they accumulated relative to standard controls. Verbal feedback was provided by the medical team by pointing out specific areas of sun damage (brown spots) visualized on the UV Reveal Camera. Any further questions were answered.

Sun protection behaviors, perceptions of tanning, and risk perception of skin cancer was based on a subset of items from the Sun Protection Awareness Questionnaire. The following are the measured items:

Both the baseline and follow-up questionnaires were completed in-person on a paper. The primary outcome of this study was intent-to-change sun protection behaviors immediately after seeing the UV photoaged image. This variable was originally collected with 3 levels (yes/no/unsure) but was dichotomized as yes versus no/unknown. The primary independent variable was ethnicity (Hispanic vs NHW). Secondary outcomes included change of pre- to postimage perceptions of tanning and risk perception of skin cancer.

The analysis included participants that identified as Hispanic or NHW (n=278), which comprised 86% of the total sample (N=322). We excluded other races because our primary research question focused on Hispanic individual’s sun protection behaviors compared to NHW individuals, and the sample size was small for other races. We excluded records with a missing pre- or postimage date. Missing data analysis using analysis of variance for continuous variables and Pearson chi-square tests for categorical variables were conducted on demographic variables to determine if any differences existed by ethnicity.

Descriptive statistics were performed on baseline demographic information and sun exposure information. The proportion of individuals who intended to change sun protection behaviors was compared using Pearson chi-square test. We compared the coefficient on the predictor from a mixed model with no fixed effect covariates to that from a model with a single covariate. If the coefficients differed by >10%, the covariate was included in the full multivariate analysis. Mixed effects logistic regression was used to assess the relationship between ethnicity and intent-to-change sun protection behaviors. Ethnicity was included in the model as the main effect, and education and age were included in the model as covariates. Location of community event that the interview took place was dichotomized (health event vs not health event) and included in the model as a random intercept. In a post hoc analysis, we included the final model stratified by season (winter vs summer) to investigate if this relationship varied by the season in which the data were collected.

Changes in tanning and skin cancer perception from pre- to postimage were compared using McNemar test and Wilcoxon signed-rank tests for ordinal, repeated data. Variables that changed the coefficient more than 10% were included in the final multivariable model. A mixed effects linear model was used to assess main effects of the intervention, estimating mean change in perception of tanning and risk of skin cancer from pre- to postimage. Ethnicity, education, and age were included in the model as fixed effects. An alpha criterion of P<.05 was used. All tests were 2-tailed. Statistical analyses were performed using Stata version 15 (StataCorp). This study was approved by the Colorado Multiple Institutional Review Board.

We recruited 278 Hispanic and NHW participants from 9 community events. Of the 278 participants, 262 (94%) completed the follow-up questionnaire. Comparisons of baseline information by ethnicity are described in Table 1. At baseline, compared to NHW participants, Hispanic participants were younger, less educated, more likely to work outdoors, had fewer self-reported past diagnoses of skin cancer, and were less likely to use sunscreen in the past 12 months. Furthermore, Hispanic participants’ perceived risk of developing skin cancer was lower; they were more likely to think a tan made them look attractive and were more likely to want a tan. By contrast, NHW participants were more likely to think they needed to cut down on tanning and felt guilty about tanning. No differences by ethnicity were observed by sex, perceptions on developing wrinkles, skin damage, skin cancer from sun exposure, importance of a tan, or use of tanning beds in the last 12 months. Missingness analyses found there were no significant differences between those who did not complete the pre- and postimage surveys by ethnicity, age, sex, or intent-to-change sun protection behaviors.

At follow-up, 213 of 262 (81%) participants reported that viewing the UV photoaged image influenced an intent-to-change sun protection behaviors. However, this differed by ethnicity. Of 141 Hispanic participants, 125 (89%) reported a likelihood of change compared to 88 of 121 (73%) NHW participants (odds ratio 2.9, 95% CI 1.5-5.6). Demographic and clinical information, sun protection behaviors, perceived risk of skin cancer, and perceptions of tanning were not associated with intent-to-change sun protection behaviors. However, these were included in the multivariable model as covariates based on clinical importance. After adjusting for age, education (high school graduate or less vs some college or more), perceiving a tan was more attractive, tanning bed use, and normal sunscreen use, Hispanic participants were significantly more likely to have an intent-to-change sun protection behaviors compared to NHW participants (adjusted odds ratio [aOR] 4.04, 95% CI 1.6-10.4; Table 2). In post hoc analysis, Hispanic participants were more likely to have an intent-to-change sun protection behaviors compared to NHW participants in both summer (aOR 3.28, 95% CI 0.5-25.3) and winter seasons (aOR 4.54, 95% CI 1.1-18.2), although not significantly in summer due to reduced sample sizes.

The most common sun protection behavior changes that participants intended to implement were increases in sunscreen use (134/262, 51%), to start wearing protective clothing like hats (39/262, 15%), and reapplication of sunscreen (26/262, 10%). These sun protection behaviors also varied by ethnicity. Of the 141 Hispanic participants, 96 (68%) reported that they intended to increase sunscreen use, compared to only 41 of 121 (34%) NHW participants (P<.001). More Hispanic participants also reported the intention to reapply sunscreen more often, while NHW participants were more likely to report the intent-to-increase wearing protective clothing like hats, but neither of these differences were statistically significant.

Hispanic participants that intended to change their sun protection behaviors after viewing the UV photoaged image (125/141) trended toward being younger (P=.09), working indoors (P=.13), and having a high school degree or less (P=.07) than Hispanic participants that did not intend to change their sun protection behaviors. Of 122 Hispanic participants, 110 (90%) with low perceived risk of skin cancer at baseline intended to change their sun protection after seeing the UV photoaged image behaviors, compared to only 15 of 19 (79%) with high-perceived risk (P=.11). Sex, previous skin cancer diagnosis, sunscreen use, perceived risk of wrinkles, and perceptions of tanning did not differ by intent-to-change sun protection behaviors among Hispanic participants.