We included a total sample of 205 German secondary school students of both sexes in the age group of 13-19 years in our cross-sectional pilot study (median 15 years, SD 1.36; 111/205, 54.1% male; 93/205, 45.9% female) attending the two most common school types in Germany (grammar school: n=136, 66.4%; general comprehensive school: n=69, 33.6%). Almost all participants (201/205, 98.5%) owned a smartphone.

From a risk profile standpoint, 50.7% (104/205) of the participants had a Fitzpatrick skin type of 1 or 2 [41]; current sunbed use was reported by 5.9% (12/205) [42], and 77.0% (157/205) remembered having at least one sunburn in the past [43].

The mirroring approach was implemented by medical students from the Education Against Tobacco nonprofit organization attending the University of Essen in Germany [31,44]. To increase familiarity with the photoaging app (called Sunface) and students’ participation in the mirroring intervention, students were asked to download the app before our visit, via a letter 1 week in advance. By this means, 36.3% (74/205) had the app on their mobile phones when we visited the schools.

In the first 15-minute phase, we used the displayed face of one student volunteer to show the app’s altering features to the peer group, providing an incentive for the rest of the class to test the app. In front of their peers and teachers, students could interact with their own animated face via touch (coughing, sneezing, etc) and display their future self based on their skin type (Figure 1) or use of sun protection (Figure 2) or tanning beds (Figure 3) at 5, 10, 15, 20, or 25 years in the future. Multiple device displays can be projected simultaneously, which we used to consolidate the altering measures with graphics (eg, to explain skin atrophy and solar elastosis). We implemented mirroring with a Galaxy Tab A tablet computer (Samsung, Seoul, South Korea) via Apple’s AirPlay interface (Apple Inc) using the app Mirroring360 (Splashtop Inc) for the Android operating system (Google Inc).

In the second 15-minute phase, we encouraged students to try the app on their own device or one of the tablet computers provided for students who did not own a smartphone or did not download the app. Both the provided tablet computers and the students’ own smartphones were connected to the projector.

We gathered the students’ sociodemographic data (sex, age, school type) and their risk profile (skin type, sex, age, sunburn in the past, sunbed use) directly after the intervention via an anonymous survey. We captured their reactions toward the intervention via 6 items on a 5-point Likert scale (ranging from 1=strongly agree to 5=strongly disagree): (1) change of intentions (2 items: indoor vs outdoor tanning); and (2) perceived reactions of the peer group on change in attractiveness (2 items: indoor vs outdoor tanning), whether they perceived the intervention as fun (1 item), and perceived effects of the app as realistic (1 item). The items have been previously used in other published studies [30,45] and were pretested in advance in accordance with the guidelines for good epidemiologic practice [46].

The study received ethics approval from the ethics committee at the University of Essen (17-7587-BO).

In our sample, we measured overall agreement with the subjective realism of the created selfies (72/202, 35.6% strongly agreed or agreed on realism, while n=44, 21.8% disagreed or strongly disagreed, and 42.6% (n=86) were not sure; Figure 4). These results varied notably in males (n=49, 44.5% agreement; n=21, 19.1% disagreement) versus females (n=23, 25.3% agreement; n=23, 25.3% disagreement; Figure 6) but not in other subgroups.

We measured more than 60% agreement in both items that measured motivation to reduce UV exposure and only 12.5% disagreement (n=126, 63.0% agreed or strongly agreed that their 3D selfie motivated them to avoid the tanning bed and n=124, 61.7% to increase use of sun protection). Only 25 (12.5%) disagreed or strongly disagreed with this item and with increased use of sun protection in our sample. The perceived effect on motivation was larger in participants with Fitzpatrick skin types 1-2 in both tanning bed avoidance (n=74, 71.8% agreement in skin types 1-2 vs n=50, 53.8% agreement in skin types 3-6) and increased use of sun protection (n=70, 68.0% agreement in skin types 1-2 vs n=52, 55.3% agreement in skin types 3-6; Figure 5), and also positively correlated with higher age (Figure 7, Multimedia Appendix 2).

The 2 items measuring the reactions of the peer group toward the individual selfie showed positive peer pressure in regard to both use of sun protection (n=117, 57.9%) and tanning bed avoidance (n=142, 72.1%; Figure 4). The subjective norm on decreasing UV exposure in order to look more attractive was perceived by a higher percentage of participants with Fitzpatrick skin types 1-2 than types 3-6, especially for tanning beds (Figure 5, Multimedia Appendix 2).

A majority of participants claimed that they perceived the intervention as fun (n=103, 51.2% agreement vs n=14, 7.0% disagreement), and the fraction agreeing that the intervention was fun was at least two times larger than the fraction disagreeing throughout all subgroups. However, this perception was notably more prevalent in male participants (n=64, 59.3%) than in female participants (n=39, 42.4%; Figure 6). A total of 79 (39.1%) of participants reported that they would try the app again later on and 57 (28.4%) planned to show the app to another person after school (Multimedia Appendix 2).

Available data on appearance-based behavioral change settings reveal that photoaging interventions appear to be more effective for girls [31]. In our sample, the female participants perceived the intervention as less realistic in comparison with males (n=23, 25.3% agreement vs n=23, 25.3% disagreement for females) versus males (n=49, 44.5% agreement vs n=21, 19.1% disagreement; Figure 6). However, it is notable that the relatively low perception of realism by both sexes was not accompanied by low scores in the global feedback, motivation to change behavior, or subjective norm categories. Even though the questionnaires were anonymous, which reduces confounding effects in surveys, we cannot entirely rule out that the self-reported data may have been influenced by a social desirability bias.

Previous publications stressed the relevance of targeting especially fair-skinned individuals with Fitzpatrick skin type of 1 or 2, as these are less protected against UV radiation and thus have a significantly greater risk for skin cancer [1]. In our sample, the perceived effect on motivation was larger in participants with Fitzpatrick skin types 1-2 in both tanning bed avoidance (n=74, 71.8% agreement in skin types 1-2 vs n=50, 53.8% agreement in skin types 3-6) and increased use of sun protection (n=70, 68.0% agreement in skin types 1-2 vs n=52, 55.3% agreement in skin types 3-6; Figure 5, Multimedia Appendix 2), which could be explained by the Sunface app having stronger altering effects in lighter skin than in darker skin types.

Our data also indicate that photoaging interventions have a higher impact in late adolescence (17-19 years) than in early adolescence (13-16 years), but both age groups appear to benefit (Multimedia Appendix 2). We hypothesize that this effect is due to the higher relevance of wrinkle formation (ie, due to solar elastosis) in late adolescence than in early adolescence.

We present a novel way of integrating photoaging in school-based melanoma prevention, which affects the students’ peer group, considers the predictors of UV exposure in accordance with the theory of planned behavior, and is particularly effective in changing behavioral predictors in fair-skinned adolescents (Fitzpatrick skin types 1 and 2). Further research is necessary to evaluate the intervention’s prospective effects on adolescents of various cultural backgrounds.