Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 Jul 1.
Published in final edited form as: Am J Prev Med. 2016 Jan 22;51(1):1–11. doi: 10.1016/j.amepre.2015.11.008

Efficacy of an Intervention to Alter Skin Cancer Risk Behaviors in Young Adults

Carolyn J Heckman 1, Susan D Darlow 1, Lee M Ritterband 2,3, Elizabeth A Handorf 4, Sharon L Manne 5
PMCID: PMC4914462  NIHMSID: NIHMS739378  PMID: 26810358

Abstract

Introduction

Skin cancer is the most common cancer, and its incidence is increasing. Young adults expose themselves to large amounts of ultraviolet radiation (UV) and engage in minimal skin protection, which increases their risk. Internet interventions are effective in modifying health behaviors and are highly disseminable. The current study's purpose was to test an Internet intervention to decrease UV exposure and increase skin protection behavior among young adults.

Study design

RCT conducted in 2014, with data analyzed in 2015.

Setting/participants

A national sample of adults aged 18–25 years at moderate to high risk of developing skin cancer by a self-report measure was recruited online.

Intervention

Participants were randomized to one of three intervention conditions: assessment only, the website of a skin cancer organization, or a tailored interactive multimedia internet intervention program (UV4.me) based on the Integrative Model of Behavioral Prediction.

Main outcome measures

Self-reported overall UV exposure and skin protection assessed at 3 and 12 weeks after baseline. Secondary outcomes were self-reported intentional and incidental UV exposure, sunburns, sunscreen use, and skin cancer screening.

Results

For the intervention arm, there were significant decreases in UV exposure and increases in skin protection at both follow-up time points compared with the assessment-only condition (p<0.001). The effect sizes (Cohen's d) comparing the experimental and assessment-only arm for exposure behaviors were 0.41 at 3-week follow-up and 0.43 at 12-week follow-up. The effect sizes for protection behaviors were 0.41 at 3-week follow-up and 0.53 at 12-week follow-up. The control condition was not significantly different from the assessment only condition. All three conditions exhibited decreased exposure and increased protection at both follow-ups (p<0.01), but the effect was much stronger in the intervention group. Secondary outcomes were generally also significantly improved in the intervention condition compared with the other conditions.

Conclusions

This is the first published report describing the results of an RCT of an Internet intervention to modify skin cancer risk behaviors among young adults. The UV4.me intervention significantly improved self-reported skin cancer prevention behaviors. Future research will investigate mechanisms of change and approaches for dissemination.

Introduction

Skin cancer is the most common cancer, with nearly five million diagnoses annually in the U.S., and its incidence has been increasing in recent years.1–5 Invasive melanoma is the second most diagnosed cancer among young adults.6 Contributing to increased skin cancer risk among young adults is the fact that U.S. adolescents have had the lowest skin protection rates of all age groups7 and also engage in increased exposure to ultraviolet radiation (UV) as they move into adulthood.8

For these reasons, it is important to have available interventions that are effective in addressing skin cancer risk behaviors among young adults. Although a few such interventions exist,9 most of these intervention studies have been conducted locally often with college (more often female) students only, have required an in-person component, and the length of follow-up has been short. Additionally, no reports on Internet interventions for skin cancer prevention among this population have been published previously. However, with approximately 97% of U.S. young adults using the Internet,10 and the evidence for the efficacy of Internet interventions,11 the Internet is an ideal mechanism with which to reach young adults and explore the efficacy of a skin cancer prevention intervention for this population.

The web-based intervention that was designed to modify skin cancer risk and protective behaviors among young adults was informed by the Integrative Model for Behavioral Prediction (IM).12 Constructs from the IM are associated with skin cancer risk and protective behaviors and include demographics, past UV-related behavior, attitudes such as appearance consciousness,13,14 other individual difference variables (e.g., knowledge and risk perception), UV-related beliefs,15– 22 norms and compliance,13,23 self-efficacy and control,24–29 and intentions,28 though the authors' group is the only one that has applied the overall model to the skin cancer domain. Individually tailored and interactive messages and material focusing on IM constructs such as norms and self-efficacy were featured in the web-based intervention. The tailored intervention emphasized appearance concerns, which are known to be the primary motivation for UV exposure and lack of skin protection among young adults. This was accomplished in part through the use of facial images showing UV damage as well as computerized age progression demonstrations.30–33

The purpose of the study was to test the efficacy of the web-based intervention to decrease UV exposure and increase skin protection behaviors among young adults at moderate to high risk of developing skin cancer in an RCT. It was hypothesized that participants randomized to the experimental intervention would report significantly less exposure and more protection than participants in other conditions at follow-up. Unlike prior research, this study was conducted with a large national sample.

Methods

Participants were recruited online by Survey Sampling International using their U.S. consumer opinion panel and partnerships with other panels and online communities. Survey Sampling International panelists were exposed to brief web banner ads about the study from which they could click to link to the study website. Once at the study website, interested candidates were asked to complete the Brief Skin Cancer Risk Assessment Tool (BRAT),34 which was scored automatically. Skin cancer risk items include sun sensitivity, sunburn history, number of moles/freckles, and climate of childhood residence. Items are weighted, resulting in a 0–78 score. The BRAT authors recommend a cut off of ≥27 to denote moderate to high skin cancer risk. Internal and test–retest reliability compare favorably to those reported in the literature for similar items/scales (BRAT).34 Eligible participants were adults aged 18–25 years, had never had skin cancer, and were at moderate to high risk of developing skin cancer based on the BRAT. Approximately 47% of the 5,015 individuals who submitted a screening form met these eligibility criteria. Figure 1 shows the CONSORT study flow diagram.

Figure 1.

Figure 1

Open in a new tab

CONSORT flow diagram.

FU, follow-up

This project was approved and monitored by Fox Chase Cancer Center's IRB, and informed consent was obtained from research participants. Eligible participants were directed to the online informed consent form, which participants signed using a computer mouse. Participants were then directed to the baseline survey. Participants were subsequently randomized to one of three treatment conditions in blocks of nine or 12: assessment only, a control website, and the intervention website (described below).

In order to intervene with participants prior to summer, participants completed assessments at baseline in the spring (March to June, 2014), 3 weeks after baseline (April to July, 2014), and 12 weeks after baseline (June to October, 2014). Of the participants who completed baseline assessments, approximately 71% completed the first follow-up survey, and approximately 72% completed the second follow-up. Participants received an Amazon e-giftcard after each questionnaire: $10 for baseline, $20 for 3-week follow-up, $50 for 12-week follow-up, plus $20 for completing all three assessments.

Intervention Conditions

This trial was registered with ClinicalTrials.gov (identifier, NCT02147080). The experimental website (UV4.me) was targeted to young adults, personally tailored, and included interactive, multimedia, and goal-setting components. The control website was the Skin Cancer Foundation (SCF) website (www.skincancer.org). According to the SCF website, the SCF “is the only international organization devoted solely to education, prevention, early detection, and prompt treatment of” skin cancer. The SCF website was chosen because it is a high-quality multimedia website on the topic of skin cancer; however, no prior study has reported on its potential impact on behavior. Major topics include skin cancer information, prevention, true stories, healthy lifestyle (e.g., sports, anti-aging, vitamin D), news, and getting involved, among others. Participants were sent automated e-mail reminders to enter the control website and to begin and engage with UV4.me (e.g., set and work on goals).

The UV4.me Intervention

Based on a synthesis of prior formative individual interviews and focus groups with the target population, the authors' expertise, and the literature,35–37 a multidisciplinary team collaborated to create the design, modules, and other activities to be included in the web program. The intervention was intended to be interactive, tailored, utilize multiple media formats (text, audio/video, images), and maximize participant engagement. Twelve modules were created, with topics determined to be important in terms of risk or protective behaviors:

  1. Why do people tan?

  2. To tan or not to tan?

  3. Indoor tanning.

  4. UV & health.

  5. Skin cancer.

  6. UV & looks.

  7. Skin damage.

  8. Shade.

  9. Clothes.

  10. Sunscreen.

  11. Sunless tanning.

  12. Skin exams.

Constructs from the IM12 were incorporated throughout the modules. For example, the indoor tanning module provided data showing that most young adults do not indoor tan in order to attempt to influence normative beliefs about indoor tanning. In addition, several other more general website sections (e.g., avatar, MyStuff—a printable summary of tailored goals and recommendations) were developed. Each module was expected to take about 10 minutes to review, and an attempt was made to have each module stand alone and be as focused as possible on encouraging behavior change.

Tailoring algorithms were created to direct participants to prioritize certain modules based on their responses to a few questions administered during the baseline questionnaire (e.g., the indoor tanning module was recommended if participants said they tanned indoors). Tailoring also occurred throughout the web program. For example, participants were asked questions (e.g., Who tans in your circle?) and were provided with feedback (You said that no one in your circle tans. You're in good company!). A number of interactive elements were created to increase engagement in the web program and successful behavioral outcomes. For example, at the end of each module was a goal-setting section (Locke and Latham38 provide a review of goal setting theory and research) in which participants could choose to set a pre-specified goal for the next 2 weeks or not (e.g., For the next two weeks, I will not use a tanning bed, booth, or sunlamp.). More detail about the development, content, and pre-testing of the intervention is provided in another paper (Heckman, et al., Internet Interventions, in press).

Measures

Demographic variables included age, sex, race, ethnicity, skin color, family history of skin cancer, education, employment status, ability to live on income, and receipt of public assistance.

The following sun protection behaviors were assessed, using a seven-item scale adapted from Glanz and colleagues42: wearing sunscreen with a sun protection factor (SPF) of ≥15 on the face, wearing sunscreen with an SPF of ≥15 on other parts of the body, wearing a shirt with sleeves that cover the shoulders, wearing long pants, wearing a hat, wearing sunglasses, and staying in the shade. Participants indicated how often they engaged in these behaviors over the past month (1=never, 5=always). This measure was internally consistent (α=0.76).

In addition, participants were asked to indicate how often they engaged in five UV exposure behaviors (wearing clothes that expose the skin to the sun, sunbathing, getting a tan just by being outdoors [i.e., unintentional tanning], tanning indoors, and using products to get a faster or deeper tan) over the past month (1=never; 5=always), using a five-item scale adapted by Ingledew et al.43 This measure had acceptable internal consistency (α=0.74).

For secondary outcomes, individual items were used. Because these data were skewed, responses were dichotomized. For sunburns, participants were asked to indicate how many times in the past month that they had a red or painful sunburn that lasted a day or more, an item adapted from Glanz and colleagues.42 Responses were dichotomized into those who reported being sunburned in the past month versus those who did not. In addition to including indoor tanning in the overall UV exposure scale, indoor tanning was also examined independently. Participants were asked to indicate the number of days in the past month that they used a tanning bed or booth, using an item adapted from Lazovich et al.44 Responses were dichotomized into those who reported tanning indoors in the past month versus those who did not. Participants were also asked to report how many hours per week in the past month they spent in the sun trying to get a tan (i.e., sunbathing) and not trying to get a tan (e.g., working, recreational activities). Responses were categorized: 0 hours, 1–4 hours, and ≥5 hours for intentional sun exposure and 0 hours, 1–4 hours, 5–10 hours, and ≥10 hours for incidental sun exposure.

Participants were asked about the SPF of the sunscreen they used most often in the past month with response options of didn't use sunscreen, <15, 15–30, >30, and I'm not sure/I don't know. Responses were dichotomized into SPF <15 versus SPF ≥15. In terms of skin cancer screening, participants were asked whether they, a partner, or a healthcare provider had checked their whole body for skin cancer using items adapted from Glanz and colleagues.42 Response options were yes and no.

Statistical Analyses

Prior to conducting outcome analyses, quality metrics derived from the baseline questionnaire were evaluated using methods described by Meade and Craig.45 The metrics selected to indicate potential poor quality responses were:

  1. unusually short or long baseline questionnaire completion time;

  2. unusually long or short strings of identical consecutive responses;

  3. low within-subject correlation for items with the strongest positive correlations across all subjects (i.e., synonyms);

  4. unusually high or low within-subject correlations for even and odd numbered items within subscales; and

  5. unusually high or low Euclidean distance from subjects' responses to the mean response on each item.

Responses were also “flagged” as indicating potential poor quality for issues such as providing a non-unique or non-working e-mail address or phone number.46

Based on interactions with participants during the course of the study, several participants with unusually similar responses on the screener questionnaire were identified who may have been either affiliated with one another through a participant-initiated Facebook post or who had enrolled in the study more than once. Hierarchical clustering was used to determine membership in this affiliated group based on the Euclidean distance matrix of responses to each screener question.47,48 A cluster of individuals was identified and are referred to as “similar screeners.”

Polytomous Variable Latent Class Analysis Version 1.42013 was used to identify groups of subjects with similar distributions of quality metrics.49 Included were all quality metrics from the baseline questionnaire, other flagged responses, and classification as a similar screener. Based on characteristics of the fitted class memberships, classes representing lower-quality responders were determined. Overall, low-quality respondents (n=269) had a greater number of “hits” on the low quality metrics.

The final sample (N=965) consisted of participants with predicted membership in the high-quality latent class. Demographics were examined to determine whether randomization was successful using chi-square tests and ANOVA. All outcomes were analyzed using generalized linear regression models. The primary outcomes of UV exposure and skin protection behaviors were analyzed as continuous variables, using linear regression. Additionally, categories for intentional and incidental UV exposure were created based on hours of reported exposure per week, and multinomial logistic regression was used to determine the effects of the intervention. Other secondary outcomes were dichotomized, and logistic regression was used to model these outcomes. All models contained intervention condition, time (categorized as baseline, 3 weeks post baseline, and 12 weeks post baseline), and the interaction between intervention and time. Generalized estimating equations50 were used for all regressions to account for within-individual correlation. For the exposure and protection outcomes at 12-week follow-up, moderation by sex and age were assessed using a linear regression, with an interaction between treatment and each of the moderating variables. Likelihood ratio tests were used to determine whether the overall interaction between each moderator and treatment was statistically significant.

The reported analyses used all observed data from the high-quality sample. To determine potential biases due to selective drop-out, another analysis of the primary outcomes was conducted with the last observation carried forward to fill in missing data. This gives a conservative estimate of the intervention effect, as it assumes no improvement in any participants who did not complete the 3- or 12-week follow-up questionnaire, which can be interpreted as a lower bound on the true intervention effects. To determine the effect of removing low-quality screeners, a sensitivity analysis was conducted in which all participants who completed the baseline questionnaire were included, regardless of quality.

Results

No demographic covariates were significantly different between the three intervention conditions, indicating successful randomization (Table 1). Demographic characteristics of the high-quality sample are shown in Table 1, and descriptive data for the outcomes at each time point are shown in Table 2. On average, the 84% of those randomized to the control website who accessed it visited the control website twice (2.1 [1.7]), whereas the 70.4% of experimental condition participants who accessed the intervention website visited it more than five times (5.8 [5.0]), completing almost half (5.7 [5.0]) of the available 12 modules.

Table 1. Baseline Demographic Characteristics by Intervention Condition (N=965).

Variable n (%) Overall Assessment only (n=340) Control (n=338) Experimental (n=287) p-value
Age [Mean (SD)] 21.8 (2.2) 21.9 (2.2) 21.7 (2.2) 21.8 (2.2) 0.498
Female Sex 637 (66.1) 224 (65.9) 216 (64.1) 197 (68.6) 0.487
White 825 (85.7) 297 (87.6) 283 (84.0) 245 (85.4) 0.397
Hispanic 92 (9.6) 30 (8.9) 31 (9.3) 31 (10.9) 0.668
Fair skin color 833 (86.3) 288 (84.7) 301 (89.1) 244 (85.0) 0.192
Family history of skin cancer 339 (35.2) 110 (32.4) 127 (37.7) 102 (35.5) 0.344
College graduate 213 (22.1) 71 (20.9) 76 (22.6) 66 (23.0) 0.792
Working >full-time 175 (18.1) 62 (18.2) 58 (17.2) 55 (19.2) 0.819
Hard to live on income 314 (32.5) 114 (34.9) 113 (34.3) 87 (31.4) 0.634
Receives public assistance 173 (18.8) 62 (18.6) 62 (19.1) 49 (18.1) 0.953
Northern U.S. 640 (66.3) 227 (66.8) 232 (68.8) 181 (63.1) 0.497
Open in a new tab

Table 2. Descriptive Data for the Outcome Variables.

Variable Baseline
(n=965)		 3-week follow-up
(n=623)		 12-week follow-up
(n=629)a
M (SD) Assessm
ent only
(n=337)b 		Contr
ol
(n=33
6)		 Experime
ntal
(n=286)		 Assessm
ent only
(n=229)		 Contr
ol
(n=20
3)		 Experime
ntal
(n=191)		 Assessm
ent only
(n=229)		 Contr
ol
(n=20
5)		 Experime
ntal
(n=195)
Overall UV exposure 1.49 (0.80) 1.50 (0.79) 1.44 (0.80) 1.34 (0.81) 1.35 (0.75) 1.02 (0.76) 1.21 (0.73) 1.19 (0.70) 0.89 (0.73)
Skin protection 1.95 (0.77) 1.83 (0.81) 1.94 (0.81) 2.09 (0.87) 2.00 (0.84) 2.45 (0.88) 2.17 (0.87) 2.17 (0.84) 2.64 (0.89)
N (%)
Sunburn 191 (56.3) 172 (51.5) 156 (54.5) 116 (50.9) 91 (44.8) 62 (32.6) 94 (41.2) 78 (38.2) 51 (26.3)
Indoor tanning 30 (8.9) 31 (9.3) 26 (9.1) 19 (8.3) 17 (8.3) 8 (4.2) 17 (7.4) 12 (5.9) 8 (4.1)
No incidental UV 84 (25.1) 80 (24.0) 78 (27.6) 38 (16.8) 27 (13.4) 46 (24.6) 38 (16.9) 43 (21.3) 60 (31.1)
No intentional UV 265 (78.2) 254 (75.4) 212 (74.1) 176 (77.2) 153 (75.0) 160 (83.8) 186 (81.6) 163 (80.3) 175 (90.7)
SPF of 15 or higher 187 (55.0) 164 (48.5) 143 (49.8) 150 (65.5) 131 (64.2) 140 (73.3) 161 (70.3) 156 (76.1) 162 (83.1)
Clinician screening 54 (15.9) 61 (18.2) 45 (15.7) 52 (22.7) 41 (20.1) 47 (24.6) 56 (24.5) 53 (25.9) 69 (35.4)
Self-screening 43 (12.6) 48 (14.3) 36 (12.5) 59 (25.8) 38 (18.6) 63 (33.0) 59 (25.8) 48 (23.4) 87 (44.6)
Open in a new tab
a

Participants who did not complete the 3-week follow-up were still permitted to complete the 12-week follow-up.

b

Note that n = the total possible sample size for each intervention group at each time-point, though the sample size for any specific outcome variable may have been slightly less due to missing data.

The main results of interest in the regression models were whether the effects of the intervention and control arms were different from the assessment-only arm at 3- and 12-week follow-up. For the intervention arm, significant decreases in exposure behaviors and significant increases in protection behaviors were found at both 3 and 12 weeks when compared with the assessment-only arm (p<0.001, Table 3). The effect sizes (Cohen's d) comparing the experimental and assessment-only arm for exposure behaviors were 0.41 at 3-week follow-up and 0.43 at 12-week follow-up. The effect sizes for protection behaviors were 0.41 at 3-week follow-up and 0.53 at 12-week follow-up. The control arm was not significantly different from the assessment-only arm at follow-ups. All three arms exhibited decreased exposure and increased protection at 3 and 12 weeks (p<0.01), but the effect was much stronger in the intervention group (results not shown). The intervention effects were somewhat attenuated when using the last observation carried forward method for missing data or when all participants, regardless of quality assessment, were included; however, effect direction and statistical significance was maintained in all analyses (results not shown). Significant moderator effects of sex and age were not identified (results not shown).

Table 3. Generalized Estimating Equation Analyses for Primary Outcomes.

Overall UV exposure (N=962) Skin protection (N=894)
Effecta SE p-value Effecta SE p-value
Baseline
Assessment only ref - - ref - -
Control 0.014 0.061 0.822 -0.127 0.064 0.047
Experimental -0.051 0.064 0.423 -0.017 0.069 0.804
3 weeks
Assessment only ref - - ref - -
Control 0.012 0.069 0.864 -0.108 0.082 0.186
Experimental -0.293 0.071 <0.001 0.349 0.086 <0.001
12 weeks
Assessment only ref - - ref - -
Control -0.034 0.067 0.609 -0.024 0.083 0.773
Experimental -0.300 0.070 <0.001 0.429 0.090 <0.001
Open in a new tab

Note: Boldface indicates statistical significance (p<0.05).

a

Treatment effects at each time point compared with assessment only condition.

For the secondary outcomes, all groups were less likely to report any sunburns (p<0.001) and to report ≥5 hours of intentional sun exposure (p=0.05) at 12-week follow-up. All groups were more likely to report use of SPF of ≥15 (all p<0.001), to report having completed a clinician skin cancer screening (all p<0.01), as well as completing a self or partner-assisted skin cancer screening (all p<0.001) at 3- and 12-week follow-up.

In terms of intervention effects on secondary outcomes, intervention participants were less likely than assessment-only participants to report sunburns at either 3- or 12-week follow-up (p=0.003 and p=0.014, respectively) or 1–4 hours of intentional sun exposure at 12 weeks than assessment-only subjects (p=0.002). Intervention participants were more likely than assessment-only participants to report clinician screening at 12-week follow-up (p=0.035) as well as self-screening (p=0.003). The control group did not differ from the assessment only group on any of these secondary outcomes (results not shown).

Finally, intervention participants were more likely than assessment-only participants to report higher SPF use at either 3- or 12-week follow-up (p=0.015 and p<0.001, respectively), as were control participants at 12-week follow-up (p=0.019). Intervention participants were less likely to report 5–10 or ≥10 hours of incidental sun exposure at 12 weeks than the assessment group (p=0.018 and p=0.016, respectively), but there was no decrease in 1–4 hours of incidental UV exposure (results not shown). There were no significant differences over time or by treatment condition for the number of days spent indoor tanning in the past month.

Discussion

To the authors' knowledge, this is the first published report on the efficacy of a web-based prevention intervention to modify skin cancer risk behaviors among at-risk young adults from a national RCT. Participants in the tailored, multimedia, interactive Internet intervention (UV4.me) improved their self-reported UV exposure and protection behaviors significantly more than they would have otherwise. The effect sizes for the primary outcomes of 0.41–0.53 were in the small to medium range according to Cohen's criteria.51 Though the experimental group decreased their indoor tanning by more than 50%, there was not enough power for this result to be significant due to the small number of individuals who had tanned indoors in the last month at baseline, approximately 9%. These findings are important because UV exposure is a definitive skin cancer risk factor, and skin protection can help prevent skin cancer.3,52-58

Upon closer examination of some of the behaviors, the experimental group reduced both their intentional and incidental exposure more than the other groups. Both the experimental and control group were more likely to report using higher SPF sunscreen than the assessment-only group. This suggests that a generic website not necessarily intended to change specific behaviors may be successful in doing so in some cases. Additionally, the experimental group was more likely to report both clinician and self skin cancer screening than the other groups.

Interestingly, participant behavior also improved in the control condition and the assessment-only condition, though less than in the experimental condition. This is notable given that one might have expected behavior to worsen during the summer. It is unknown whether the improvement reported in the non-experimental groups was due to potential increased awareness, behavior tracking, attrition by less motivated participants, social desirability, or other factors related to longitudinal “panel conditioning.”59,60 Though the assessments were not intended to serve as an intervention, participants from each condition commented that they found the questionnaires “fun,” “thorough,” “interesting,” or “informative.” Some prior skin cancer prevention studies have found knowledge and behavior to be unrelated or demonstrated minimal short-term effects of increased awareness and knowledge.61-65 Thus, it would be interesting to see whether the changes in the non-experimental groups would have decreased over a longer time period. Many studies have demonstrated behavioral tracking to be an effective behavior change intervention, though none have been reported in the area of skin cancer prevention.66-69 Participants who were less interested in or motivated by the topic may not have completed the follow-up questionnaires, resulting in the appearance of a certain amount of behavioral improvement across all groups, although improvements for all groups were maintained based on the last observation carried forward method. Finally, though prior studies have found self-report of UV exposure and protection to be reliable and valid,39–41 future intervention researchers may want to include a measure of social desirability.

Next steps for this line of research will be to investigate additional moderators, mediators, and mechanisms of intervention effects as well as dissemination and enhancement strategies. It will be useful to learn whether certain subgroups of individuals are more or less impacted by the intervention based on demographic characteristics such as skin color, past behavior (e.g., having indoor tanned or not), or traits such as concern about appearance. Though several participants commented that the intervention seemed to be designed primarily for women or adolescents, and women and older participants were more likely to complete the study assessments, intervention effects were not moderated by sex or age. Similarly, it will be important to know which constructs from the IM (e.g., norms, self-efficacy) change with intervention and whether certain modules or components of the intervention are more or less impactful than others. Additionally, a dissemination trial should investigate strategies for engaging at-risk young adults with the intervention without the help of a market research company and unsustainable incentives such as by using Google, Facebook, and snowball sampling. Finally, the intervention could be enhanced to facilitate reach and engagement through the use of a mobile version and social media/networking. Although intervention engagement seemed relatively strong compared with other Internet intervention trials,70-75 16% of control participants and 30% of experimental participants did not access the interventions. Among participants in the experimental condition, only about half of the modules were viewed/completed. Innovative techniques and strategies will need to be added to future versions to address matters of adherence. In addition, a longer follow-up period should be employed to better understand issues of long-term behavior change and maintenance.

Limitations

The strengths of the study include a focus on an important public health issue, an RCT design, a theoretically informed intervention, and participation of a large national sample. The study also has several limitations. First, in order to enhance efficiency and geographic diversity, a commercial research panel was used for recruitment rather than other more-traditional methods. However, two large studies have demonstrated similar demographic characteristics and long-term follow-up rates comparing multinational consumer panels and traditional national surveys.76,77 Participants were disproportionately female. This may be because women were more interested in this study or because they were more likely to be eligible owing to their risk behaviors. Second, data quality procedures are necessary for online studies including incentives.45,46 Thus, not all of the data collected for the current study was deemed high quality and able to be included in the outcome analyses. However, though including only high-quality data attenuated the effects somewhat, this did not change the direction or statistical significance of the effects. Third, the SCF website front page is updated regularly with news stories, thus potentially changing the nature of the control condition during the course of the study. However, the study was completed within a period of about 6 months; thus, it is likely that few major changes were made to the website during this time period that would have significantly altered its impact. Fourth, self-report methods were used for outcome assessments. However, several studies have been conducted demonstrating the reliability and validity of self-report questionnaires of UV exposure and protection compared to observation and objective measures with no systematic bias identified among various populations.39,41,78-80

Conclusions

The UV4.me intervention was found to be successful in increasing skin protection and decreasing UV exposure and sunburns, which are known to be associated with skin cancer and its prevention. Future studies designed to modify behaviors associated with skin cancer development will investigate moderators, mediators, and mechanisms of intervention effects as well as dissemination and enhancement strategies.

Acknowledgments

This trial was registered with ClinicalTrials.gov (identifier, NCT02147080). This work was funded by R01CA154928 (CH), T32CA009035 (SD), and P30CA006927 (Cancer Center Support Grant). The study sponsor had no role in study design; collection, analysis, or interpretation of data; writing the report; or the decision to submit the report for publication. We thank the following for their assistance with this project: Carolyn Caruso, Mary Grove, Peter Braswell, and Gabe Heath from the staff of BeHealth Solutions, Inc. for creating the website and online assessment and data management system, Teja Munshi for her assistance with intervention development and data management, Stephanie Raivitch for her assistance with intervention and assessment development, and Helene Conway for her assistance with manuscript preparation.

Footnotes

Dr. Ritterband is an equity holder of BeHealth Solutions, Inc., which developed the data management system and helped develop the intervention described in this paper. Dr. Ritterband's conflict of interest is being managed by a conflict of interest committee at the University of Virginia, in accordance with their respective conflict of interest policies.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

RESOURCES