Digital Educational Strategies to Teach Skin Self-examination to Individuals at Risk for Skin Cancer

Abstract

Skin cancer is the most common cancer in the United States, and early detection of melanoma may lead to diagnosis of thinner and more treatable cancers, resulting in improved survival rates. This study examined the effects of message interactivity (high vs. low) and imagery (cartoon, real human character, or customized imagery preference) on accuracy of identifying abnormal skin lesions (ASL)and skin self-examination (SSE) intention. This study employed a 3 (cartoon character vs. real person vs. customization) x 2 (high interactivity vs. low interactivity) between-subjects online experimental design. Participants at risk for skin cancer were randomly assigned to one of the six conditions and completed a survey after reviewing the educational materials. Univariate analyses were conducted to detect group differences on the accuracy of identifying ASL and intention to conduct SSE in the next 3 months. Among 321 participants who completed the study, the mean age was 36.61 years, 56.7% were females, 76.1% had a college or higher degree, and over60% self-identified as non-Hispanic White. Individuals in the high interactivity and customization group (compared to the low interactivity and cartoon group) were more likely to accurately identify ASL. Individuals in the high interactivity and customization or low interactivity and real person imagery groups (compared to the low interactivity and cartoon group) reported higher intention to conduct SSE in the next 3 months. These results suggest that customization and interactivity may be beneficial for educational programs or intervention design to improve both melanoma identification and SSE intention.

Introduction

Skin cancer is the most common cancer in the United States (U.S.), and more people are diagnosed with skin cancer each year in the U.S. than all other cancers combined (Gershenwald & Guy, 2016). Melanoma is a growing public health concern in the United States. Melanoma has the highest fatality rate of all skin cancers (Gershenwald & Guy, 2016; American Cancer Society, 2008, 2018). On average, one individual dies of melanoma every hour in the U.S. (American Cancer Society, 2008, 2018). The American Academy of Dermatology has reported that the cost burden of skin cancer in the United States is heavy. From 2002 to 2011, the annual treatment of skin cancer increased from 3.4 million to 4.9 million (US dollars) and the average annual cost increased from 3.6 billion to 8.1 billion (US dollars) (Guy et al., 2015). The annual cost of treating new melanoma patients is projected to increase from $457 million in 2011 to $1.6 billion (US dollars) by 2030 (Guy et al., 2015).

The incidence of melanoma has increased over the past several decades (e.g., 2.8% increase per year since 1992). The incidence of melanoma is highest among non-Hispanic white individuals and those with a higher socioeconomic status (Little & Eide, 2012 Russak & Rigel, 2012 Singh et al., 2011). Risk factors for melanoma also include the presence of many moles, fair skin, tendency to sunburn, and individuals with a personal or family history of melanoma are at high risk for melanoma (English & Armstrong, 1988; Glanz et al., 2003; Hanrahan et al., 1998 Kelly, 1998).

Early detection of melanoma is vital to improve cancer out-comes and may lead to diagnosis of thinner and more treatable cancers, resulting in improved survival rates (Leiter et al., 2010). Melanomas are often visible on the skin’s surface, and more than half of melanomas are self-detected (Francken et al., 2007; Moore et al., 2008). Skin self-examination (SSE) is a technique for promoting early self-detection of melanoma. Thorough SSE involves deliberate, systematic inspection of all areas of the body to identify suspicious lesions using established criteria, often using a mirror or the assistance of another person to examine hard to view areas. There is evidence that SSE can lead to better outcomes related to melanoma. Berwick and colleagues found that the survival rate was higher among melanoma patients who had performed SSE than those who had not, and SSE could reduce the risk of advanced disease among melanoma patients and may decrease melanoma mortality by 63% (Berwick et al., 1996). Among melanoma patients, the routine performance of SSE was associated with detection of thinner melanomas at diagnosis (Carli, 2003 Pollitt et al., 2009). Early detection by SSE in high-risk individuals has the potential to reduce melanoma mortality and morbidity and thus relieve the burden of melanoma on both the patients and the health care system in the United States.

Most studies on skin cancer prevention have focused on either evaluating the efficacy of the SSE intervention program as a whole or have not tested any advanced features of digital media instead of identifying specific effective components (see reviews Ersser et al., 2019; Yagerman & Marghoob, 2013); moreover, studies examining specific components of skin cancer-related educational programs often investigate sun protection behaviors instead of SSE related outcomes (Bernhardt, 2001; Ersser et al., 2019; Niu et al., 2022; Niu et al., 2021). Therefore, there is a lack of research focused on key learning elements of SSE programs. As a result, little is known about the most effective ways to design educational skill-building components to ensure that individuals are able to accurately identify features of suspicious skin lesions during SSE that are consistent with melanoma. The goal of the proposed project is to investigate the effectiveness of various educational approaches in teaching the skills necessary to correctly identify abnormal skin lesions that have features of melanoma.

Health education about SSE

Health education programs to promote skin self-examination have demonstrated promising results. The documented effects of various educational methods of melanoma early detection (Yagerman & Marghoob, 2013) indicate that education can increase knowledge, awareness, self-efficacy, and compliance regarding SSE, and individuals can be motivated to do SSE (Phelan et al., 2003). For example, Robinson and colleagues used face-to-face sessions to train melanoma patients and their partners to learn about SSE and identify abnormal lesions. Their studies found that the intervention led to better SSE performance as well as increased self-confidence/efficacy (Robinson et al., 2007).

Despite the value provided by programs that can effectively motivate at-risk individuals to perform SSE, there remains a lack of research to examine how to best present the educational SSE materials and train at-risk individuals to learn SSE and improve the accuracy of identifying suspicious moles. The typical approach for training involves teaching participants the “ABCDE” approach, which is a decision rule designed to distinguish the features of melanoma from benign skin lesions. The ABCDE abbreviation corresponds to assessing lesions for asymmetry, border irregularity, color variation, diameter greater than 6 mm and evolution over time (Friedman et al., 1995; Robinson & Ortiz, 2009). Several studies have demonstrated the efficacy of educational programs in instructing and motivating individuals to conduct SSE to check their skin for early signs of skin cancer (Gaudy-Marqueste et al., 2011; Hamidi et al., 2008; Yagerman & Marghoob, 2013).

Previous studies have presented the ABCDE rules in a variety of approaches and examined the effects of imagery materials on skin cancer-related outcomes compared to text/verbal presentation; however, little research has investigated the best way to present these results or reinforce learning (Hamidi et al., 2008; Yagerman & Marghoob, 2013). Therefore, more studies are needed to investigate the effects of educational methods that aim to promote better learning of SSE among at-risk individuals for melanoma. The findings can also provide valuable information to guide the design and development of comprehensive interventions that aim to promote sun protection and skin surveillance behaviors that are related to skin cancer.

Theoretical background and health education

Visual communication theories and Interactivity Theory were used to guide the design of the proposed experiment. Visual images can draw extra attention, improve comprehension, help viewers recall health information and subsequently lead to health behavior changes (McWhirter & Hoffman-Goetz, 2013). Understanding visual attention, which is how individuals allocate attention to visual stimuli in informational messages, is critical in health communication. Visual attention allocation indicates users’ interest, cognitive processing, decision-making, perceptions and behaviors, etc. (Calitri et al., 2009; Kim et al., 2022; Sutton & Fischer, 2021). Theories from visual communication have the potential to guide the development of health interventions but have not been fully applied to health campaigns related to skin cancer (McWhirter & Hoffman-Goetz, 2013; McWhirter & Hoffman-Goetz, 2014). Some studies mentioned using visual communication elements in guiding their education programs related to promoting skin protection and avoiding tanning (Girardi et al., 2006; Robinson & Ortiz, 2009; Gaudy-Marqueste et al., 2011; Emmons et al., 2011). The Pictorial Superiority Effect and Dual Coding Theory (Paivio & Csapo, 1973; Paivio, 1991) suggest that visual images can reinforce memory recall and enhance cognitive processing during learning. Thus, this project used visual image theories (i.e., Pictorial Superiority Effect and Dual Coding Theory) to guide the design of the SSE educational experiment.

Different types of characters such as cartoons and real human images in messages have been used to promote health information in a variety of health domains (e.g., healthy eating, smoking cessation, and sexual health). For example, cartoon characters have been used in numerous health messages designed for children (Kraak & Story, 2015) as well as for adults (Sheer et al., 2018; Hendriks & Janssen, 2018; Mukherjee & Dubé, 2012; Ahrens et al., 2006).A number of studies suggest that graphic, vivid, and/or relatable health messages may increase psychological reactance to desired health behaviors (LaVoie et al., 2017; Quick & Stephenson, 2008; Wolburg, 2006). Some people may have psychological reactance to real medical images of cancer or other illness which could trigger fear (Shen, 2011).The information expressed in cartoons tend to be vivid and easy to remember, which may promote attitudinal and behavioral changes (Weinert, 2013). Jee and Anggoro (2012) illustrate that cartoons integrating words and images facilitate learning. Therefore, it is important to examine whether we should use cartoons or real images as educational materials for cancer education especially skin cancer self-detection, which requires significant review of cancer images.

From a communication perspective, the Modality, Agency, Interactivity and Navigability (MAIN) model posits that interaction between humans and a system or message could influence information processing (Sundar, 2008). Interactivity is defined as how responsive a system is to a user (Rafaeli, 1988) and has been tested by health communication researchers as a key component of interactive digital communication (Atkinson, 2004; Lu et al., 2014; Lustria, 2007; Oh & Sundar, 2015). Interactivity plays an important role in health communication due to its influences on attitudinal and behavioral changes (Lu et al., 2014; Lustria, 2007; Oh & Sundar, 2015; Niu et al., 2019; Niu et al., 2021). Previous work has found support that interactivity that allows users to alter the source, medium, and message increased the sense of interaction (mutual communications) between the message and users(Rafaeli, 1988; Liu & Shrum, 2002). Customization refers to the ability to self-tailor the mediated environment and allows users to serve as the communication source (Kalyanaraman, & Sundar, 2006; Sundar, Bellur, & Jia, 2012; Sundar & Nass, 2001). For example, changing features or choosing certain content to match one’s individual preferences is part of customization, which empowers the users to have the sense of controlling the communication source or what they see. This process could increase user engagement with and depth of processing of the medium or message and may subsequently improve health-related outcomes (Lu et al., 2014; Lustria, 2007; Oh & Sundar, 2015).

Although some studies have found that higher levels of modality and message interactivity can improve knowledge (Lu et al., 2014), attitude toward health websites (Lustria, 2007; Oh & Sundar, 2015), and intentions to use a health information resource (Oh & Sundar, 2016; Willoughby & L’Engle, 2016), some studies have found that greater interactivity did not influence knowledge (JAFFE, 1997; Camerini & Schulz, 2012) or self-efficacy (JAFFE, 1997) positively. Overloading of interactivity with technological features on media interfaces that exceed cognitive processing may lead to a distraction and reactance toward the persuasive messages (Bucy, 2004; Sundar & Marathe, 2010). Since there are both arguments for using interactive technology and against it in terms of information processing, it is important to further examine the role of interactive features in an online environment.

Users on different digital media platforms access information using various technological features. Moreover, the variety of content has also been designed in different ways (e.g., video, interactive media, augmented reality) to influence users’ perceptions and behaviors. It is important to consider the joint impact of digital features and content elements when evaluating the effects of digital educational programs on users. Interactivity and visual elements can both impact users’ learning process and their attitudes and behaviors. How interactivity and visual factors interact in real world situation warrants further empirical investigation. Thus, the combined effects produced by different factors together are also significant to evaluate. We conducted a pilot study to evaluate the effects of interactivity and types of imagery on SSE learning and intention. In addition, we aimed to assess the attention allocation and the acceptability and interest of this program to obtain preliminary data for a future skin self-examination education. We proposed the following research questions:

RQ1: How will (a) interactivity level and (b) types of imagery (i.e., cartoon characters, real human characters, and customization) influence accuracy of identifying abnormal skin lesions and SSE intention?

RQ2: Is there an interaction effect of interactivity and imagery type on accuracy of identifying abnormal skin lesion and SSE intention?

RQ3: What are the differences of outcomes between the six groups?

RQ4: Which part of the health messages will be paid attention to first?

RQ5: What is the acceptability of and interest in the skin cancer educational material?

Method

Experiment and Survey Design

To test the effects of interactivity and character imagery on teaching SSE for identification of abnormal skin lesions and SSE intention, a2 (interactivity: high vs. low)× 3 (characters in imagery: cartoon vs. real person vs. customized imagery preference) between-subject experimental design was conducted online via Qualtrics. Participants at risk for skin cancer were randomly assigned into one of six conditions and completed a survey after viewing the educational materials.

Interactivity was provided by both using the “Hot Spot” feature of Qualtrics and asking for qualitative input from the participants about the educational materials they viewed (See Figure 1 for examples). Hot Spot questions allow participants to have interactions with the content by clicking on any part of the graphics presented. Quantitative ratings and qualitative input can increase deeper processing of the content and can be used to improve educational material design for future campaigns and/or interventions. Participants were also randomly assigned into a cartoon character, a real human doctor character, and a customized condition where the participants could choose from which of these avatars they would like to learn (see Figure 1 for examples).

Figure 1.

Examples of experimental groups

Study Procedure

Amazon Mechanical Turk (MTurk, Amazon’s online marketplace for human intelligence tasks) was used to recruit participants. A “work assignment” in MTurk was created, and individuals with MTurk accounts were able to view the information about the study on a task page. After reading the brief description of the study, interested participants completed a brief online screening assessment. Eligible individuals were aged 18 years or older and met one or more of the following skin cancer risk criteria: having more than 2 moles on the body that are larger than a pencil eraser (1/4 inch), reporting very fair or fair skin type, or having any close blood relative (parent, child, brother, sister) with a melanoma diagnosis. Eligible participants were invited to participate in the full study conducted using Qualtrics, a secure online data collection program provided to researchers by the University.

Interested participants were directed to the study consent form and, if they gave their consent, were assigned into one of the six SSE training conditions. After reviewing the educational materials in the condition, participants were asked to complete a survey to examine their knowledge and other variables related to the experimental condition. After completing the Qualtrics survey, participants were presented with a completion code to get an incentive directly from Amazon. This allowed the researchers to provide the incentive to participants without directly obtaining their contact information and thus retaining the anonymous nature of the data collection. The entire study took approximately 20–30 minutes for the participants to complete. Each participant was paid $3.50 for their participation in the study. This study was approved by the university institutional review board.

Sample

334 participants started the baseline survey and 321 participants completed the baseline survey. Participants who had not finished the experiments or did not complete any of the dependent variable assessments were not included in the final analysis. 56.7% of the sample were female, and 76.1% of the participants reported having a college or higher degree. More than 60% (n=194) identified themselves as non-Hispanic White. Participants ranged from 18 to 76 years old (M = 36.61, SD = 10.57). The summary statistics of participant demographics and outcome variables are reported in Table 1.

Table 1.

Descriptives of Main Variables under Study

Variable	na	%	Mean (SD)	Range
Accuracy of identifying abnormal skin lesions	315		4.05 (1.72)	0-7
SSE intention	315		3.10 (1.18)	1-5
Sex
Male	138	43.0
Female	181	56.0
Age	315		36.61 (10.57)	18-76
Education
High school or less	31	9.7
College degree or some college	223	69.5
Graduate degree or above	64	19.9
Marital Status
Married/partnered	333	50.1
Not married/partnered	332	49.9
Race/ethnicity
Non-Hispanic whites	193	60.1
Other	124	38.6
Melanoma risk factors			1.76 (1.09)	0-4

^a.Variations in the sample size are due to missing data.

Measures

Dependent variables.

Seven items were used to assess accuracy of identifying abnormal skin lesions (knowledge). Each item had an image of a mole presented and asked the participants whether they thought this is a potential melanoma. For each item, the response options were yes, no, and don’t know. Five out of seven images depicted abnormal lesions. A total score (ranging from 0 to 7) was calculated for each participant by summing the number of correct answers (with a don’t know response being coded as incorrect) (M = 4.05, SD = 1.72).

SSE intention was measured by one question asking the participants to answer “How likely do you think it is that you will check any part of your body for early signs of skin cancer in the next 3 months?” on a 5-point response scale (from 1= very unlikely to 5 = very likely) (M = 3.1, SD = 1.18) (Friedman et al., 1995; Janda et al., 2004).

Covariates.

Age, education, sex, race/ethnicity, marital status, and melanoma risk factors were controlled for in the analyses. Melanoma risk factors were measured with four questions. Based on prior skin cancer research (Glanz et al., 2003;Coups et al., 2013; Coups et al., 2012; LeBlanc et al., 2008; Bränström et al., 2010), we categorized individuals as having a melanoma risk factor if they reported “yes” to any of the screening questions: fair or very fair skin color; having 2 or more moles on the body that are larger than a pencil eraser (1/4 inch or 6 millimeters); having any close blood relative (parent, child, brother, sister) with a melanoma diagnosis; or severe or moderate sunburn if stayed outdoors in the midday summer sun for one hour without sun protection. Answering “yes” to any of the above questions was counted as 1 for each factor, and we calculated the total number of melanoma risk factors for each participant. The sum score was used to measure skin cancer risk factors ranging from 0 to 4 (0=no skin cancer risk, 4=greatest skin cancer risk).

Acceptability.

The current study also had questions regarding the acceptability of the educational program. One question asked how interested the participants would be in taking part in a skin cancer-related behavioral intervention that teaches you skin self-examination and sun protection behaviors (1 = Not at all interested, 7 = Very interested) (Lindsay et al., 2020). Four items asked the participants how well the adjectives “interesting”, “relatable”, “insightful”, “informative” described the skin self-examination information they read (1 = Not at all, 10 = Extremely) (Crawley et al., 2018). One set of ten questions asked whether the participants would be interested in such a program if it involved different mobile or web-based platforms (Table 2).

Table 2.

Interests in educational program via different platforms

Channels	na	Yes %	No %	Maybe %
Using an app on a mobile device	315	65.9	22.3	11.8
Receiving information from your healthcare providers?	314	62.4	22.6	15.0
Reading information on a computer-based website?	314	62.1	19.1	18.8
Reading information on a mobile-based website?	314	61.8	22.9	15.3
Receiving information by email	314	59.2	21.3	19.4
Receiving information by text messages?	314	53.2	34.4	12.4
Using WhatsApp?	314	51.9	39.2	8.9
Reading information on Instagram?	314	43.3	38.2	18.5
Using Facebook groups or public accounts?	314	41.4	40.4	18.2
Using Facebook Messenger?	314	39.5	44.6	15.9

^a.Variations in the sample size are due to missing data.

Data Analysis

Univariate analyses of covariance (ANCOVAs), with age, gender, education, marital status, race/ethnicity, and melanoma risk as covariates, were conducted to explore the effects of interactivity and character imagery on participants’ accuracy of identifying abnormal skin lesions and SSE intentions. An additional ANCOVA was conducted to examine the hot spot data and test the effects of attention allocation to different parts of the health messages on accuracy of identifying abnormal skin lesions and SSE intentions. Bonferroni correction was used to adjust the p-values for multiple group comparison. All analyses were conducted using SPSS 24.

Results

RQ1. There was no main effect of interactivity level or different types of characters on either accuracy of identifying abnormal skin lesions or SSE intentions (descriptive information of means and SDs is shown in Table 3). RQ2. Further, the interaction effects of interactivity x character on the accuracy of identifying abnormal skin lesions was not significant. RQ3. However, according to the results of the post-hoc comparison, individuals in the high interactivity and customization group (M = 4.37, SE = 0.22), compared to the low interactivity and cartoon group (M = 3.71, SE = 0.22) were more likely to accurately identify abnormal skin lesions (p< 0.05). Details are shown in Figure 2.

Table 3.

Descriptive Statistics for experimental conditions on dependent outcomes

Outcomes	Accuracy of identifying abnormal skin lesions			SSE intention
Group	Mean	SD	N	Mean	SD	N
High interactivity x Customization	4.52	1.64	52	3.35	1.20	52
Low interactivity x Customization	4.30	1.73	53	3.25	1.14	52
High interactivity x Cartoon	3.87	1.59	53	2.94	1.17	53
Low interactivity x Cartoon	3.70	1.55	54	2.83	1.08	54
High interactivity x Human	4.12	1.81	52	2.98	1.30	51
Low interactivity x Human	4.06	1.70	53	3.25	1.15	52

Figure 2. Accuracy of identifying abnormal skin lesions by different groups.

Note. Age, gender, education, marital status, race/ethnicity, and melanoma risk were controlled in the model as covariates.

Individuals in the high interactivity and customization (M = 3.28, SE = 0.16) or low interactivity and real human character group (M = 3.27, SE = 0.16), compared to the low interactivity and cartoon group (M = 2.80, SE = 0.16), reported higher intention to conduct SSE in the next 3 months (ps< 0.05). Participants in the low interactivity and customization group (M = 3.25, SE = 0.16) also has a marginally significant difference compared to the low interactivity and cartoon group (M = 2.80, SE = 1.08, p = 0.051). See Figure 3 for the mean score for each group.

Figure 3. Skin self-examination intention by different groups.

Note. Age, gender, education, marital status, race/ethnicity, and melanoma risk were controlled in the model as covariates.

Among the controlled covariates, individuals who reported higher melanoma risk(p< 0.05)and who were married (p< 0.001) also scored higher on accuracy of identifying abnormal skin lesions. Additionally, individuals who reported higher melanoma risk (p< 0.05), were married (p< 0.05), or had a college degree compared to high school or less (p< 0.05) also reported a stronger intention to conduct SSE.

RQ4. Results of the initial attention analysis (Hot Spot data)showed that individuals who paid attention to the character first (M = 5.91, SE = 0.23)compared to those who paid attention to the mole part first (M = 4.44, SE = 0.59) reported significantly higher accuracy of identifying abnormal skin lesions (p< 0.05) after controlling for age, gender, education, marital status (p< 0.01), race/ethnicity (p< 0.05), and melanoma risk factors as covariates. Married participants and non-Hispanics whites reported higher accuracy than those who reported other marital status and other races/ethnicities.

RQ5. On average, participants were interested in taking part in a skin cancer-related behavioral intervention that would teach them skin-examination and sun protection behaviors (M = 5.07, SD = 1.43). Moreover, participants gave positive feedback on the adjectives describing the SSE information they read during the educational program: interesting (M = 7.87, SD = 1.72), relatable (M = 7.92, SD = 1.74), insightful (M = 8.11, SD = 1.69), and informative (M = 8.47, SD = 1.49).Participants reported similar levels of interest in such a skin cancer-related behavioral intervention using an app on a mobile device (65.9%), by receiving information from their healthcare providers (62.4%),reading information on a computer-based website (62.1%), and reading information on a mobile-based website (61.8%). They reported lower levels of interest in receiving training via social media (i.e., Facebook and Instagram). More details are shown in Table 2.

Discussion

Digital media interventions and education programs have been utilized to promote skin cancer-related behaviors such as sun protection and skin examination, with different levels of effectiveness (Sar-Graycar et al., 2021; Finch et al., 2016; Ersser et al., 2019).What technological features and design elements will be more effective for learning the health information and changing skin cancer-related health behaviors warrants further investigation. The present study aimed to explore certain message features (i.e., interactivity, human character in imagery, customization), which have been shown to have an impact in digital skin cancer educations and interventions (Niu et al., 2019; Niu et al., 2021). The study innovatively explored interaction effects since the combination of different features can influence users’ interaction with and perceptions of digital media content on different media platforms. Findings included that individuals who reported higher melanoma risk and who were married scored higher on accuracy of identifying abnormal skin lesions and SSE intention. Additionally, individuals who had a college degree compared to high school or less also reported a stronger intention to conduct SSE. These findings are consistent with previous studies regarding skin cancer preventive behaviors (Niu et al., 2022). Future educational programs should consider building messages for individuals who are not married or are less educated.

The current study found that to increase people’s ability to identify suspicious skin lesions, health messages or programs with high interactivity and customized features worked together the most effectively. This means that allowing users to be able to interact with the digital platform or content and to modify certain features to suit a particular individual or task may improve learning about SSE and increase the accuracy to identify potential skin cancers.

High interactivity and customized features also worked effectively in increasing participants’ intentions to conduct SSE in the next 3 months. Real doctor characters in the educational materials with low interactivity could improve intentions to conduct SSE in the future as well. This finding is consistent with a previous study showing that human character(s) present in skin cancer interventions would increase receivers’ intention to use sunscreen (Niu et al., 2019). It is worth noting that the low interactivity and customization group also showed a marginally significant difference in participants’ intention for SSE compared to the low interactivity and cartoon group. Interactivity has been used in skin cancer research, and findings are mixed (Niu et al., 2021; Lustria, 2007). Further, interactivity may synergize with different features, and the effects on user engagement and behaviors may vary by context (Oh et al., 2019; Oh, 2017; Pajor et al., 2020). Therefore, more empirical research is needed to identify the effective combination of features and designs for digital education and intervention for different health contexts and target audiences.

Cartoon characters in general did not work as effectively as human characters. Although previous studies have shown that cartoon or animated materials work effectively in health education or intervention programs(Sheer et al., 2018; Hendriks & Janssen, 2018; Mukherjee & Dubé, 2012; Ahrens et al., 2006), it was not the case in the current study. Future studies should examine whether cartoon or animated characters can effectively influence behaviors related to skin cancer prevention and control.

Another important finding of the present study showed that participants who liked the human characters more than the moles/lesions in the educational materials tended to report greater intention to engage in SSE behaviors in the future. Characters in the educational materials could imply social settings and increase social aspects of motivation for participants when learning about the health messages. Mimicking the social setting, especially through presenting a presumed expert doctor character, in a learning environment may affect health behaviors (Bandura, 1991; Tougas et al., 2015). Future interventions could attempt to increase attention to characters using evidence-based persuasive social psychology strategies, e.g., by making them seem more credible, similar to the participant, or attractive.

Previous studies have examined interactivity and customization individually in skin cancer related research (Lustria, 2007; Niu et al., 2019; Niu et al., 2021; Noar et al., 2009). However, no study has explicitly assessed the influences of interactivity and customization together in a skin self-examination education program. Employing interactivity and customization in health interventions still requires additional empirical research. With many digital features and components integrated in the digital age, theories may be integrated to better assess one complex situation. This study lends support to the application of both visual communication theories and interactivity theory in skin self-examination educational programs using the digital media platforms and provided preliminary data for future interventions and/or educational programs that target both SSE knowledge and intention.

Limitations

This study used MTurk to recruit participants. M-Turk data provides a more socio-economically and demographically diverse sample than college samples and traditional Internet samples. However, there is a concern about the data representativeness and quality, with some prior evidence of the penetration of bots (Stokel-Walker, 2018). Although we had a reasonably-sized sample and used random assignment of participants to groups, we did not measure outcomes at baseline before their engagement in the education program. Future studies should assess variables at baseline and check if there is any significant difference between groups. We measured intention instead of actual behaviors. The identification of potential skin cancers employed digital pictures describing suspicious skin lesions. Future studies can focus on the actual identification on one’s skin. The measure of SSE intention used one item. Although this has been used by previous studies (Coups et al., 2011; HINTS), future studies should use a more robust scale. Finally, the current study used cartoon characters for a general adult population. Cartoon characters may work more effectively on the outcomes among younger populations such as teens or children.

Conclusions

When designing an educational program to teach SSE, customization and interactivity could be incorporated to improve both melanoma identification and SSE intention. Customized characters and using a real human character (i.e., a doctor) with high interactivity could increase accuracy and intention to conduct SSE in the future. The present study addressed a gap in the skin cancer prevention research literature related to identifying optimal approaches for teaching early melanoma detection skills in SSE training programs. Results from this research provided new insights on developing and designing future skin cancer education and intervention programs. However, lesion detection accuracy and SSE intentions were moderate even after brief training. Thus, research on more intensive interventions is needed.