We examined data from HINTS 4 cycle 4 (2014) of the National Cancer Institute. HINTS is a cross-sectional, nationally representative household survey conducted annually, which collects information on demographics and the perceptions and use of health information, and provides information on how cancer risks are perceived. HINTS 4 cycle 4 had a sample of 3677 participants. Data from 2014 were used because the outcome of interest, “How interested are you in exchanging digital images or video (eg, photos of skin lesions) with a health care provider electronically?” was only available in this cycle. Details about HINTS data collection, including weighting methodologies, is described elsewhere. The institutional review board of Emory University exempted this study from review.

The covariates include sociodemographic variables as well as access and use of the internet such as the following: self-reported gender, age, education, race and ethnicity, annual household income level, seeing a health professional regularly, having health insurance, using the internet, using electronic devices to share medical information with a health professional, and using various devices (if any) with or without health apps on them. The survey question on devices was, “please indicate if you have one of the following electronic devices: tablet, smartphone, cellphone, etc.” The responses were as follows: “(1) tablet computer like an iPad, Samsung, Galaxy Tab, Motorola Xoom, or Kindle Fire, only; (2) smartphone such as an iPhone, Android phone, Blackberry device, or a Windows phone; (3) basic cellphone only; and (4) multiple devices listed.” The survey question regarding health apps contained on one’s electronic device was, “on your tablet or smartphone, do you have any software apps related to health?” The answer choices were as follows: “yes,” “no,” and “don’t know.” These 2 questions were combined to generate 1 variable to specify the type of electronic device a participant owned and if there were health apps installed on that device.

We also assessed cancer information–seeking behavior, including, but not limited to, skin cancer. Cancer-related covariates included the following: seeking cancer information, trusting web-based information on cancer, and using the internet to obtain cancer-related information for oneself in the past 12 months.

The primary outcome was defined by the following question: “how interested are you in exchanging information like digital images or video (eg, photos of skin lesions) with a health care provider electronically?” We dichotomized the outcome by a high level of interest (responding with “very”) and those who did not have a high level of interest (responding with “somewhat,” “a little,” or “not at all”) in exchanging images and videos with health care providers.

Statistical analyses were conducted using SAS (version 9.4, SAS Institute). Nationally representative prevalence estimates were obtained using jackknife replicate weights that accounted for the complex survey design. For bivariate analyses, the associations between specific sociodemographic groups and interest in sharing photographs or videos were assessed using unconditional logistic regression. We used complete cases analysis for logistic regression owing to the low proportion of missing data (6%). A multivariable logistic regression model was developed through backwards selection, with all final covariates associated with varying levels of teledermatology use at P<.05. Furthermore, sensitivity analysis was performed by changing the outcome dichotomy to model those who were “not at all” interested. Two-sided tests were performed with P<.05 indicating statistical significance.

High levels of interest in using teledermatology were associated with modifiable behaviors such as the use of devices with health apps, trust in web-based information on cancer, and prior experiences in exchanging health information with physicians on the internet. Sociodemographic factors such as young age, female gender, high education, and high household income were not associated with an increased interest in exchanging images of skin lesions with health care providers after adjusting for these modifiable variables. Future implementation of teledermatology should address these identified factors within the digital divide to provide equitable access to care across diverse patient populations.

Access to devices and how they are used are aspects of the digital divide, which can be adjusted. Physical access to the internet was found to be the most significant predictor of web-based patient-provider communication [9]. Once access is established, usage becomes the rate-limiting factor, which depends on the ability to retrieve and search for information on the internet and to use mobile health apps [11,12]. Users of mobile health apps are more likely to exhibit health-promoting behaviors than those who own similar devices but do not use health apps [11]. A potential barrier to the use of health apps is privacy concerns with inputting personal data into digital devices. Digital health information requires high levels of eHealth literacy to effect action [13]. Mobile device and health app usage is associated with characteristics previously linked to increased eHealth literacy, such as young age, higher education, and high income [7,14]. However, unlike age, education, and income, access and usage of devices can be modified to enhance teledermatology implementation.

Trust in web-based information on cancer can be directed to mediate interest in teledermatology usage. Trust is a necessary antecedent to the development of eHealth literacy and engagement with health information [15,16]. Trust in web-based health information is associated with higher education and the disclosure of health information on the internet, which are factors linked to eHealth literacy [17]. Even among groups with increased eHealth literacy, young individuals have higher trust in web-based health care services than their older counterparts [18]. Patient trust in a telemedicine service can be broken down into their trust in the organization, treatment, care professional, and technology [19]. Increased trust in the telehealth service can be gained when patients are referred by other health care professionals [20]. Moreover, face-to-face interactions with the provider prior to a web-based consultation also increases patient trust [21]. The provider should ensure that patient concerns are being addressed as this will increase trust in the telehealth provider and the treatment plan [21].

Prior experience in exchanging health information on the internet is an adjustable factor that can be targeted to increase interest in exchanging digital images. Lack of knowledge and experience with web-based patient-provider communication is found to impede its use [9], while prior experience with sharing medical information electronically is associated with higher interest in exchanging digital images with providers [15]. As patients expand their experiences with digital technology within and outside of the health care context, patients will have the opportunity to develop trust in teledermatology services [19].

Teledermatology interventions are implemented and expanded across populations to bridge the digital divide [22]. For example, the US Department of Veteran Affairs has expanded the reach of teledermatology by loaning electronic devices to veterans and provided training in using the devices so that veterans can more easily connect with the existing telehealth networks [23]. They have also attempted to evaluate the effectiveness of these measures through a survey on patient satisfaction with teledermatology use and its contributory factors [24]. Our study found that factors malleable to influence—use of health apps, trust, and experience—are barriers that can be mediated to increase the reach, adoption, and effectiveness of teledermatology.

There were some limitations to this study. The cross-sectional nature of the HINTS data precluded the establishment of causal relationships between the usage, trust, and prior experience in using health apps and the interest in teledermatology. The measured outcome was available only in HINTS 4 cycle 4 (2014); more recent data were not available to address the study question. All survey responses were self-reported and subject to information bias. We could not exclude residual confounding variables from additional unmeasured or unexamined variables. We were unable to distinguish the history of skin cancer from that of other cancers when controlling for covariates related to information-seeking behaviors associated with cancer. Interest in exchanging images on the internet might differ if the patient worked with GPs to send images to a teledermatology service or if the service is directly patient-initiated, and this aspect should be examined in future studies. Future studies should explore how well patients and GPs follow teledermatology guidelines on taking adequate images. We were unable to assess if interests in teledermatology translated directly to teledermatology usage or adherence to recommendations from teledermatology services.

In conclusion, modifiable access barriers to teledermatology adoption included experience with exchanging health information on the internet, trust in web-based information on cancer, and the use of mobile health apps. Future implementation of teledermatology should address these identified factors within the digital divide to provide equitable access to care across diverse patient populations.