Improving Recruitment Through Social Media and Web-Based Advertising to Evaluate the Genetic Risk and Long-Term Complications in Stevens-Johnson Syndrome and Toxic Epidermal Necrolysis: Community-Based Survey

Introduction

Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are rare, life-threatening drug-induced illnesses affecting 1-5 people per million each year, with mortality rates of 50% or higher in high-risk populations [1,2] SJS/TEN is strongly associated with human leukocyte antigen (HLA) class I alleles, providing an opportunity for preprescription risk prediction, as well as understanding genetic variation across racial and ethnic groups [2,3]. Although studies suggest that the presence of a genetic risk factor has equal implications across populations, the relevance of specific HLA risk alleles may differ due to differences in the HLA carriage rate [3-6]. The study of genetic risk and long-term sequelae of SJS/TEN across racially diverse populations has been burdened by its rarity, social disparities, and trust or access to health care [4,7]. Social media has been advocated as a strategy to improve study diversity and recruitment in rare diseases and increase accessibility for study participation [8,9]. To enhance awareness and enrollment in our community-based study, we implemented social media and web-based advertisements, measuring the utility of each channel in increasing interest and diversity.

Methods

Overview

Using historical recruitment data from our community-based study of SJS/TEN survivors, we compared data both before and after implementation of social media and web-based recruitment methods. We recruited SJS/TEN survivors from the United States, aged 7-90 years, to participate in a remote study to assess genetic risk and long-term outcomes in SJS/TEN. Survivors across the United States, who saw or heard about our study, filled out an online interest survey. Data on recruitment and enrollment for this study were collected from September 2019 through December 2023 using the online tool Research Electronic Data Capture (REDCap) and analyzed using Microsoft Excel.

Recruitment

Baseline recruitment methods included advertisements on the SJS Foundation website and in the American Burn Association newsletter for physician referrals. Two years into the study, enrollment consisted of 147 (72.8%) female participants and 156 (77.2%) participants identifying as non-Hispanic White. In June 2021, we consulted institutional resources to design and implement advertising campaigns across several social platforms to increase interest and diversity. Advertisements were introduced on the Vanderbilt University Medical Center (VUMC) Facebook and Instagram pages using flyers, and 60-second SJS/TEN survivor video vignettes were later added. Next, we launched a nationwide Google Ad campaign. Advertisement-derived interest surveys identified potential participants who were then contacted by phone. Interest surveys reported the mode of recruitment through a single-choice answer of “a website,” “Facebook,” “Instagram,” “Google,” “referral,” or “other.” More detailed responses for choices like “other” were obtained through free response options. Referrals are defined as participants referred to the study by family, friends, or a physician. Ethnicity, race, gender, and age categories were self-reported and followed the National Institute of Health’s reporting guidelines.

To understand the impact of each recruitment channel, we measured rates of study interest before and after the ad implementation, as well as the “enrollment effectiveness” of each channel. Furthermore, these were compared to highlight differences in race, ethnicity, gender, and age.

Ethical Considerations

The community-based SJS/TEN Survivor Study was approved in August 2019 by the VUMC institutional review board (IRB#191350). Written informed consent was obtained from all enrolled participants. All participants were given the opportunity to withdraw from the study at any time. Participants were given a US $50 Amazon eGift-card as compensation for participation.

Privacy and confidentiality of participants were maintained. All data collected were deidentified before analysis. Data were stored on REDCap, a secure Health Insurance Portability and Accountability Act (HIPAA)-compliant database.

Results

Prior to implementing ad campaigns, 6.9 unique interest surveys were completed, and 3.4 participants enrolled per month. Following the launch of social media and Google Ads, 26.9 unique interest surveys were completed, and an average of 5 participants enrolled per month, representing a 289.5% increase in interest and a 48.6% increase in enrollments. Ineligibility reasons included lost to follow up (n=253), not diagnosed with SJS/TEN (n=156), duplicate survey (n=73), inability to obtain records (n=65), hospitalized outside of North America (n=38), and declined to participate (n=38; Figure 1).

‎

Completed interest surveys and enrollment rate were broken down by recruitment channel. The SJS Foundation website produced 149 responses/69 enrolled (46.3% enrollment rate). Google Ads followed with 201 responses/56 enrolled. VUMC Facebook ads contributed 163 responses/25 enrolled (15.3% enrollment rate). Instagram had 7 responses/1 enrolled (14.3% enrollment rate). Participants were also referred by family, friends, and physicians, with 134 responses/41 enrolled (30.6% enrollment rate; Figure 2).

‎

Discussion

Principal Findings

Recruitment in studies involving human subjects is often inadequate due to factors including limited patient awareness, lack of trust, and logistical barriers [7-10]. These challenges are frequently compounded when conducting trials on rare diseases like SJS/TEN, partly due to the small number of diagnosed patients. Most studies on SJS/TEN have been country-specific and may not be widely representative of heavily diverse populations like the United States [3,11]. This study is unique as we pulled survivors from the community within the United States, hoping to gain diverse representation. To increase accessibility, enrollment, and knowledge of our trial on SJS/TEN, we implemented several social media-based recruitment methods.

The addition of social media and Google Ads increased interest and enrollment as well as racial, ethnic, and age diversity in recruitment. Google Ads brought in the highest number of responses to the study and the second-highest enrollment rate after the SJS Foundation. Emotional appeals from the survivors’ video testimonials were vital in capturing interest. Supporting this, one study found that after conducting 37 in-depth interviews, patients preferred study recruitment messages through infographics, emotional appeals through images, or educational posts [10]. However, Facebook ads produced inferior responses to Google Ads. Instagram did not gain traction for recruitment, even though the same video testimonials were used.

When evaluating the effectiveness of advertisements on gender, we saw increased male interest, but decreased eligibility and enrollment. Therefore, there was no true impact on gender in terms of enrollment with the increase of social media and web-based advertisements.

An important finding was the increase in responses from people aged over 65 years, in contrast to other studies showing less online engagement from older adults [12]. Older age increases the risk of SJS/TEN-associated morbidity and mortality, making it vital to access older age groups [11,13].

Even with our advertising efforts and the increase in diverse responses, our enrolled participant population remained largely non-Hispanic White females. While a more nuanced understanding is needed, if there is a known HLA allele risk for SJS/TEN for a medication, the implications for carrying the risk allele are the same regardless of country of origin or background [4]. However, the prevalence of different risk alleles does differ by specific population, driven by genetic race. Therefore, the proportion of individuals who develop a reaction to a specific drug can be very population-specific [2,4,5,11]. Given the US population diversity, we need to see more equal representation in SJS/TEN research from all strata in the population to holistically understand the impact of SJS/TEN across the United States.

Limitations

For the first 10 months of the study, participants were not asked how they heard about our study, resulting in missing recruitment data on 232 unique interest survey responses and 121 enrolled participants. Main recruitment avenues at that time consisted of the SJS Foundation and American Burn Association physician referrals. Another significant limitation is the ongoing cost incurred when using paid advertising. Although we found that social media and Google Ads increased diversity and inclusion, we acknowledge that not everyone can access them. In some cases, older participants will not seek information via the internet [12]. Therefore, exclusive use of web-based search tools could limit accessibility to the study. However, increased diversity associated with referrals and Google Ads supports the use of these advertisements as a valuable source of communicating study information to SJS survivors and their families.

Conclusions

Our results demonstrated that social media and online recruitment approaches can increase recruitment interest and enrollment in a community-based study of a rare disease such as SJS/TEN while improving inclusivity, diversity, and age ranges. Increasing representation of enrollment across age and different strata of the population supports the unmet need to understand the risk and impact of SJS/TEN on diverse populations. Social media and web-based search tools can differ in enrollment effectiveness and target audiences. Our partnership with the SJS Foundation paired with Google Ads was most effective in reaching SJS/TEN survivors. This study highlights the use of a variety of social media and web-based advertising methods and their potential impact on reaching diverse and hard-to-reach populations. We improved overall enrollment, diversity, and age inclusivity, overcoming recruitment hurdles commonly seen in the rare disease community.