YouTube Videos as a Source of Information About Clinical Trials: Observational Study

Background: Clinical trials are essential to the advancement of cancer treatment but fewer than 5% of adult cancer patients enroll in a trial. A commonly cited barrier to participation is the lack of understanding about clinical trials. Objective: Since the internet is a popular source of health-related information and YouTube is the second most visited website in the world, we examined the content of the top 115 YouTube videos about clinical trials to evaluate clinical trial information available through this medium. Methods: YouTube videos posted prior to March 2017 were searched using selected keywords. A snowballing technique was used to identify videos wherein sequential screening of the autofill search results for each set of keywords was conducted. Video characteristics (eg, number of views and video length) were recorded. The content was broadly grouped as related to purpose, phases, design, safety and ethics, and participant considerations. Stepwise multivariable logistic regression analysis was conducted


Introduction
The release of new cancer treatments to market has outpaced all other therapeutic areas [1], with the introduction of 70 oncology treatments for more than 20 different tumor types over the past 5 years alone [2]. Bringing a new cancer treatment to the clinical setting is a complex process that extends over many years from the initial discovery and development in the laboratory through Food and Drug Administration (FDA) approval for use [3]. Integral to this process are the heavily FDA-regulated clinical trials that rigorously determine the safety and effectiveness of new and promising treatments among humans in an experimental setting [4].
Clinical trials are designed to answer specific research questions and are, thus, governed by strict protocols and eligibility criteria. Informed consent documents, which provide potential participants with detailed information about the purpose of the study, procedures to be performed, potential harms and benefits of the experimental agents used, and the voluntary nature of participation, are federally mandated to ensure that participants make informed decisions regarding enrollment. Much of the information presented to a patient is complex, incorporating translational research, biomarker selection, and sophisticated study designs into trials [5], but the level of health literacy remains low in the adult US population [6]. Fewer than 5% of adult cancer patients enroll in clinical trials [7], the most commonly cited barriers being lack of awareness or knowledge regarding clinical trials [8][9][10][11][12][13][14] and the reluctance to be randomized [15][16][17][18][19].
In one study, 92% of cancer patients reported the internet as the resource that empowered them when making treatment decisions and provided them with information with which to engage their physicians in discussion [20]. Cancer patients also use the internet to seek clarification, more detailed information, or reassurance about what was learned after a clinical encounter about clinical trials [21,22]. Much of the clinical trial information available online has been characterized as variable in quality with poor readability [23]. While clinical trial search tools are relatively easy to locate on the internet using various search engines, both content and functionality were also highly variable, and users needed a fair amount of knowledge about their condition and good web navigation skills to access the relevant information [24]. In a study that simulated the search for treatments of four common cancers by naïve cancer patients without clinical trial knowledge, only 85% of cancer-treatment sites mentioned clinical trials on the landing page and only 68% provided links to trials [23]. Another study that evaluated the navigability of comprehensive cancer center websites to clinical trial information observed that clinical trial content is narrow in scope with trial descriptions written at a college reading level, thereby limiting understanding for the average user [25]. When the written word proves to be challenging, consumers may turn to video-based information.
First introduced in 2005, YouTube is the second most visited website worldwide, and it has become an increasingly important medium through which health information is exchanged between and shared by consumers and health care professionals, government and nongovernment agencies, and industries [26]. Recent statistics indicate that currently, 300 hours of videos are uploaded to YouTube every minute and almost 5 billion videos are watched by 30 million visitors every day [27]. Despite the extensive reach and pervasive use of YouTube videos, little is known about videos related to cancer clinical trials; thus, the aim of this study was to evaluate the content of the most widely viewed YouTube videos related to clinical and cancer clinical trials.

Methods
YouTube videos posted prior to March 2017 were searched for using the keywords "clinical trial(s)" (426,000 videos), "cancer clinical trial(s)" (352,000 videos), "clinical trials cancer" (611,000 videos), and "oncology clinical trial" (619,000 videos). To reduce bias introduced in the display of videos by the search engine due to the location and search history of the study computer, searches were conducted using the incognito mode of Google Chrome in a single day, with results captured for later assessment [26]. A snowballing technique was used to select videos for review wherein sequential screening of the autofill search results for each set of keywords was conducted. A total of 25 search term options were initially identified ( Figure  1); 6 search term options were deemed irrelevant and excluded. The first 30 videos from each of the 19 remaining search term options were recorded; duplicate videos were removed, yielding 291 cancer clinical trial videos. YouTube uses a complex algorithm to rank video quality that is based on the duration the video has been watched. Longer viewer time indicates that the video is most likely appropriate for the search terms employed, which results in a higher ranking and greater likelihood of the video appearing on top of a search list [28].
Videos with <200 views (n=77) and those deemed irrelevant (n=80) were removed. Of the remaining 134 videos, additional 19 videos were found to be irrelevant upon viewing and were excluded from the final analysis. The remaining 115 videos were reviewed by 4 independent reviewers (GCH, SAM, KMS, and MB). Interrater reliability of the video characteristics and content-related variables, excluding the number of views, video length, and "thumbs up" and "thumbs down" was assessed by a fifth reviewer (CHB) and was found to be high among a randomly selected 10% sample (Cohen kappa=0.85). The following video characteristics were assessed: number of views, length of the video in minutes, year published, country of origin, video source (nonprofit organization, health care provider/organization or cancer center, school/educational organization, pharmaceutical or biotech company, clinical research organization, or other), and YouTube video category (nonprofits and activism, science and technology, education, people and blogs, or other). Style of communication was categorized as affective versus cognitive.
Videos were coded as affective in style if the content was presented verbally with overt positive or negative emotion that was persuasive in nature, whereas videos in which the content was delivered nonpersuasively, verbally or nonverbally, and without overt emotion were coded as cognitive in style. Also evaluated was the tone of the video (positive, negative, or neutral). Themes arising in the video (altruism/positive emotions, risks and dangers of clinical trials, advancing science, importance of volunteering for clinical trials, and other), the intended audience (patients, general public, caregivers, health care professionals, students, and research personnel assessed as to whom the information was being addressed), and the manner of presentation (lecture/course format/webinar, interview, testimonial, advertisement for paid participation, and other) were also evaluated. Viewer perceptions concerning the videos were also captured by assessing the "thumbs up," "thumbs down," and comment counts. Using the National Cancer Institute (NCI) Clinical Trials Information for Patients and Caregivers [29] series of documents as the reference standard, the content of each video was grouped a priori into five broad categories: clinical trial purpose, trial phases, study design, safety and ethics, and participant considerations. Reviewers derived the themes embedded in the content throughout the video viewing process; categorization of themes (eg, altruism/positive emotions, risks/dangers, advancing science, importance of volunteering, and other) was based on reviewer consensus.
Descriptive analyses, including calculation of frequency distributions, means (SD), and ranges, were performed. To assess video characteristics and content differences between videos for cancer clinical trials and clinical trials in general or videos with a focus on a disease other than cancer, univariable analyses using chi-square test for categorical variables and analysis of variance for continuous variables were conducted.
Stepwise multivariable logistic regression models, controlling for the year of publication, were constructed to assess the associations between video type (cancer vs noncancer) and variables with P values <.05 in the univariable analysis, controlling for the year of upload. All analyses were performed using IBM SPSS (version 24) [30]. Institutional review boards of each author's respective institution deemed nonhuman subject research exempted from review.

Results
The 115 videos related to clinical trials were viewed by approximately 11 million viewers with a mean number of 94,360 (SD 827,883) views each (Table 1).  With regard to the content, the most commonly mentioned item was the purpose of a clinical trial (62/115, 53.9%), followed by the fact that clinical trials are conducted in phases (44/115, 38.3%), information about specific phases (Phase I=46/115, 40.0%, Phase II=37/115, 32.2%, and Phase III=40/115, 34.8%), there is eligibility criteria for entering a trial (37/115, 32.2%), and that there are benefits to participating in a clinical trial (38/115, 33.0%; Table 2). Cancer-related clinical trial videos more often mentioned that Phase I studies evaluate how the drug effects the body and are used to determine potential side effects (4/46, 8.7% vs 0/69, 0%, P=.02) and that they may be offered in cases when no standard treatment options exist (5/46, 10.9% vs 0/69, 0%, P=.009) compared with noncancer-related videos. Cancer-related videos also differed from noncancer-related videos in that cancer-related videos more frequently mentioned the benefits of clinical trial participation, such as better care and monitoring (

Principal Findings
Our review of the 115 top viewed YouTube videos revealed that a large proportion of these videos are devoted to cancer clinical trials. Overall, clinical trial videos convey information that is aimed at both patients and the general population audiences. The majority of the videos presented the viewer with the overall purpose of a clinical trial and many discussed the phases of clinical trials and the fact that criteria are used to determine a patient's eligibility for enrollment. Beyond these topics, the video content varied widely, with most touching upon selected topics (eg, phases of clinical trials, federal regulations, informed consent or benefits of enrollment, and the importance of communication with a physician). Interestingly, none discussed the concept of clinical equipoise.
Cancer clinical trial videos were more positive in tone and more frequently used an affective communication style. They tended to emphasize altruism, the importance of volunteering to participate in a trial, and the benefits of participation more so than did non-cancer videos. Further, cancer clinical trial videos were nearly six times as likely to be created by a health care organization or a cancer center and were much more likely than noncancer trial videos to communicate practical information about clinical trial participation costs and to encourage dialogue with one's physician.

Limitations
Much attention was taken in the selection of the videos reviewed in this study to represent the most commonly viewed YouTube videos about clinical trials however, selecting the top 30 videos with greater than 200 views may have introduced a selection bias. Since the YouTube video ranking algorithm places videos with longer user viewing times at the top of the list and overlap was found in the videos in the top 30 for the search terms, our inclusion criteria likely captured the most widely viewed YouTube videos related to clinical trials. Further, we postulated that any bias introduced by the algorithm would similarly influence the videos displayed when a consumer uses the same search term and that the impact of less-viewed videos would be minimal. Despite using search terms specific to cancer clinical trials, we found that a large proportion of clinical trial videos were not related to cancer. While this was an unexpected finding, a consumer using our search terms would likely have the same experience. Whether or not viewers were engaged for the full duration of any video is unknown as 30 seconds of YouTube watching is considered a "viewing" [31]. This study is also limited in that it was cross-sectional in design and is further compounded by the fact that new videos are continually being uploaded on YouTube. Finally, this study focused solely on English language videos.

Comparison With Prior Work
To date, no other study has evaluated the contents of YouTube videos regarding clinical trials. More than 800 peer-reviewed publications reporting on the quality and content of YouTube videos relating to public health topics ranging from anorexia [32] to Zika virus [33] now exist, which is cause for concern regarding the power of this medium to communicate information accurately and responsibly to the general public. The decision to take part in a cancer clinical trial is a complex one, and the most common barrier to participation is lack of knowledge about cancer clinical trials [33]. Two separate studies found that the internet and media are the primary sources for learning about clinical trials [34] and that the information "read, saw, or heard" about a study was a major influence on the decision to participate. Further evidence supports that seeking information about one's illness can be viewed as a key coping strategy, which may lead to health-promotive activity and facilitate psychosocial adjustment to illness [35]. Use of the internet as a source of health-related information, however, has been likened to drinking from a fire hose and not knowing the source of the water [36], a sentiment that can easily be applied to YouTube video viewing. Because there exists no arbiter of the truth or accuracy of the material posted on YouTube, many question both the credibility and accuracy of the information and find that the content is influenced by perspectives of the video source [37][38][39][40][41][42]. Currently, the NCI acknowledges the importance of social media as a source of health-related information, and through its Cancer Moonshot Initiative, seeks to leverage this platform to provide patients with reliable information by developing a social media best practices toolkit. Information learned in this study showed that the majority of the clinical trial information communicated was accurate, as determined using NCI information for comparison [29] and was conveyed in a positive and compassionate manner. The coverage of topics however, was spotty, and the sufficiency and quality of information was lacking many times.

Conclusions
Overall, YouTube clinical trial videos provided information on many aspects of clinical trials, particularly cancer clinical trials. Few covered the full range of concepts needed to make an informed decision about participation; the majority focused on selected topics and provided varying levels of detail, leaving the viewer with an incomplete view of key concepts and partially informed. Given the abundance of clinical trial videos and relative ease of access to this information, care must be taken by patients and their families to verify and supplement YouTube video information with consultations with their healthcare professional to obtain a full and accurate picture of cancer clinical trials, thus, to make an adequately informed decision about participation.