The association between YouTube use and knowledge of human papillomavirus-related cancers

Objective To examine the association between YouTube usage and HPV-related cancer knowledge (cervical, anal, oral and penile). Study design Cross-sectional study using data from the Health Information National Trends survey conducted between 2017 and 2020 (N = 16,092). Logistic regression was used to analyze the independent effect of YouTube use on cancer knowledge, controlling for sociodemographic characteristics. Results Respondents' knowledge of HPV-related cancers varied: 49.9% about cervical, 18% anal, 20.1% oral and 20.4% penile cancers. YouTube use was associated with increased knowledge for all cancers (cervical: OR 2.66, 95% CI 2.04, 3.46; anal: OR 1.83, 95% CI 1.32, 2.53; oral: OR 1.89, 95% CI 1.37, 2.61; penile OR 2.00, 95% CI 1.44, 2.77) in models adjusted for all covariates. Other independent predictors of HPV-related cancer knowledge included female gender, younger age, a higher income, and higher education. Conclusion(s) YouTube could play an important role in educating people about HPV-related cancers and should also target other populations, such as males and those with less formal education. Innovation The study provides novel insights into the potential of YouTube as an educational tool for promoting cancer knowledge with the goal of cancer prevention.


Introduction
In 2018, the prevalence of human papillomavirus (HPV) infections in the United States was estimated to be 43 million, with 13 million being new infections [1,2]. Although most HPV subtypes are cleared by the body without consequence, oncogenic HPV subtypes cause most cervical, anal, oropharyngeal, vaginal, vulvar, and penile cancers [1][2][3][4][5][6][7][8]. Widespread HPV vaccination has the potential to reduce cervical cancer incidence worldwide by as much as 90% while also reducing the need for screening and subsequent medical care, biopsies, and invasive procedures [5]. Vaccination also has the potential to protect against penile, anal and oropharyngeal cancers [1].
The Healthy People 2020 goal to have at least 80% of adolescents aged 13-15 years receive the Advisory Committee on Immunization Practices recommended number of appropriately spaced doses of HPV vaccine fell short, with only a rate of 54.5% by 2020 [9]. One possible approach to address this concern is through increasing awareness and knowledge about HPV infection and vaccination. Several research studies have demonstrated that adequate understanding regarding HPV can lead to an increase in interest, acceptance, intention to get vaccinated, as well as actual uptake rates for the HPV vaccine [10][11][12][13][14][15][16]. With social medial playing a significant role among those who access health information online, it presents an avenue where programs could reach out specifically aiming towards educating more people about the need for receiving the HPV vaccine. In fact, a survey conducted by Pew Research Center in 2021, found that YouTube is the most commonly used social media platform (81% of US adults, effectively 95% of those 18-29 years old use the platform) [17]. In addition, a study using Health Information National Trends Survey (HINTS) data from 2020 found that 40.8% of US adult respondents use YouTube to watch healthrelated videos [18]. A systematic review showed that many individuals are willing and interested in receiving accurate and helpful information regarding HPV and HPV vaccination via social media and are more comfortable doing so online than in-person [19]. As such, YouTube and the Internet impact the awareness, knowledge and attitudes in which individuals consume information, including HPV and HPV vaccination information [19].
Our objective was to examine the association between Internet and YouTube usage on knowledge of HPV-related cancers and to explore the role of underlying sociodemographic characteristics on this association.
Understanding these characteristics will allow for the creation of interventions geared towards specific sociodemographic groups to improve HPV vaccine uptake.

Data source
We used data obtained from HINTS 5 (Cycles 1-4, total sampled N = 66,723), a cross-sectional study of civilian, non-institutionalized adults 18 years or older living in the United States. HINTS examines trends in health information usage over time and provides data for conducting fundamental research on associations among cancer-related communication, knowledge, attitudes, and behavior at the population level. Data were obtained from HINTS and prepared for the National Cancer Institute by Westat. Cycle 1 data (total respondents n = 3285) were collected from January 2017 until May 2017; Cycle 2 data (n = 3504), from January 2018 until May 2018; Cycle 3 data (n = 5438) from January 2019 until May 2019; and Cycle 4 data (n = 3865) from January 2020 until June 2020. All households received a mailing of the survey either in English or in Spanish, and one adult in the household completed the survey. Details about the HINTS data source and methodology are available on the HINTS website [20].

Measures
The primary predictor was a trichotomous variable created as a combination of the Internet use and YouTube use variables. Internet use status was determined by answering the following question: Do you ever go online to access the Internet or world wide web. Or to send and receive email? A "yes" answer equated to an Internet user, a "no" answer equated to a non-Internet user. The YouTube use status was determined by the following question: In the past 12 months, have you used the Internet for any of the following reasons? Watched a health-related video on YouTube? A "yes" answer equated to a YouTube user, a "no" answer equated to what we subsequently referred to as a non-YouTube user. By combining these two variables, we created a trichotomous variable with the following subgroups: 1) non-Internet users/non-YouTube users; 2) Internet users/ non-YouTube users; 3) Internet users/YouTube users. For simplicity, we use the term "YouTube user" to describe this last group henceforth.
Sociodemographic characteristics (covariates) assessed were gender (male/female); race and ethnicity (non-Hispanic White [NHW], non-Hispanic Black [NHB], Hispanic, or Other); age, marital status, education level, income level, insurance, whether they had regular health provider, general health, and number of children under 18 in the household. Table 1 details additional categorization of these covariates.

Outcome
The primary outcome was knowledge of HPV-related cancers, assessed by a "yes" response to each of the following questions: 1) Do you think that HPV can cause cervical cancer? 2) Do you think that HPV can cause anal cancer? 3) Do you think that HPV can cause oral cancer? 4) Do you think that HPV can cause penile cancer? Answering either "no" or "not sure" on any question equated to not being knowledgeable about the respective HPV-related cancer; answering "yes" equated to being knowledgeable to HPV-related cancers. We restricted our analysis to respondents who answered "yes" to the preceding question "Have you heard of HPV?"; respondents who answered "no" were not asked any of the four HPV-related questions and were categorized as "Have not heard of HPV" (Tables 2-3, Figs. 1-2).

Statistical analysis
All analyses were performed using STATA version 17.1. Unweighted and population-weighted estimates (n, %) were calculated for all participant characteristics and outcome prevalence. A weighted multivariate logistic regression model was performed to assess the independent association of HPV knowledge among YouTube users compared with non-Internet users/non-YouTube users and to assess whether sociodemographic characteristics independently predicted HPV knowledge regarding HPVrelated cancers. We included all aforementioned covariates in this analysis. All weighted analyses accounted for the complex survey design using jackknife replication as documented in the HINTS technical manuals. Statistical significance was a two-sided p < 0.05 with no formal adjustment for multiple comparisons. Table 1 shows the demographic characteristics of the sample and the weighted population estimates. The average age of sample respondents was 54.8 years. Approximately 60% of respondents were female and 63% identified as NHW; 18% were non-Internet/non-YouTube users and about 31% used YouTube. A similar trend was observed across the weighted population estimates.

Results
Figs. 1 and 2 show the knowledge of HPV-related cancers by cancer type in the unweighted and weighted sample populations respectively. As seen in Fig. 1 (and similarly for Fig. 2), half of respondents (49.9%) knew that HPV causes cervical cancer, but fewer were knowledgeable about whether the other three cancers were HPV-related (18.3% for anal, 20.1% for oral, and 20.4% for penile). Approximately one third of all respondents had not heard of HPV. Table 2 shows sociodemographic characteristic differentials among those knowledgeable regarding HPV-related cancers. For all cancers, knowledge was highest among YouTube users, as well as those of younger age, NHW race and ethnicity, female gender, higher education, a combined household income $75,000 or more, having a regular provider, and having more than one child under age 18 in the household. Comparatively, across all sociodemographic characteristics, most respondents were not knowledgeable regarding HPV-related anal, oral, or penile cancers. Table 3 shows results of multivariate analyses identifying independent predictors of knowledge about HPV-related cancers adjusted for all other covariates. For cervical cancer, being a YouTube user was an independent predictor of knowledge after controlling for potential confounders (odds ratio [OR] 2.66, 95% confidence interval [CI] 2.04-3.46). Compared to NHB participants, NHW had higher odds of knowing that HPV causes cervical cancer (OR 1.43, 95% CI 1.11-1.83) and "Other" race and ethnicity had lower odds (OR 0.62, 95% CI 0.44-0.87).
Independent predictors for HPV-related anal and oral cancer knowledge were similar with a few exceptions. For anal cancer, those with one or more children under age 18 in the household (OR 1.39, 95% CI 1.17-1.65) had increased knowledge compared to those without children, but no effect was seen with race and ethnicity. For oral cancer, both an effect of having children in the household (OR 1.23; 95% CI 1.00-1.51) and NHW race and ethnicity (OR 1.40, 95% CI 1.07-1.84) were noted; no knowledge association was apparent for age. For HPV-related penile cancer, no significant associations were noted between knowledge and age, race/ethnicity, income or having a regular provider. However, having one or more child in the household was associated with increased odds of knowledge (1.39, 95% CI 1.16-1.67). Similar to HPV-related cervical cancer, YouTube use was associated with the highest increased odds of knowledge regarding HPV-related cancers (anal cancer OR 1.83, 95% CI 1.32-2.53; oral cancer OR 1.89, 95% CI 1.37-2.61; penile cancer OR 2.00, 95% CI 1.44-2.77).

Principal findings
We found that most respondents who have heard of HPV knew that HPV causes cervical cancer but relatively few knew of its relationship to anal, oral, and penile cancers. Though our analysis showed statistically significant knowledge gaps across various sociodemographic characteristics, this overall knowledge is consistent with previously published studies [21]. Our study uniquely highlights YouTube usage as an independent predictor of HPV-related cancer knowledge among respondents who know that HPV causes the respective cancers. By doing so, we associate YouTube as an important health information resource contributing to the knowledge regarding HPV-related cancers. This research is also important for exploring the characteristics of those with the most knowledge regarding HPVrelated cancers and their association to health information both on and off the Internet.

Results in the context of what is known
The emphasis on YouTube to distinguish levels of knowledge regarding HPV-related cancers stems from previous studies showing that YouTube is the most used social media platform and approximately 41% of the US adult population watch health-related videos on YouTube [17][18][19]. A systematic review by Ortiz et al. analyzing the quality of data on various social media platforms regarding HPV found that most videos about the HPV vaccine viewed in 2014 were anti-vaccine videos [19]. Consequently, this online information can be subject to misinformation and/or anti-vaccine rhetoric that could negatively influence vaccine uptake [19]. Nevertheless, the impact of social media on indivdiuals' awareness, knowledge, attitudes and behaviors related to HPV and HPV vaccination is both positive and negative. Though other studies have used HINTS data to explore the use of the Internet and/or social media with its association to HPV awareness and knowledge [22][23][24][25][26], little information is available exploring the association of all four HPV-related cancers with YouTube use while controlling for various sociodemographic characteristics among knowledgeable respondents.
The addition of YouTube use shows the highest odds of knowledge across all four HPV-related cancers. Unlike a recent study using 2020 HINTS data that did not find an association between HPV-related cancer knowledge and social media (including YouTube) [22], our analysis showed YouTube use was an independent predictor of HPV-related cancer knowledge. These differences in findings can be attributed to several reasons: 1) we used four years of HINTS data, thereby increasing the sample size and power of our study, 2) we specifically analyzed YouTube usage as a predictor of Internet-based health-information, and 3) we controlled for a greater number of covariates we believed could confound our analysis. Though we cannot conclude increasing YouTube use will lead to increases in HPV vaccine uptake, our study model uniquely depicts the role of  YouTube use as an independent health-information resource and its impact on HPV-related cancer knowledge. Furthermore, we also highlight gender and racial differences in HPVrelated cancer knowledge due to the persistent knowledge gap that exists in these subgroups [23,[25][26]. Unlike previous studies, we wanted to assess if these sociodemographic characteristics were independent predictors of knowledge of HPV-related cancers when controlling for YouTube usage. Our finding that females were more knowledgeable about HPV-related cancers compared to males was not entirely unexpected given the prevalence of HPV-related cervical cancer and deaths (as of 2018, approximately 311,00 women died from cervical cancer worldwide) 1 and from other studies [24][25]. This association is further strengthened by the fact that the initial licensing of HPV vaccines targeted cervical cancer prevention [23], and the recommendation to vaccinate males was made 5 years after the recommendation for females [21]. Highlighting the gender gap is critical because >40% of all HPV-related cancers in the US occur in men. Previous studies have shown that men are more willing to accept the vaccine when malespecific HPV-related health outcomes are emphasized [27]. As such, comprehensive information regarding HPV-related cancers is not being widely disseminated and needs to be inclusive of male-predominant cancers as well.
Our study also showed racial and ethnic disparities in knowledge regarding HPV-related cancers, a novel finding that was not shown in a previous study [26], perhaps due to our more expansive dataset and analyzing our data among YouTube users. Though statistical significance varied across each HPV-related cancer for race and ethnicity, overall, NHW was independently associated with higher odds of knowing that HPV causes cervical cancer, anal cancer and oral cancer compared to NHB.

Limitations and strengths
This study has several limitations. First, we had to create a trichotomous variable that combined Internet use and YouTube use because these questions were not mutually exclusive in the dataset. As such, across all 4 survey years, n = 324 respondents (2.1% of the total analyzed sample) identified themselves as non-Internet/YouTube-users. Rather than reclassifying these individuals as Internet users, since, by definition, one must use the Internet to access YouTube, we categorized them as data entered in error and excluded them. Our study was also limited in the depth of knowledge associated with the HPV-related cancers. Specifically, we could only assess whether knowledge was present or absent, but not the quality of that knowledge (whether informative or non-informative). As such, we cannot draw any conclusions regarding whether increased knowledge will lead to increased HPV vaccine uptake. Despite these limitations, our study had significant strengths. Primarily, it is a national, publicly available database designed to represent the United States population which allowed us to extrapolate data at a larger population level. Secondarily, it had a relatively large sample size to explore the association of sociodemographic characteristics with our outcome of interest.

Innovation
The novelty in our study lies in our detailed investigation of YouTube use as an important resource for patient education. This methodological approach innovatively utilized the most recently available HINTS data spanning four years, allowing for an increase in power and statistical significance of our analysis not seen in previous studies. By using such a large sample size, not only did we capture a representation of the population Have not heard of HPV" (blue) was considered a skip question in the overall survey. As such, for those who have heard of HPV, each respondent was then asked whether they know that HPV causes the respective cancer. These results are also depicted in row 1 of Table 2. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) HPV, human papillomavirus Fig. 2. Knowledge of HPV-related cancers by cancer group in the weighted population. This figure depicts the weighted distribution of respondents for each HPV-related cancer question. "Have not heard of HPV" (blue) was considered a skip question in the overall survey. As such, for those who have heard of HPV, each respondent was then asked whether they know HPV causes the respective cancer. These results are also depicted in row 1 of Table 2. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) HPV, human papillomavirus that can vary seasonally or cyclically for any number of reasons, but it also helped mitigate any potential confounding variables that can vary from year to year, such as changes in health policies or demographic shifts. In addition, to gain a more comprehensive understanding of the factors that affect knowledge about HPV-related cancers, we focused our attention to those with prior knowledge regarding the cancerous effects of HPV. Using this methodology enabled us to pinpoint particular factors linked with enhanced comprehension of these types of cancers, especially among those with prior knowledge. This information can be valuable in formulating specialized educational strategies aimed at strengthening cancer prevention efforts. In fact, our results show various independent predictors of HPV-related cancer knowledge other than YouTube / Internet use as a reflection of individuals also seeking their health information outside of the Internet. For instance, having a regular health care provider was an important independent predictor of HPV-related cancer knowledge. This finding emphasizes that the onus remains with the healthcare provider to ensure discussing these important health care topics with their patients, especially for those individuals without access to YouTube and/or the Internet. In addition, despite the misinformation that can plague YouTube and the anti-vaccine rhetoric even in the pre-Covid 19 era, there is evidence that knowledge regarding HPV-related cancers is retained and could impact HPV vaccine uptake in the future.

Future direction
Future directions in research include the use of implementation sciences to evaluate which content on platforms such as YouTube is most beneficial in increasing knowledge regarding HPV-related cancers and how that information might best be disseminated. It is a call to researchers to investigate the quality of HPV-related content online, improve the quality of health information disseminated offline since many individuals still seek their health information from other sources (e.g., their community, healthcare provider), and to ensure these messages are culturally appropriate. Additionally, with the introduction of new social media applications, such as Tiktok and Snapchat, it would be innovative for future studies to investigate knowledge and awareness of HPV and HPV vaccine across these platforms as these viewers compromise the predominant HPV-vaccine eligibility age group. Finally, it would be novel to see the association of HPV-related cancer knowledge, YouTube use and HPV vaccine uptake in the time of the Covid-pandemic where the anti-vaccine rhetoric was at an all-time high across social media.

Conclusion
Efforts to educate the public about human papillomavirus-related cancers should include both Internet and non-Internet platforms to target populations such as older individuals, males, and those with low incomes and less formal education. The innovative nature of this research lies in its ability to provide broader insights into the relationship between HPV-related cancers and knowledge while also suggesting that internet-based platforms such as YouTube could play an important role in educating people about HPV-related cancers.

Funding
No funding source was provided.

Declaration of Competing Interest
None.