Despite widespread availability and known effectiveness, Pap smear screening and human papilloma virus (HPV) vaccination rates remain below national targets. New modes of online communication through social media - which have transformed information sharing, especially for younger individuals - present opportunities to target health communication more effectively to improve cervical cancer prevention. Although the Internet has yet to realize its full potential for health-related communication, about 87% of US adults have used the Internet to look for health information, and a large majority of younger Americans report they would share personal health information on social media sites. Thus, pushing the current boundaries to capitalize on social media to inform decision-making about cancer screening and prevention offers great potential. The long-term goals of this research are to 1) generate validated methodologies to systematically study social media content, and 2) promote cervical cancer screening and prevention more effectively using social media. In pursuit of these goals, we will used a novel methodological approaches to characterize cervical cancer content present in social media and then to create and test messages within an experimental online social network.
In Aim 1, we will identify and rank the key words related to cervical cancer screening and HPV vaccination in social media messages. We will identify statistically significant phrases that are co-occurring in messages drawn from six months of publically available content from Twitter and Facebook. This technique will allow us to synthesize the entire breadth of content on screening and vaccination on social media, and to extract the full-text of messages corresponding to the most common significant words/ phrases.
In Aim 2, we will categorize and validate the range of cervical cancer screening and HPV vaccination discussion on social media websites. We will use qualitative content analysis informed by grounded theory to analyze a random sample of the findings from Aim 1, coding a minimum of 1500 messages. This will allow us to gain an in-depth understanding of the range of real-life discussions online. This will also serve as the validation of the machine learning result from Aim 1, as well as the basis for the development of new cervical cancer prevention messages that will be tested in Aim 3. Building upon both of these aims, in Aim 3, we will create tailored cervical cancer screening and HPV vaccination messages and track responses and spread in an online social network. Using the content knowledge from Aims 1 & 2, we will create tailored messages about cervical cancer screening and prevention and then pre-test them using virtual focus groups to gain insights from our target audience. We will establish a simulated online network to evaluate the spread of these messages, as each network member can directly share health opinions with others in the study. Within this setting, we will be able to determine the diffusion of specific, effective screening messages, as well as determine what types of network structures and message authors have the biggest impact.
We plan to study publicly available messages on online social media websites (for example, Twitter and Facebook) related to cervical cancer screening and prevention, such as comments about Pap smears. We will determine the major discussion topics on these websites, and then use this information to craft new messages about cervical cancer prevention. Finally, in an experiment, we will test how these messages are shared between individuals - in order to learn how public health agencies and researchers can improve their online health-related communication strategies to promote cervical cancer screening.
| Lyson, Helena C; Le, Gem M; Zhang, Jingwen et al. (2018) Social Media as a Tool to Promote Health Awareness: Results from an Online Cervical Cancer Prevention Study. J Cancer Educ : |
| Cherian, Roy; Westbrook, Marisa; Ramo, Danielle et al. (2018) Representations of Codeine Misuse on Instagram: Content Analysis. JMIR Public Health Surveill 4:e22 |
| Lyles, Courtney Rees; Godbehere, Andrew; Le, Gem et al. (2016) Applying Sparse Machine Learning Methods to Twitter: Analysis of the 2012 Change in Pap Smear Guidelines. A Sequential Mixed-Methods Study. JMIR Public Health Surveill 2:e21 |
| Chan, Brian; Lopez, Andrea; Sarkar, Urmimala (2015) The Canary in the Coal Mine Tweets: Social Media Reveals Public Perceptions of Non-Medical Use of Opioids. PLoS One 10:e0135072 |
