**1. Introduction**

Hearing loss is the most common human sensory deficit, a ffecting normal communication. Hearing loss is a multifactorial disorder caused by both genetic and environmental factors [1]. According to recent reports from the World Health Organization (WHO) [2], the global prevalence of hearing loss has increased rapidly over the last decade. The WHO estimated that, in 2018, 6.1% of the world's population or 466 million people experienced some degree of hearing loss. Furthermore, it is expected that the number of people with hearing loss will rise to 900 million by 2050 [2]. Hearing loss can negatively a ffect the quality of life as it has adverse e ffects on communication performance as well as emotional and social functions [3,4]. Thus, understanding the patients' perspective of their hearing loss is an essential step toward patient improvement.

Developing adequate healthcare and public health promotion interventions requires not only an in-depth knowledge of diseases and traits, but also a comprehensive understanding of patients' perceptions, misconceptions, concerns, and needs about the diseases. Clinical and behavioral interventions are more e ffective when seeking to improve outcomes that are central to patients' experiences and perspectives [5]. A review of prior literature showed that patient-centered interventions are responsive to patients' needs and thus, patients are encouraged to actively participate in the research process, especially regarding the identification of salient issues and concerns. By recognizing

the importance of patient engagement, scholars have increasingly paid attention to electronic patient-authored texts as sources of valuable information [6].

With the advances in Internet and communication technologies, healthcare professionals and patients can easily communicate about patient-provided data. In addition, patients not only can easily generate and share their health data and concerns with others, but also look for relevant medical information from peers who have experienced a similar diagnosis, set of symptoms or treatments. Fox and Duggan [7] investigated the use of online resources and found that 72% of Internet users used the Internet to search for health-related information, and 39% of these online health seekers looked for health information related to their own health or medical situation [7]. Research about health communication has demonstrated that health information exchanged on the Internet is a useful resource to examine patients' perceptions, concerns, and needs. Among Internet-based platforms, community-based question-and-answer forums are increasingly becoming popular as a medium for exchanging health information [8,9].

It is not surprising that patients with hearing loss or hearing difficulties use social media channels and social Q&A, such as Yahoo! Answers in English, Baidu Knows in Chinese, and Naver Knowledge-IN in Korean, to exchange information to address their concerns about hearing problems. Prior research has demonstrated that a collection of questions that patients post on social Q&A provides not only interesting but also important information to health professionals [9–11]. Topic analysis of patients' questions on the topic of hearing loss allows for achieving a thorough understanding of patients' perspectives, thereby leading to better development of the patient-centered intervention.

In this paper, we investigate the electronic patient-authored questions on social Q&A sites by applying the natural language analytics. This study adopts a multi-component semantic and computational linguistics method to discover and analyze themes or topics from hearing loss-related health conversations in the social Q&A sphere. Computational linguistics analysis is suitable to process unstructured textual data and identify hidden patterns in the data. Notably, a probabilistic topic modeling is applied to answer our research questions—topic prevalence, topic correlation, and topic evolution: what are hearing loss-related topics being asked or communicated in the social Q&A sphere? What are the prevalent topics in the hearing loss health community? How are those topics interrelated?

First, to the best of our knowledge, no prior studies have examined questions and conversations on social media Q&A as a viable source of understanding patients' prevalent concerns on hearing loss. By applying LDA topic modeling for text analysis techniques on electronic patient-authored hearing loss questions, our present study extracts latent topics and their dominant words, which might provide useful insights for understanding patients' perspectives and concerns on hearing loss. Second, our results reveal that patients post about the relationship between hearing loss and other diseases such as otitis media, tinnitus, and chronic kidney disease. These topic relations between categories have implications for the development of a patient-centered intervention. Lastly, our present research demonstrates evidence that topic model analysis techniques applied to electronic patient-authored questions are effective in studying patient-focused engagemen<sup>t</sup> strategy. The extant literature on hearing loss indicates that survey-based methods have been a primary study design to assess patients' concerns and perceptions about hearing loss and related issues. Using this traditional engagemen<sup>t</sup> strategy is costly to capture and analyze a massive volume of patients' concerns. Data analytics on patient-authored texts on social Q&A have the potential to improve patient-centered outcome research.

### **2. Related Literature**
