Unlocking Patient Voice on Social Media through AI

Social Media has become an information goldmine for extracting patient insight and opinion. Listening to what patients are talking about can lead to better engagement and patient centricity. This has become a priority for many healthcare organizations. According to a recent survey by Deloitte of various Technology Executives and Leaders of Health Systems were hoping to achieve better patient experience (92%) as the top desired outcome [1] from digital transformation.

Patient voice can be defined as the gathering and analysis of first-hand experiences of patients with the intention of improving patient representation in the development of treatment options and support systems. In a business context, Patient voice can be leveraged to improve go-to-market efforts and customer intelligence. Engaging with patients online can improve brand perception and loyalty, especially with younger generations. This is important since public perception is an area Pharma has traditionally struggled with.

Key points discussed in this article:

• Unlocking patient voice

• Online vs traditional

• Types of platforms used

• Motivations to use social media

• Types of discussions

• Cultural and regional nuances

• Using AI for scale and insight

What does unlocking patient voice on social media mean?

Online patient voice can be harnessed to build social intelligence that can aid in answering key business questions. Whether it be identifying reasons for treatment switch; or identifying fears and unmet needs for crafting an awareness campaign; the answers to these questions can be supported by data collected by analyzing patient voice at scale. However, synthesizing patient discussions into data driven decision making requires capabilities that are beyond what traditional social listening tools can currently offer.

There is a lot of potential business value and collective insights in online discussions that can aid various stakeholders in the healthcare ecosystem. Assuming we have the right tools for the job, these are now accessible. Improvements in accuracy, capability and scale in text analytics, combined with the customization of these in a medical context is leading to a new breed of tools. Some examples of these capabilities are:

• Inferring motivations and behavior of patients

• Mapping out key trends and topics of discussion

• Inferring biases in perception

• Detecting detailed emotional charges in online discussions

• Tracking the impact of adverse events or industry or legal challenges

• Observing patterns in the proliferation of information across the network

• Segmenting the online population for precise analysis and targeting

Doing this at scale, seamlessly with ease of access to the data is a key challenge. Moreover, it is essential to design specialized tools for specific medical use cases, as opposed to relying on traditional social listening tools that are more generic and cross-industry.

Why online as opposed to traditional methods?

The pandemic has accelerated the reliance of patients on online peer-networks for healthcare support. This presents an opportunity for mining useful insights from online patient voices and keeping track of emerging behavioral patterns and topics of interest. Furthermore, the anonymity that virtual environments present leads patients to be more honest with their opinions. Social media also represents an easily accessible, geographically and demographically diverse set of people. This can lead to higher degrees of objectivity and remove biases that traditional methods of assembling focus groups might fall prey to.

Another challenge with traditional methods is that of scale. With easier and cheaper access to technology, social media is being adopted at a rapid rate across the globe, averaging 227 million new users per year [2]. However, analyzing text from social media discussions and gathering insights at scale is highly time consuming without any automation in place. This has changed with the advent of AI and advanced text analytics capabilities. The quality and detail of insights available with these technologies can be compared to a human expert analyzing the data. This is a paradigm shift in the possibilities to understand patient behavior and needs at a scale which was previously impossible.

What type of social media platforms do patients use?

Social media networks use by patients typically fall under two categories: general-purpose online social networks (Twitter, Reddit, Facebook, Instagram, TikTok) as well as specifically designed networks for patients such as PatientsLikeMe, TreatmentAction Campaign, etc. Either type of network is likely to be used to disseminate knowledge about health and influence health behavior.

Each platform has differences in intent of use[3]. For example, Reddit is used for peer interaction and community building. Discussions here are experience driven and less scientific. The community relies on informal education of its peers. Facebook and Instagram have medical activity in closed groups, but public discussions are in the form of image and video-based storytelling. More specialized online forums are championed by patients that are already well educated and are looking for advanced information on trials, cutting edge medication, etc. Twitter is predominantly used by healthcare professionals, patient advocates and medical organizations sharing news.

Why are patients motivated to using social media?

Information on social media is immediately accessible and vetted by peers. This community-driven inclusive nature of social media grants a greater perceived sense of control to patients over their own health [4]. Moreover, digital environments have benefits of anonymity, providing further control on identity and condition. This is especially useful for overcoming cultural or social stigmas that might impede access to care (for example, in the case of mental health of sexually transmitted infections) [5].

Patients express their needs, issues, concerns, opinions, etc. often seeking emotional support through a sense of community. For example, it has been demonstrated that social ties in the virtual environment reduce risks of depression among the elderly and boosts their self-confidence[7].

People find meaning in assisting each other through their journeys and the sense of belonging to a wider community. These drives can be understood and addressed by organizations to create more meaningful relationships with patients online. Moreover, such psychosocial behavior and motivational frameworks can give us powerful models in organizing and distilling information from these online communities at scale.

What kind of discussions are patients having on social media?

There is vast information diversity on social media, patients seek advice and information on various topics such as symptoms management, lifestyle challenges or treatment selection. Patients often express unmet needs and frustrations, these discussions can be quite emotionally charged and are motivated towards seeking support.

For example, this study analyzed a Reddit forum for discussing issues related to chronic pain [10]. It was found that “back pain” was the most discussed body part. Patients were focused primarily on discussing treatment trajectories proposed by medical physicians.

Patients often use social media to celebrate achievements such as receiving effective treatment which boosted their quality of life. They talk about coping techniques that could help them lessen symptoms and restrict future progression of their condition.

Cultural and regional nuances to analyzing Patient voice

Informational ecosystems that are endorsed and run by peers provide a trustworthy environment to seek support. However, there are certain differences when it comes to regions. For example, while in the US patients are more likely to trust in certified HCPs, in Korea and Hong Kong there is more trust in information provided by peers. There is a tendency in Koreans and Hongkongers to trust and use experience-based knowledge to a greater extent than Americans [11].

A number of papers have also addressed cultural characteristics (Hofstede’s framework) such as masculinity, collectivism and uncertainty avoidance can impact social media use and adoption. Cultural factors have an impact on perceived usefulness, importance and ease of use of social media. Research in Pakistan and Taiwan [8,9] with this framework has validated this. Both are masculine cultures and have a significant positive impact on perceived usefulness of social media. Collectivist cultures perceive that social media use would be beneficial not only for themselves but for overall society. Cultural models can be used to calibrate messaging that is more in line with a particular region.

Translating Patient voice from Social Media through AI

Translating patient voice at scale from social media has been made possible through advancements in AI. Text analytics is an area that has had recent breakthroughs with large language models such as GPT3 being made accessible to the public. The art of the possible has expanded to human-expert-level analysis to intuiting nuances in written and spoken word.

Now we can focus on the bigger questions are around what information would be useful to detect and gather from patient discussions at scale. The first form of assistance that AI can provide is to organize and categorize discussions into various taxonomies that are conducive to extracting specific insights. For example, knowing which phase of their journey are patients seeking out assistance from social media communities; and furthering an analysis of their emotional state can help understanding expectations both from an informational and emotional perspective.

A lot can be inferred from a little information gathered from each actor when the number of actors are large. A combination of detection and classification at a per post level can lead to population level insights that are quite useful when making data-driven decisions. For example, knowing that patients are anxious about getting themselves scanned — “Scanxiety” is something that is being seen in therapeutic areas apart from Oncology.

Things become a lot more interesting when we consider interactions between actors from a network perspective. Notions of how information diffuses through networks are being challenged as new insights are being discovered.

Conclusion

There are various nuances when it comes to why patients use social media and culture, platform type, demographic and the dynamics of a given therapeutic area have a big role to play in the successful interpretation of patient voice. Being aware of this context can lead to more accurate insights and analysis. Unlocking patient voice is a complex task. Some of the key factors of doing this effectively are highlighted in this article. Doing so automatically requires a sophisticated system designed specifically for this purpose.

As we see advancements in AI leading to advanced text analytics, we can rethink how current Social Listening tools can be enhanced and used. Once individual voice is decoded successfully, this can lead to population level insights that are only possible through the scale that social media provides.

At Innate, we are involved in a number of experiments in collaboration with industry leaders, data scientists and academics to uncover insights that are innately present in these networks of voices but are yet to be seen and discovered. Moreover, our product for patient voice is now in beta. Please feel free to get in touch.

Please follow us on LinkedIn for our latest efforts🔥:

https://www.linkedin.com/company/innate-ai/

For a demo ⚡of our product for AI driven Patient Voice contact us on: contact@innate-ai.com

Or call us directly, for a real or virtual meeting/drink☕🍺: +358 44 936 6352

References:

1. https://www2.deloitte.com/us/en/insights/industry/health-care/digital-transformation-in-healthcare.html

2. https://www.smartinsights.com/social-media-marketing/social-media-strategy/new-global-social-media-research/#:~:text=More than half of the,social media is 2h 29m.

3. https://mpfs.io/assets/cmi2020/2022/07/Social-Listening-Perspective-Patients-HCPs-Using-Social-Platforms-for-Different-Purposes.pdf

4. Pretorius C, Chambers D, Coyle D. Young people’s online help-seeking and mental health difficulties: systematic narrative review. J Med Internet Res 2019 Nov 19;21(11):e13873 [FREE Full text] [CrossRef] [Medline]

5. Zhang X, Wen D, Liang J, Lei J. How the public uses social media wechat to obtain health information in china: a survey study. BMC Med Inform Decis Mak 2017 Jul 05;17(Suppl 2):66 [FREE Full text] [CrossRef] [Medline]

6. Naslund J.A., Aschbrenner K.A., Marsch L.A., Bartels S.J. The future of mental health care: Peer-to-peer support and social media. Epidemiol. Psychiatr. Sci. 2016;25 :113–122. doi: 10.1017/S2045796015001067. [PMC free article ] [PubMed ] [CrossRef ] [Google Scholar ]

7. Banegas J.R., Cabrera R.H., García E.L., Pérez-Regadera A.G., Rodríguez-Artalejo F. Social network and health-related quality of life in older adults: A population-based study in Spain. Qual. Life Res. 2005;14 :511–520. doi: 10.1007/s11136–004–5329-z. [PubMed ] [CrossRef ] [Google Scholar ] [Ref list ]

8. Lin HC, Ho WH. Cultural effects on use of online social media for health-related information acquisition and sharing in Taiwan. Int J Hum-Comput Int 2018; 34: 1063–1076.

9. Social media use for health, cultural characteristics, and demographics: A survey of Pakistani millennials, https://journals.sagepub.com/doi/10.1177/20552076221089454#bibr10-20552076221089454

10. Goudman, L.; De Smedt, A.; Moens, M. Social Media and Chronic Pain: What Do Patients Discuss? J. Pers. Med. 2022, 12, 797. https:// doi.org/10.3390/jpm12050797

11. Song H, Omori K, Kim J, Tenzek KE, Morey Hawkins J, Lin WY, Kim YC, Jung JY. Trusting Social Media as a Source of Health Information: Online Surveys Comparing the United States, Korea, and Hong Kong. J Med Internet Res. 2016 Mar 14;18(3):e25. doi: 10.2196/jmir.4193. PMID: 26976273; PMCID: PMC4810010.