Social media platforms have great potential for research on public health initiatives, as they have a wide reach and accessibility: 4.9 billion users and 227 million new users per year. Furthermore, social media platforms contain a diverse population sample: people from different backgrounds, regions and age groups use social media, providing a broad and varied pool of perspectives and experiences related to healthcare, which can lead to higher levels of objectivity. Additionally, social media can eliminate biases that could arise from conventional methods to create focus groups. Finally, social media can offer real-time information and updates on various healthcare topics.
Case studies on AI and social media for public health initiatives
An American study from 2016 used AI to understand the factors that drive the spread of information regarding cervical cancer prevention on social media. It found that individual messages about personal experiences were less likely to be effective regarding the spread of information on positive health behaviour, compared to organizational social media accounts. The key strategy that was developed as a result of this initiative, was to reinforce the credibility of organizational accounts and to develop messages that directly convey information and resources through organizational accounts.
Another study discovered that, while in the U.S. people are more likely to trust certified healthcare providers and organizations, in Korea and Hong Kong, however, there is a tendency to trust first-hand experience-based knowledge more. This necessitates localizing the research and information, in which AI can play an important role.
The role of AI in analyzing social media
One of the key challenges humans face while processing data on social media is the vast scale of information, which makes it impossible to objectively analyze, measure and extract insights from data. Furthermore, there is a lot of variability regarding the types of social media content and accounts, which makes it difficult to manually codify identification methods. Finally, the incompleteness and low quality of data on social media can cause the data to be overlooked or misunderstood.
In contrast, AI can offer powerful capabilities to address the incompleteness and quality of data. It can also help segment the online population by identifying patterns and characteristics of conversational styles for different user groups, allowing for tailored analysis and targeted interventions, as well as finding key trends and topics of discussion. Additionally, it can provide insights into the motivations behind people’s actions and decisions. AI can also identify underlining biases that influence the way information is interpreted, shared and discussed. Moreover, it can detect detailed emotional charges in online discussions. Understanding these nuances can help us appropriately engage in online discourse. Finally, as mentioned in the study above, AI can identify the cause and effect of why certain types of information spread across networks, which can be useful in propagating public health initiatives.
The Global Antibiotics Resistance Foundation (TGAR) and social media analysis by AI
In collaboration with the TGAR, the Innate platform has conducted AI-powered research on social media, addressing two questions regarding AMR (antimicrobial resistance): Can we identify topics for TGAR to spread awareness to improve engagement with the healthcare community? And: Can we use AI to find patients’ experiences with AMR to evangelize through TGAR’s patient platform?
The research involved collecting a sample set of social media posts from Reddit and Twitter and analyzing each post for account type and classifying them into different categories, which led to the emergence of specific themes. An interesting finding was that doctors who were overworked and fatigued caused an increase in AMR, as they were more likely to give antibiotics to desperate patients. Another contributing factor to an increase in AMR was the frustration and helplessness felt by patients with symptoms that kept persisting despite receiving multiple rounds of antibiotics. The research also found trending topics, such as efflux mechanisms and microbes, and bacteria resistance due to increasing temperatures. Increased awareness of these topics could lead to organizations like TGAR to increase their influence and credibility.
Nikhil Banerjee – founder and CEO of Innate
Written by Christine de Zwart