Prediction – the next frontier for social listening

, ,

Prediction – the next frontier for social listening

By Elizabeth Fairley, Ph.D.

A patient’s journey is a complex equation, with multiple factors affecting both the course of the journey and the eventual health outcome. Utilizing social listening data to understand patterns and trends in patient behaviors can now make that journey easier, improve quality of life and result in better outcomes for patients.

At present, we use social listening data in the medical advertising industry to gain insights into how, when and where patients are talking about their medicines. By listening to the real patient voice, we can gather intelligence about their feelings towards treatment, which we can then feed back into the brand planning process with the aim of more effectively targeting advertising campaigns, increasing patient uptake of medicines and, ultimately, improving outcomes for patients. 

While the current status quo for social listening provides valuable insights into how patients think about their medicines, many see prediction as the next frontier, allowing advertisers to more effectively and efficiently target patients based on predicted behaviors, needs and responses. 

What are the benefits of prediction for advertisers and patients? 

As we’ve already highlighted, patient journeys – from diagnosis through to treatment – are highly complex. Understanding those journeys is absolutely key to making treatment more efficient and effective. Through AI we can more accurately predict common touch points and shared experiences within those journeys. 

Take, for example, the issue of ensuring patients take the correct medicines dosage through the entire course of their treatment. We know that 50 percent of medicines are just not taken as prescribed (World Health Organization) – forgotten, wasted, waylaid – so there is a disconnect between those making medicines and those taking them.

In this instance, prediction could provide a solution. By predicting future patient behaviors, AI could enable advertisers to warn patients with a specific diagnosis and treatment that they’re about to fall into a common trap of, for example, only taking one tablet a day rather than two. Helping patients to avoid this trap could have a huge impact on their treatment and health outcome. 

What if we could recognize the chances of a particular health outcome occurring to predict disease progression over time? With AI-based prediction, we can do just that. Take for example, a woman in her 20s who begins to take a contraceptive medication. This might be for birth control or another condition like endometriosis. She is then diagnosed with an underactive thyroid in her late 20s and comes off birth control to have two children. At aged 55 she is diagnosed with rheumatoid arthritis and starts to take calcium supplements. At 75 years she stops taking calcium, is still taking thyroid medication for her thyroid condition and makes the decision to start to take an immunosuppressant for her arthritis.

By using data to extract patients and trends, we can begin to more accurately map out patient journeys such as this. However, it’s not just the patients who benefit (although this is clearly the ultimately goal for all involved). There are clear benefits to the medical advertising industry as well. AI is the new frontier for the creative/marketing industry to create more time in order to use their existing expertise to create copy, artwork, and campaigns. 

The challenges of prediction for medical advertisers

While the benefits of prediction are becoming much clearer to see, there are also several challenges that need to be addressed along the way.  

For prediction to be accurate it is important that the data being used is gathered from a wide range of sources. Social listening is just one piece of this data jigsaw. Data from health insurers and other third-party sources will be key to building a critical mass of data from which we can structure and extract meaning through the use of our AI/ML modeling to make accurate predictions. 

Regulation of influencer populations is also set to make prediction harder. Influencers and patient advocates in the medical space are, rightly, held to a higher standard. With new regulations incoming there may be limits on what they are able to share or say. These influencers generate online discussions from which we are able to gather data and without them we will lose a valuable source of data.

Finally, clinical trials will also need to change if the data gleaned from them are to accurately reflect the population. Increasing the number and diversity of patients involved in trials, as well as the locations, will almost certainly help us to develop a more representative understanding of how patients from different backgrounds use medicines. 

Next steps for prediction

While medical advertisers will undoubtedly need to grapple with these challenges, the benefits of prediction are too enticing to ignore. Using AI to automate the effectiveness of patient campaigns has the potential to save industry millions in misspent marketing budgets. More importantly, it can also deliver the very purpose of what medical advertising sets out to achieve – from increasing uptake of medicines and a better quality of life for patients to improving patient health outcomes.  

Elizabeth Fairley, Ph.D. is founder, chief operating and data officer, Talking Medicines.