Social media listening in R&D

Bringing social media listening into R&D

After her daughter was diagnosed with Prader-Willi syndrome, Maria Picone used social media listening and AI analytics to find out the best ways to manage the rare disease. Now the insights her company is gathering are helping pharma and regulators fulfil unmet needs in difficult conditions. We spoke to her to find out more.

As pharma comes to understand the importance of real world data, it has started looking for insights in all kinds of places it would never have considered before – especially in rare diseases, when information is usually scarce.

Social media listening company TREND Community thinks that some of the most valuable insights into rare disease can be found in private social media groups such as those on Facebook.

“The conversations that are happening in private groups for rare diseases are often much more informative than in public social media spaces,” says Maria Picone, TREND’s co-founder and CEO. “A group of people who have been talking together about their experiences for five to ten years can tell us a lot about a disease even when there is very little published research.”

TREND partners with these communities to gather insights using qualitative research and AI algorithms, identifying unmet needs amongst patients and caregivers.

“We give them the tools that they need to explain to their community members who we are and what we want to do,” Picone says. “We make sure that we’re sensitive to issues around privacy and security – when we’re working with a group on a public forum like Facebook, everything we do is anonymised.”

She says that the information they are gathering can be useful for patients and caregivers looking to understand how to manage a rare disease, and has also garnered interest from pharma.

Hope for rare diseases

Picone and her husband had worked in data analytics and digital health for several years before their daughter was born with the rare disease Prader-Willi syndrome in 2012.

“When you have a child with a rare disease that you know nothing about, life just freezes,” she says.

“Fortunately, because of my experience in clinical research and our connections, we had the resources to dig into the literature, connect with leading specialists, and understand the standard of care.

“Even so, it wasn’t until I joined a private group on social media of other caregivers that I understood how much hope there was for a child born today with Prader-Willi syndrome. That was when we realised there was a huge gap between what our doctors knew and what was actually happening in these patient communities.”

At that point, Picone and her husband realised that their daughter could have a much better life than they had previously thought possible.

“But we had to figure out how to give that to her,” she says. “The original idea for TREND Community came about because we wanted to take the technology that we had been building for our clients – like clinical trial tools – and share them with our new patient community.”

The initial goal was simply to understand what Prader-Willi patients and caregivers were doing, what was working, and why some things worked for some people and not for others. This was done by collecting anecdotal experiences, such as how caregivers managed diet, what drugs they were trying off label, and what outcomes they were seeing.

“It wasn’t a startup idea,” Picone says. “We weren’t getting paid. We really just thought that we could use this technology and this approach to collect data that could help us figure out how to treat our daughter.

“It was really successful. Our daughter and our family have benefited directly from the work that we’ve done. For example, we put her on a ketogenic diet, which has helped her tremendously in terms of cognitive development.”

Along the way, the team learnt some important lessons about when and how people are most likely to provide data.

“We learned that people were more likely to share data if they felt like there was an immediate benefit to them – not just participating for the good of someone else in the future.

“We experimented with asking questions in different ways, and even boiling it down to just five questions they had to answer once a week. People still struggled with that, but we noticed that they would nevertheless write everything about their week in a journal or in private groups. Everything we wanted to know, they were sharing – they were just doing it organically in private discussions.”

In 2016 the US passed the 21st Century Cures Act, through which the FDA began to embrace the idea of bringing patient perspectives into drug development – through mechanisms such as drug development meetings where patients and caregivers are invited to take part.

The TREND team therefore thought that they could develop the technology to capture information about patient experiences from social media conversations. In 2017, TREND Community was born.

“We want to make it possible to gather more perspectives from patient communities, and to validate what patients are saying one-on-one with industry and with FDA,” says Picone.

“The FDA can use our data to demonstrate that there is an unmet need in the community, and it is a problem that patients and caregivers are talking about. On the pharma side, there are opportunities to use the data to identify new indications.”

Bringing the patient perspective to R&D

One example of TREND providing insights for an FDA drug development meeting was with the Friedreich’s ataxia community.

“In advance of the meeting, they asked us if we could analyse their private social media group to generate some evidence for the FDA,” Picone says. “The FDA was interested in learning which symptoms affect patients and caregivers the most, what treatments they’ve tried, and what their unmet needs are.”

These kinds of questions are where TREND focuses their AI analytics algorithms, and where Picone believes the most useful insights can be gathered.

“We’re looking at what people are talking about around disease burden and disease management. We then identify the other concepts that are trending and see if there’s an unmet need that we can identify. We started out using an existing database called the Unified Medical Language System to guide our engine, and continue to train it in that way, developing a taxonomy that understands and defines which terms fall into those three buckets.”

She gives an example from their work in Prader-Willi syndrome: “Prader-Willi is most commonly associated with two major things – hyperphagia, which is chronic insatiable hunger, and obesity. Most clinical trials were focused on finding treatments for those two symptoms. In fact, for many clinical trials, hyperphagia was an inclusion criteria.

“My daughter, however, doesn’t have hyperphagia. That doesn’t mean that she won’t ever have it, but there’s a misunderstanding that every person with Prader-Willi syndrome is in hyperphagia. That means she wouldn’t qualify for most of these trials.”

Through social media listening on private groups in Prader-Willi syndrome, the team found that issues around sleep trended highly – even more so than conversations around hunger and obesity.

“That tells us that this may be an issue that a lot of caregivers are dealing with. At that point, our engine was looking for phrases like, ‘I’m tired’, ‘I’m sleepy’, ‘I’m fatigued’. It’s smart enough to know that there are a bunch of terms that can be grouped together that are all related to issues of sleep.”

TREND then partnered with paediatric neurologists Daniel Glaze and Ameeben Patel at Texas Children’s Hospital/Baylor College of Medicine, who helped describe what disordered sleep looks like from a clinical perspective in this patient population.

“We were then able to create a codebook and use that to train our engine to look for things beyond ‘I’m tired’ – such as head bobbing, imbalance issues, narcolepsy or cataplexy. Then we went back out to social media and looked for conversations where people were talking around this new definition of disordered sleep. We used that as a way to quantify the problem in the community.

“We also helped community members access a drug called pitolisant that was recently approved in Europe for the treatment of narcolepsy through FDA’s personal importation programme. Scientist Lara Pullen identified pitolisant as a possible treatment for her son who was experiencing episodes of debilitating cataplexy. We asked caregivers to track their children’s experiences while taking this experimental treatment on our data collection platform.”

This work has led to a pharma company that acquired the US rights to pitolisant, Harmony Biosciences, connecting with TREND and pursuing an indication for Prader-Willi syndrome.

“In this example the drug company already had a treatment for narcolepsy, pitolisant, that they thought could help in Prader-Willi syndrome, but there was a lack of evidence and literature on sleep issues.

“I requested an informal meeting with the FDA and asked them how our data could support or accelerate the approval process. They suggested doing a PK study on the children we had identified to have a baseline in Prader-Willi syndrome and better understand the safety. I brought this idea back to the sponsor and they proposed it in their application to the FDA, the FDA approved the protocol and they opened their investigational new drug application with the study.”

Picone says that TREND would like to publish as much as possible with the communities on the insights gleaned from their shared stories and experiences, so that they can become part of the scientific literature.

“In that way, doctors and researchers have access to the information and can consider it when trying to identify therapeutic targets or choose their next community to focus on.

“Our goal is to really understand what is happening right now in these patient communities and what their most pressing unmet needs are. Then we can use social insights to empower the community members to drive forward initiatives that are important to them.”

Maria Picone

Maria Picone is the co-founder and CEO of TREND Community. After more than ten years managing operational and strategic ventures in the biotech and pharmaceutical industries, Maria and her husband, Christopher DeFelice, started a consulting company that built custom digital health platforms for biotech/pharma, research organisations and health systems, and provided strategic and publishing services for FDA regulatory submissions. In 2012, their daughter was born with Prader-Willi Syndrome; their family’s journey was the inspiration for TREND Community.

George

George Underwood is a senior member of the pharmaphorum editorial team, having previously worked at PharmaTimes and prior to this at Pharmafocus. He is a trained journalist, with a degree from Bournemouth University and current specialisms that include R&D, digital and M&A.

Don’t miss your complimentary subscription to Deep Dive and our newsletter

Bringing social media listening into R&D

Hope for rare diseases

Bringing the patient perspective to R&D

About the interviewee

About the author