A patient preference study of Crohn’s disease treatments, carried out by Semalytix and partner Boehringer Ingelheim, has won a European Pharmaceutical Market Researchers Association (EphMRA) market research excellence award. The study was carried out together with Ben Collins (Semantic Analytics Manager, BI) and Wael Ayad (Sr. International Customer Insights Manager, BI).
The prize, featured in the “Future Leaders” category, recognizes and celebrates Semalytix’s AI-powered market research approach as a promising new tool for pharmaceutical companies to better understand the treatment burden facing patients and their families.
The goal of the study published at ISPOR was to leverage automated analysis of social media as a basis to identify how children with Crohn’s disease and, critically, their parents, perceive the relative benefits and drawbacks of different treatment modes of administration: IV infusion at the hospital, versus a sub-cutaneous injection-pen or syringe used at home.
Crohn’s disease – a chronic inflammatory disorder that mainly affects the bowel – presents significant challenges for patients and their families, especially in severe cases requiring injectable therapies called TNF-alpha blockers. This study, which analyzed thousands of online narratives, was uniquely powered to generate insights about parents’ experiences supporting their children through treatment, as well as those of the children themselves.
The study uncovered a preference among some children and their parents for hospital-administered IV infusions, as these were generally less painful, more appropriately dosed, and less frequent than the at-home injections used by a majority of adult patients. For many parents, nurse-administered treatment in a clinic at a scheduled time every three months removed the burden of persuading a reluctant child to undergo treatment at home on a weekly or three-weekly basis. It also offered more flexibility to travel in between the quarterly infusions.
Capturing the impact of a particular disease not just on individual patient outcomes but on their broader family and social life is critically important in understanding treatment burden and, ultimately, in meeting unmet need. This broader focus – centered on the patient, but not limited to the patient – helps differentiate Semalytix’s and Boehringer Ingelheim’s study from traditional market research methods.
It also underlines the power of harnessing social media and online health-focused discussion groups as a source of patient insight -- not only during the unique circumstances of a pandemic lock-down, when focus groups are impossible, but also well beyond.
The study partners used Semalytix’ Pharosâ platform to identify and analyze over 3000 online patient narratives around preferred mode of administration of Crohn’s disease treatments. These highly relevant stories were singled out from an initial search across many millions of international websites and blogs; the study data set included information from a forum dedicated specifically to parents with children suffering from Crohn’s.
Semalytix’ approach combines top-down data-driven concept elicitation– searching for pre-specified ideas or concepts within the data – with a process that captures hidden, or un-prompted factors or concerns from that same data, revealing insights beyond the study sponsors’ preconceived notions.
The research applied trained algorithmic models, alongside human experts, to single out reliable and relevant text or comments relating to known categories like “unmet need”, “treatment burden,” or “quality of life,” alongside more emotive concepts like “wishes”, “requests” or “coping”. Data quality control involved continuously sifting for relevance, removing as much irrelevant data – “noise” – as possible.
To elicit hidden, un-prompted concepts or feelings – the unknown unknowns – Semalytix used a proprietary Knowledge Graph comprising millions of pharma-relevant terms and entities, including symptoms, treatment as well as quality of life concepts, benefit-risk statements and demographic variables. Both the targeted queries, and the bottom-up concept elicitation, capture and control for the many nuances in how language is used around particular topics or ideas.
Analyzing the thoughts and conversations of thousands of individual patients and families will not often generate clear-cut conclusions. Each circumstance is unique; each treatment formulation offers pros and cons whose relative weighting will vary among different patient groups. Nor is any market research method free of bias – dissatisfied customers are more likely to express their views online than are happy ones, for instance. But accessing unsolicited patient narratives around a particular formulation – both positive and negative – provides unparalleled insight into real-world disease burden and treatment experience.
By combining artificial intelligence with human expertise and interpretation capabilities, Semalytix and Boehringer Ingelheim were able to listen to the authentic voices of those living with Crohn’s disease. This provided a picture of the impact of the condition on the lives of patients and their families that is more comprehensive and more accurate than that offered by most traditional market research – even if it is not perfect.