Natural Language Processing (NLP) to Optimize Clinical Trials: I2E Hackathon at the Linguamatics Text Mining Summit - Using text mining to address healthcare information challenges
(Cambridge, UK & Boston, USA – 8 November 2016) The tenth Linguamatics Text Mining Summit took place on 17-19 October 2016 at the Chatham Bars Inn, MA. The Summit was attended by leaders in the Life Science and Healthcare Industries, and provided an opportunity to share use cases and best practices on using Natural Language Processing (NLP) for text analytics to address critical information challenges. Speakers included Agios Pharmaceuticals, AstraZeneca, Atrius Health, Drexel University College of Medicine, GlaxoSmithKline, Kaiser Permanente, Lilly, National Cancer Institute (NCI), Novo Nordisk, and the OHSU Knight Cancer Institute.
Top pharma organizations and healthcare providers have a joint interest in using NLP to improve clinical trial design, such as site selection and patient recruitment. Identification of patient populations is difficult to scale due to complex eligibility criteria and the need to scan through large numbers of unstructured medical records, and manual chart review is a long and laborious process. Many trials suffer from slow accrual that can lead to missed recruitment targets.
For the second year, the Summit featured a text mining Hackathon. This year, the Hackathon sought to mirror the problems around clinical trials recruitment in a two-part challenge. The first part was to find typical inclusion and exclusion criteria for a diabetes-related trial, such as gender, age, body mass index and blood pressure. The data used was semi-structured clinical trial descriptions available from ClinicalTrials.gov. The second part was to mine a set of anonymized medical records for patient matches to trial criteria (such as age greater than 18), to identify potential trial subjects to ensure successful recruitment.
The Hackathon was judged according to the success in correctly identifying the most common upper and lower limits for the trial criteria, along with the precision and recall achieved in identifying potential trial subjects. Honors went to Pentavere Research Group, a Canadian based health insight company, for the second year in a row.
Jason Evans, co-founder and President of Pentavere Research Group, commented, “We are humbled to have won this extremely difficult competition. Our company’s purpose is to transform healthcare data into healing wisdom to improve health care outcomes for all. This competition validates our approach. It is wonderful to see how far we have all advanced our capabilities to un-tap data’s potential to improve lives.”
David Milward, Linguamatics co-founder and CTO, said “The Hackathon showed how NLP can be successfully used both in trial design and in patient recruitment. The speed and accuracy with which participants were able to address the challenge in this half-day event illustrates how the I2E text mining platform can cut costs and improve patient outcomes in healthcare.”
Linguamatics is the world leader in deploying innovative natural language processing (NLP)-based text mining for high-value knowledge discovery and decision support. Linguamatics I2E is used by top commercial, academic and government organizations, including 17 of the top 20 global pharmaceutical companies, the US Food and Drug Administration (FDA) and leading US healthcare organizations. I2E can be used to mine a wide variety of text resources, such as scientific literature, patents, Electronic Health Records (EHRs), clinical trials data, news feeds, social media and proprietary content. I2E can be deployed as an in-house enterprise system, or as Software-as-a-Service (SaaS) on the cloud.
Pentavere Research Group is a Canadian-based company whose mission is to leverage insights from real world evidence in order to improve healthcare outcomes for all. Pentavere’s technology platform daRWEn ™ contains a vast and growing repository of insights from the primary care clinical setting, structured through the integration of natural language processing on a big data platform. daRWEn ™ allows customers to combine Pentavere’s real world evidence with their own data assets to visualize insights into population health.