What’s Wrong With Clinical Trial Management & What Can NLP Tools Do To Speed Read Data
March 31, 2021, 2 Minute Read
Recently in the news, we have seen so much information about Clinical Trials due to COVID-19; however, the news forgets to highlight that Clinical Trials are a big responsibility. Patients are given experimental drugs, procedures, or treatments. Patients put faith and trust in their providers and trial sponsors, who have created the experimental products and protocols. We have seen straightforward data reporting, but recently there have been Public Relations blunders and broad media misinformation about what is contained in trial data packages. We trust in the accuracy of data, protocol designers, and patient relationships with their doctors. We believe that whenever patients place their lives into a trial, the team is collecting useful data to save lives and to advance scientific frontiers.
At the conclusion of a trial, I am always humbled by the team that designed, carried out, and cared for patients in the trial. I am thankful for patients who volunteered their lives so that new products could be deemed safe and efficacious. So if you are in the clinical trials industry, ask yourself: “Did I use all available tools at my discretion to conduct this trial to the best of our ability?”. Several examples of firms who sponsor trials – e.g. Pfizer, Moderna, Johnson & Johnson, Astrazeneca to name several – write protocols that specify who is able to participate in a trial and what data to collect. The providers who conduct trials at their facilities or “sites” – e.g. HCA Hospitals, NYU Langone, Rochester Regional Health – must be able to complete the trial protocol, while also caring for the patients, noting their circumstances in their records.
Recently, Lymba used its NLP Platform technology to create a Medical Playbook demonstration. The goal of the demonstration concept is to support patient record and trial analytics, extract severe adverse event information, and collate information during clinical trial severe adverse event adjudication. In this demonstration, we show that:
We can extract information from patient medical records such as Histories and Physicals using a common medical Ontology (see my previous blog post!).
Then we show that using definitions for clinical trial Severe Adverse Events, we can extract information related to clinical trial study results from clinicaltrials.gov.
We then make data available to help determine if patients are experiencing drug reactions, surfacing past medical conditions, or actual severe adverse events caused by the experiment.
Would you find this service and information useful for conducting your clinical trials in the future? Is this a use case you would consider for modernizing clinical trials with more information visibility, Natural Language Processing, and machine learning capabilities?
Contact us to continue the discussion and step into our Medical Playbook demonstration!