Patient Recruitment For Clinical Trials
Challenge
The hospital's existing system was ill-equipped to sift through the unstructured data in patient records, making it difficult and time-consuming to find suitable candidates for clinical trials. Manually reviewing each record was not only resource-intensive but also posed a risk of overlooking potential candidates.
Solution
We took a two-pronged approach to tackle the issue:
- LLM for Data Extraction: A Large Language Model (LLM) was trained on medical and clinical datasets to specialize in reading unstructured doctor's notes. This allowed the model to identify and pull out diagnoses with high accuracy.
- Automated Recruitment Pipeline: The extracted diagnoses were then fed into an automated recruitment system, which matched the patients to appropriate clinical trials based on a set of predefined criteria.
Result
The implementation of the LLM and automated recruitment pipeline revolutionized the client's patient recruitment process. The system could now quickly and accurately identify candidates for various clinical trials, significantly reducing manual review time and increasing the pool of eligible patients. This not only accelerated the pace of ongoing research but also opened the door for more diverse and extensive clinical studies in the future.
Your AI-OPs Partner.
