The full title of this project is "Improving Prediction Models for Diagnosis and Prognosis of COVID-19 and Sepsis with Natural Language Processing of Clinical Text". The goal of the project is to develop prediction models for diagnosis and prognosis of infectious diseases based on data in electronic health records. The primary focus is on developing multimodal models that incorporate information from clinical text, in addition to structured data.

We use machine learning and natural language processing to develop multimodal clinical prediction models for diagnosis and prognosis of infectious diseases, focusing in particular on COVID-19 and sepsis. Prediction models that are able to detect a disease early (early prediction) or provide disease progression prognoses (outcome prediction) are critical in improving patient management by facilitating early and appropriate interventions, as well as effective allocation of healthcare resources.

An important component of the multimodal prediction models will be the use of pre-trained language models specifically developed for, or adapted to, the clinical domain. By leveraging clinical language models, fine-tuned to perform a particular downstream task (such as early prediction of sepsis), for creating representations of clinical text, while also incorporating structured clinical data, we aim to develop better prediction models.