Foresight -- Deep Generative Modelling of Patient Timelines using Electronic Health Records

AI-generated keywords: Healthcare Electronic Health Records (EHRs) Temporal Modelling Foresight Biomedical Concept Modelling

AI-generated Key Points

  • Electronic Health Records (EHRs) store detailed longitudinal information about patients' health status and clinical history.
  • Temporal modelling of medical histories is a powerful tool for forecasting future events, estimating risks, suggesting alternative diagnoses, and predicting complications.
  • Foresight is a groundbreaking solution that processes the entire free-text section of EHRs for longitudinal modelling using NER+L tools like MedCAT.
  • Foresight provides probabilistic forecasts for future medical events such as disorders, medications, symptoms, and interventions with impressive precision rates.
  • Foresight offers a granular view of patients by analyzing textual data while introducing minimal additional noise.
  • The model has been rigorously tested at prominent healthcare institutions and demonstrated high relevancy in predicting candidate disorders.
  • Foresight is easy to train and deploy locally as it only requires free-text data at a minimum.
  • It can simulate follow-on disorders, medications, and interventions over multiple steps as needed, making it versatile for various biomedical concept modelling scenarios.
  • Funding from prestigious organizations like the NHS AI Lab, National Institutes of Health Research, and Health Data Research UK supported the development of Foresight.
  • The availability of the user-friendly Foresight web app with interactive features is set to revolutionize healthcare analytics and decision-making processes worldwide.
Also access our AI generated: Comprehensive summary, Lay summary, Blog-like article; or ask questions about this paper to our AI assistant.

Authors: Zeljko Kraljevic, Dan Bean, Anthony Shek, Rebecca Bendayan, Joshua Au Yeung, Alexander Deng, Alfie Baston, Jack Ross, Esther Idowu, James T Teo, Richard J Dobson

License: CC BY 4.0

Abstract: Electronic Health Records (EHRs) hold detailed longitudinal information about each patient's health status and general clinical history, a large portion of which is stored within the unstructured text. Temporal modelling of this medical history, which considers the sequence of events, can be used to forecast and simulate future events, estimate risk, suggest alternative diagnoses or forecast complications. While most prediction approaches use mainly structured data or a subset of single-domain forecasts and outcomes, we processed the entire free-text portion of EHRs for longitudinal modelling. We present Foresight, a novel GPT3-based pipeline that uses NER+L tools (i.e. MedCAT) to convert document text into structured, coded concepts, followed by providing probabilistic forecasts for future medical events such as disorders, medications, symptoms and interventions. Since large portions of EHR data are in text form, such an approach benefits from a granular and detailed view of a patient while introducing modest additional noise. On tests in two large UK hospitals (King's College Hospital, South London and Maudsley) and the US MIMIC-III dataset precision@10 of 0.80, 0.81 and 0.91 was achieved for forecasting the next biomedical concept. Foresight was also validated on 34 synthetic patient timelines by 5 clinicians and achieved relevancy of 97% for the top forecasted candidate disorder. Foresight can be easily trained and deployed locally as it only requires free-text data (as a minimum). As a generative model, it can simulate follow-on disorders, medications and interventions for as many steps as required. Foresight is a general-purpose model for biomedical concept modelling that can be used for real-world risk estimation, virtual trials and clinical research to study the progression of diseases, simulate interventions and counterfactuals, and for educational purposes.

Submitted to arXiv on 13 Dec. 2022

Ask questions about this paper to our AI assistant

You can also chat with multiple papers at once here.

AI assistant instructions?

Results of the summarizing process for the arXiv paper: 2212.08072v1

In the realm of healthcare, Electronic Health Records (EHRs) play a crucial role in storing detailed longitudinal information about patients' health status and clinical history. A significant portion of this data is often captured within unstructured text, making it challenging to extract valuable insights. To address this issue, temporal modelling of medical histories has emerged as a powerful tool for forecasting future events, estimating risks, suggesting alternative diagnoses, and predicting complications. Traditionally, prediction approaches have relied on structured data or limited single-domain forecasts and outcomes. However, a groundbreaking solution called Foresight has been developed to process the entire free-text section of EHRs for longitudinal modelling. This innovative GPT3-based pipeline leverages NER+L tools like MedCAT to convert document text into structured, coded concepts. By doing so, Foresight can provide probabilistic forecasts for future medical events such as disorders, medications, symptoms, and interventions. The beauty of Foresight lies in its ability to offer a granular and detailed view of patients by analyzing textual data while introducing minimal additional noise. Through rigorous testing at prominent healthcare institutions like King's College Hospital in South London and Maudsley as well as the US MIMIC-III dataset, Foresight has demonstrated impressive precision rates for forecasting biomedical concepts. Furthermore, validation on synthetic patient timelines by clinicians showcased the model's high relevancy in predicting candidate disorders. One of the key advantages of Foresight is its ease of training and deployment locally since it only requires free-text data at a minimum. As a generative model, it can simulate follow-on disorders, medications,and interventions over multiple steps as needed. This versatility positions Foresight as a general-purpose tool for biomedical concept modelling that can be applied across various scenarios such as risk estimation in real-world settings, conducting virtual trials for clinical research purposes, studying disease progression dynamics through simulations of interventions and counterfactuals. The development of Foresight was made possible through funding from prestigious organizations like the NHS AI Lab, National Institutes of Health Research, and Health Data Research UK. Additionally, infrastructure support from King's College London and other research centers played a pivotal role in bringing this cutting-edge technology to fruition. With the availability of the user-friendly Foresight web app offering interactive features for evaluating model outputs and understanding forecasted concepts in depth using gradient-based saliency methods; this groundbreaking tool is poised to revolutionize healthcare analytics and decision-making processes worldwide.
Created on 16 Feb. 2026

Assess the quality of the AI-generated content by voting

Score: 0

Why do we need votes?

Votes are used to determine whether we need to re-run our summarizing tools. If the count reaches -10, our tools can be restarted.

Similar papers summarized with our AI tools

Navigate through even more similar papers through a

tree representation

Look for similar papers (in beta version)

By clicking on the button above, our algorithm will scan all papers in our database to find the closest based on the contents of the full papers and not just on metadata. Please note that it only works for papers that we have generated summaries for and you can rerun it from time to time to get a more accurate result while our database grows.

Disclaimer: The AI-based summarization tool and virtual assistant provided on this website may not always provide accurate and complete summaries or responses. We encourage you to carefully review and evaluate the generated content to ensure its quality and relevance to your needs.