Researchers have developed an AI tool called Foresight that can create digital twins of patients and predict future health outcomes. Utilizing generative pre-trained transformers, similar to ChatGPT, the tool was trained on medical records and then used to generate virtual duplicates of patients using fresh healthcare data. These digital twins can forecast various health outcomes, including disease development and medication needs.
The tool showed promising results in the US, accurately identifying patients’ next conditions with 88% accuracy. However, its performance was slightly lower with UK data, correctly predicting subsequent conditions in NHS patients 68% to 76% of the time. Despite this, researchers are optimistic about the potential of digital twins in healthcare.
The research team believes that these digital twins can revolutionize diagnoses, treatment plans, and clinical research. With forecasts described as “possible multiverses” for understanding diseases, the tool can provide insights and scenarios for different health conditions and interventions.
Further improvements and testing are needed before widespread implementation of the technology. The team is already working on an enhanced version called Foresight 2. For more details on the study, the findings are available in Lancet Digital Health.
HEALTH
WORLD AFFAIRS
TECHNOLOGY
CREATIVE
