Jakarta, INTI – Amid the ongoing challenge in the medical world where patient deterioration is often detected too late, a breakthrough was announced on Monday, November 17, 2025, at the University of Melbourne, Australia, offering new hope. Researchers introduced an AI-based “digital twin” technology claimed to predict individual health changes more quickly and accurately. This innovation emerged as a solution to the frequent issue of critical cases being identified only after a patient’s condition has already worsened. By leveraging generative artificial intelligence, the tool named DT-GPT drew public attention for its ability to replicate a patient's health profile in a virtual form. The findings were presented during a research briefing that attracted interest from academics and healthcare practitioners.
A New Breakthrough in Health Data Analysis
The study led by the University of Melbourne relied on three large datasets containing thousands of electronic medical records. These datasets were used to train a large language model capable of reading patient health patterns in greater detail. DT-GPT was then tested on patients with various conditions, including Alzheimer’s disease, non-small cell lung cancer (NSCLC), and those receiving treatment in intensive care units (ICUs).
By analyzing medical histories such as laboratory results, diagnoses, and treatment records the model can generate a patient’s “digital twin” that presents projections of health changes over time.
Outperforming 14 Advanced ML Models
One distinctive aspect of the research is that the model was not provided with the patients’ eventual health outcomes. This approach ensured that prediction validation could be conducted objectively. The results were striking: DT-GPT proved more accurate than 14 other state-of-the-art machine learning models.
This advantage has led many to recognize DT-GPT as a potential “gamechanger” in clinical trials and data-driven medical decision-making.
Shifting from Reactive to Predictive Medicine
Lead researcher Michael Menden stated that this technology has the potential to shift medical practice from reactive treatment to predictive and personalized care. With its ability to detect signs of patient deterioration earlier, physicians can intervene before conditions worsen.
Moreover, the model can also predict drug side effects, assist doctors in tailoring therapies to each patient’s unique profile, and ultimately improve treatment outcomes.
Chatbot-Like Interface and Zero-Shot Predictions
Another advantage of DT-GPT is its ability to interpret complex data quickly through a chatbot-like interface that allows medical professionals to explore AI predictions with ease. The technology employs a zero-shot prediction approach, meaning it can generate forecasts based on laboratory values even without prior specific training.
Conclusion
The “digital twin” developed through DT-GPT marks a major step toward a future of faster, more accurate, and more personalized medical care. With its capability to assess patient conditions predictively, this technology holds tremendous potential to enhance global healthcare standards and advance the integration of AI into clinical processes.
Read More:China Utilizes AI to Reduce Healthcare Service Gaps at the Community Level