Salt Lake City – For a long time, an objective of medicine has been better to better understand the long trajectories of diseases in the hope of engaging in better prevention and early intervention.
“Collectively, these are (chronic and progressive diseases) responsible for approximately 90% of health care costs in this country and the vast majority of morbidity and mortality,” said Nina de Lacy, professor of psychiatry and member of the executive committee of the AI initiative responsible for AI.
Now, researchers at the University of Utah have taken a crucial step to do so, by revealing a new open source software tool case that uses artificial intelligence to predict whether individuals will develop progressive and chronic diseases of years before the symptoms appear.
Enter Riskpath, a new technology that analyzes models in the health data collected over several years to identify people at risk with “unprecedented precision” of 85% to 99%, according to the National Institute of Mental Health, research sponsored by mental health published last week by the Department of American Psychiatry and Huntal Mental Health Institute.
The program exploits the explainable AI, which is designed to explain complex decisions in a way that humans can understand.
“The explanation means, can I explain enough about how AI has accomplished this prediction so that it becomes understandable for humans?” Said of Lacy. “It would be things like what Riskpath does.”
De Lacy explained something that has always been a challenge in biomedicine is to build models and analyze longitudinal data, which means that it is collected over many periods.
“One of the main cases of use in the use of longitudinal data is the development of courses, understanding how people grow and develop over time,” said Lacy. “And one of the others is what Riskpath is intended, which includes progressive or chronic diseases. There are many progressive and chronic diseases, and some of the big ones are things that are the main diseases that affect humans.”
Research shows that current medical prediction systems for longitudinal data often lack the brand, correctly identifying patients at risk only about half in three quarters of the time. Unlike existing prediction systems for longitudinal data, Riskpath uses AI algorithms of advanced chronological series which provide crucial information on how risk factors interact and change importance throughout the pathological process.
“By identifying high-risk individuals before the symptoms appear or at the start of the evolution of the disease and the risk factors have the most at different stages of living, we can develop more targeted and effective preventive strategies. Preventive health care may be the most important aspect of health care at the moment, rather than treating problems after their materialization,” said Lacy.
De Lacy and the rest of the research team validated Riskpath in three large cohorts of long -term patients involving thousands of participants to successfully predict eight different conditions, including depression, anxiety, ADHD, hypertension and metabolic syndrome.
Technology offers several key advantages:
- Increased understanding of the progression of the disease: Riskpath can map how different risk factors change importance over time, revealing critical windows for the intervention. For example, the study has shown how screen time and executive function become increasingly important risk contributors for ADHD as adolescence approaches.
- Rationalized risk assessment: Although Riskpath can analyze hundreds of health variables, researchers have found that most conditions can be predicted with similar precision using only 10 key factors, which makes the implementation more feasible in clinical environment.
- Practical risk visualization: The system provides intuitive visualizations showing what time periods in a person’s life contribute the most at the risk of illness, helping researchers identify optimal times for preventive interventions.
Although Riskpath is mainly a research tool to help researchers create better risk stratification models, Lacy hopes that it will eventually be used in a health care establishment to improve disease management.
“Some may use it to create models that can be implemented in health care, and we hope in a way they would do it. But … a large part of what my laboratory is interested is to create tools that do a better risk stratification job. We are very interested in prevention,” said Lacy. “The ultimate goal of Riskpath and tools like Riskpath is to help people create better risk stratification tools and decision -making tools.
“And what these do is help clinicians, and perhaps one day, patients can understand their risk of chronic or progressive illness better and earlier,” she said.
