Collaborative AI Project Aims to Revolutionise Health Data Integration
Chicago, IL – In a transformative endeavour set to redefine the landscape of health care, the University of Illinois Chicago (UIC) is spearheading an interdisciplinary project, utilising artificial intelligence (AI) to integrate data from a multitude of health care professions. This project aims to create comprehensive datasets that could significantly enhance patient outcomes.
The initiative involves collaboration with the University of Iowa, University of Missouri, and Loyola University, with technical partnerships from Microsoft and Tackle AI. Funded by up to $10 million from the federal Advanced Research Projects Agency for Health (ARPA-H), this project marks UIC's inaugural receipt of ARPA-H funding.
Enhancing Health Records with Diverse Data
The project's core objective is to devise novel methodologies for amalgamating structured data and free-text notes from various health professionals including nurses, physical and occupational therapists, speech and language pathologists, and physicians, facilitating more effective usage within electronic health records (EHRs). These notes often furnish crucial, additional insights into a patient’s progress, especially as their care extends beyond hospital settings.
The project will primarily address two intricate patient populations: individuals recovering from fall-related injuries and infants transitioning from neonatal intensive care units (NICU) to home environments. Both groups are heavily dependent on multidisciplinary health care services.
Holistic Approach to Health Data
Professor Andrew Boyd, one of the principal investigators and a professor of biomedical and health information sciences at UIC, emphasised the interdisciplinary nature of health care, lamenting that existing data tools and infrastructure often overlook contributions from many health team members. He pointed out that data from professions such as nursing and therapy offer detailed, continuous insights rather than mere episodic snapshots.
The innovative integration of this diverse data will enable the creation of all-encompassing care summaries and the development of robust AI applications. Additionally, this unified data will pave the way for new scientific discoveries, aimed at enriching patient care and treatment protocols.
Improving Communication and Outcomes
Catherine K. Craven, another principal investigator and a biomedical informatician at the University of Missouri School of Medicine, highlighted the communication barriers posed by fragmented documentation. By consolidating data from different health care providers, the project aims to enhance communication and generate groundbreaking scientific insights to better patient outcomes.
Karen Dunn Lopez, professor of nursing at the University of Iowa, further noted that addressing the complexities of falls and NICU transitions could result in solutions applicable to less complex health issues. This project reinforces the idea of patient-centred decision-making, driven by comprehensive multidisciplinary data.
AI's Promise for Complex Cases
AI's potential in health care lies in its capacity to automatically extract meaningful insights from EHRs. Including observations from nurses and therapists in patient data can lead to more precise predictions regarding risks and treatment effectiveness. For instance, adult fall injuries, which are notoriously difficult to prevent, could be better managed through detailed AI-driven analysis of multidisciplinary data.
Physical and occupational therapists’ detailed reports on aspects like strength and balance are crucial yet traditionally challenging to integrate due to their subjective nature. Unifying this information with other data sources could help identify the causes of falls and tailor appropriate interventions.
Advancing Patient Involvement
Professor Mary Khetani of UIC's occupational therapy and rehabilitation sciences department, highlighted the importance of involving patients in health care decisions. By organising narrative notes from therapists, the project aims to create a continuum of care that keeps patients and their families well-informed as they navigate post-hospitalisation care.
Overcoming Technical and Linguistic Barriers
The project will leverage advanced text-mining and language-processing tools to bridge the technical gaps that hinder current data integration from various health disciplines. Attention will be given to developing AI tools capable of accurately interpreting the specialised jargon inherent in medical data.
Once integrated, data from diverse health professionals will contribute to the training of more nuanced predictive models for health risks and treatment efficacy. AI tools will also assist in generating concise, accessible summaries of extensive text and data, making it easier for both patients and primary care providers to manage and understand complex health information.
Collaborative and Open-Source Innovation
The team plans to engage data scientists and software developers through hackathons and other activities using deidentified data to develop additional clinical and research applications. All tools created through the project will be open-source, incorporating feedback from health domain experts.
With a strong foundation of multidisciplinary health sciences and a team skilled in machine learning, natural language processing, and data science, UIC is poised to lead this groundbreaking project. Key contributors from UIC include Samantha Bond from the College of Applied Health Sciences, Miiri Kotche from the College of Engineering, and David Chestek from the College of Medicine.
“This initiative capitalises on UIC’s diverse expertise,” Boyd said. “Such collaborative opportunities enable us to transform health care data management, benefiting both our institutions and the broader health care system.”
As the project progresses, it holds the promise of not only improving health care delivery for falls and NICU patient transitions but also setting a precedent for the integration of diverse health care data to enhance patient care across various medical fields.
Source: Noah Wire Services