Recent advances in AI offer tremendous potential for enhancing patient healthcare outcomes and controlling costs. Our goal is to advance the use of AI technologies to yield new breakthroughs in fundamental research into diseases, and apply them across the care continuum for diagnosis, prognosis, treatment planning, patient monitoring, and personalized care.
Teams of UB researchers from across many disciplines are using Machine Learning to discover predictive patterns in large data sets of genetic information, high resolution medical images, and medical records which model changes in cellular, tissue, organ, and organism systems to help improve diagnosis and inform patient outcomes. UB researchers have created an electronic tumor board application which incorporates AI into its logic for ‘Precision Oncology’; where codified clinical data, genomic, proteomic and image data come together to enhance care for cancer patients. Our next goal is to use this ability to mine and synthesize vast quantities of health data, medical images, and patient records (cross-modal data fusion) to allow clinicians to design AI assisted treatment plans for patients that also facilitate nursing professionals in planning care transitions.
UB is pioneering the field of Digital Pathology that enables the acquisition, management and interpretation of pathology information generated from a digitized glass slide to improve diagnosis, treatment decisions and patient care. This has the potential to enhance clinical research and student training, reduce laboratory costs and improve efficiencies. The next step is to develop the infrastructure that would allow UB researchers to combine high performing computers with state-of-the-art imaging techniques (e.g. atom probe tomography, super-resolution optical microscopy, and high-resolution cryogenic microscopy), deep sequencing, and other “‘omic” techniques to create fused data signatures of disease diagnosis, prognosis and response to therapy. This would be revolutionary in our ability to use AI tools to improve ‘patient encounter’ – to translate the use of better diagnostics for patient care, to minimize risk, and improve the overall health care at multiple levels.
AI technologies are helping researchers and clinicians discover predictive patterns in large data sets of genetic information, high-resolution images, and medical records. For example, UB researchers have used clinical text coded with ontologies for identifying patients with non-valvular atrial fibrillation to determine their risk score for stroke and anticoagulation medication induced bleeding. Our goal is to move this capability to the next level by bringing together researchers from diverse disciplines to advance engineering-driven-medicine for a broad range of applications. For example, to understand the environmental contributions to disease; predict patient responses to medical interventions; and track and help halt the spread of infectious diseases in large populations.
AI technologies are being used for testing medical interventions and drug discoveries. At UB, faculty have created a new capability to identify novel drug-target pairs - the Computational Analysis of Novel Drug Opportunities (CANDO) to improve drug discovery. This platform creates a matrix of compound- or drug-protein interactions based on predicting their strengths using a variety of simulation tools and all-atom knowledge based potentials that has shown to be useful for designing and validating drug targets. UB is well positioned to move to the next level of drug discoveries and other medical interventions; in accelerating the development of pharmaceuticals through rapid prototyping simulations without incurring the high cost of clinical trials, also improve our abilities to rule out implementation of substandard treatment and health care delivery modalities.
Chronic conditions and co-morbidities are highly prevalent in older adults; they lower the quality of life and contribute to the leading causes of death among this population. According to AARP’s estimate, only four potential family caregivers will be available for every person requiring care by 2030, indicating an oncoming “caregiving cliff”. Innovative technology has great potential to promote health spans (the length of time that a person is healthy) and alleviate caregivers’ burden. UB faculty in the School of Nursing are exploring the use of AI to empower older adults, especially those who still live at home, to self-manage their chronic conditions and engage in social interactions that can lead to greater independence, less social isolation, and better quality of life. Voice activated personal devices are being used to remind older adults to take their medications, keep their doctor appointments, encourage physical activity, monitor diets, and connect with their family members. Motion sensors have been adopted to track movements of older adults to determine if they have fallen or may need help, as well as detect changes in their gait, affect, interactions and communication. Such information can assist with early identification of the warning signs of depression and dementia. As next steps, UB faculty will explore more individualized technological innovation to promote health within older adults, ease the family caregivers’ burden, thus reducing the burden of diseases and health care costs in a rapidly aging population.