Impact Stories

Using AI to Save Lives in New York

April 8, 2025

Einstein-Montefiore CTSA-enabled Artificial Intelligence Technology Saves Lives

https://pubmed.ncbi.nlm.nih.gov/30373653/

The Einstein-Montefiore Institute for Clinical and Translational Research (ICTR) has as its core mission the improvement of human health through research. In partnership with our NIH-funded CTSA together with investments from the technology sector and from the Montefiore Health System, a new artificial intelligence (AI) tool has been developed to help doctors better identify and treat patients, including a deadly form of lung failure in acute illness. This AI technology is named “Patient-Centered Analytic Machine” (PALM). 

Each year, about 200,000 Americans suffer from acute respiratory distress syndrome (ARDS), which occurs when the lungs fill with fluid and organs are deprived of oxygen. The challenge is that up to 40 percent of the time ARDS can be missed because these patients are often extremely ill and have other life-threatening conditions, such as shock, pneumonia, or trauma. Thus, it is easy for clinicians to attribute symptoms to another condition and miss ARDS. Dr. Michelle Gong, chief of Critical Care Medicine and an Einstein CTSA core director, led the patient-centered clinical study at Montefiore and other US hospitals (funded by grants from NIH and the Agency for Healthcare Research and Quality). PALM screened patients throughout the health systems and flagged patients at risk of developing ARDS. Once identified, clinicians received guidance for treating patients with the condition. Compared to conventional tools to predict ARDS, PALM alerted clinicians earlier and more accurately—actually saving patients’ lives (1,2).

Dr. Parsa Mirhaji, director of the ICTR’s Center for Health Data Innovations and of the Einstein CTSA Informatics Core, has developed a host of innovative technologies that enabled the ARDS trial as well as many other projects aimed at using AI and machine learning to improve health care. His team of data and computer scientists use years of patient data with hundreds of critical data points to train computers to devise prediction algorithms. After creating the algorithms, PALM runs in the background of the electronic health record (EHR) system, flagging any patients who match the risk profile. Other applications of PALM include preventing life-threatening infections with analytically-driven sepsis management, transforming the patient experience from ‘seeking care’ to ‘receiving care’ at home, and cancer care using PALM to integrate MRI images, radiology reports, and clinical data to develop precision medicine pathways to prevent and intervene earlier in cases of spinal cord compression from metastatic cancer which is a devastating event for patients and their families. Of note, the PALM technology maintains very high levels of privacy for patient data (3), a vital feature since the future of health care rests on safe and ethical use of EHR data (4). 

The power of AI and predictive analytics to accurately pinpoint patients at risk for many critical conditions is only one aspect of machine learning that can help clinicians provide better, faster, and less expensive health care. The CTSA vision and NIH funding bringing together human ingenuity with essential computational infrastructure has laid the foundations for this innovative R&D to be helping patients today.

PUBLICATIONS

  1. Multicenter derivation and validation of an early warning score for acute respiratory failure or death in the hospital. Dziadzko MA, Novotny PJ, Sloan J, Gajic O, Herasevich V, Mirhaji P, Wu Y, Gong MN. Crit Care. 2018 Oct 30;22(1):286. doi: 10.1186/s13054-018-2194-7. PMID: 30373653
  2. Early intervention of patients at risk for acute respiratory failure and prolonged mechanical ventilation with a checklist aimed at the prevention of organ failure: protocol for a pragmatic stepped-wedged cluster trial of PROOFCheck. Gong MN, Schenk L, Gajic O, Mirhaji P, Sloan J, Dong Y, Festic E, Herasevich V. BMJ Open. 2016 Jun 10;6(6):e011347. doi: 10.1136/bmjopen-2016-011347. PMID: 27288382
  3. Preserving temporal relations in clinical data while maintaining privacy. Hripcsak G, Mirhaji P, Low AF, Malin BA. J Am Med Inform Assoc. 2016 Nov;23(6):1040-1045. doi: 10.1093/jamia/ocw001. PMID: 27013522

Leveraging electronic health records for clinical research. Raman SR, Curtis LH, Temple R, Andersson T, Ezekowitz J, Ford I, James S, Marsolo K, Mirhaji P, Rocca M, Rothman RL, Sethuraman B, Stockbridge N, Terry S, Wasserman SM, Peterson ED, Hernandez AF. Am Heart J. 2018 Aug;202:13-19. doi: 10.1016/j.ahj.2018.04.015. PMID: 29802975