🟡 Dr. Nadkarni; 🔵 Dr. Chan; 🟠Dr. Hofer; 🟢 Dr. Sakhuja; ⚪ Dr. Abbott
Project Aim and DescriptionÂ
The All of Us (AoU) Research Program seeks to establish one of the most diverse health databases in history, and to advance this goal, the proposal introduces the New York Coalition, comprising several prominent academic medical centers and partners experienced in engaging underrepresented populations in biomedical research. This coalition includes Mount Sinai Health System, Weill Cornell Medicine, New York City Health+Hospitals, the Institute for Family Health, and NYU Langone, which aims to recruit 88,500 new participants and retain 35,400 existing participants through effective strategies. These strategies involve creating a diverse stakeholder engagement board, utilizing a research equity toolkit, offering pilot grants for community organizations to serve as recruitment sites, and implementing a nationally recognized training program for coordinators and community health workers. With a proven track record in participant recruitment for AoU and other studies, the coalition aims to develop coordinated strategies to effectively engage underrepresented populations in New York City, thereby enhancing the success of the AoU program. This initiative is part of the NIH’s Helping to End Addiction Long-term (HEAL) initiative, which focuses on accelerating research to improve treatment for opioid misuse and addiction.
Date: 2024-2025
Award Organization: NIH Office of the Director [3OT2OD037643-01S1]
PI: KRAFT, MONICA; Co-PI: NADKARNI, GIRISH NITIN
Project Aim and DescriptionÂ
This project aims to address the challenges of managing obstructive sleep apnea (OSA), a condition affecting over one billion adults and linked to cardiovascular disease (CVD). Despite the use of continuous positive airway pressure (CPAP) therapy, randomized clinical trials have not demonstrated significant benefits on CVD event rates, potentially due to suboptimal CPAP adherence and variability in OSA presentations. The proposal seeks to utilize machine learning (ML) on multimodal datasets to develop predictive tools for identifying patients at higher risk for CVD events and understanding the heterogeneity of CPAP treatment effects. By validating these models with real-world electronic health records, the project aspires to personalize OSA management, improve patient outcomes, and reduce healthcare costs through informed treatment decisions regarding CPAP use.
Related Publications:
- The Role of Artificial Intelligence in Nephrology Clinical Trials
- Heterogeneous Effects of Continuous Positive Airway Pressure in Non-Sleepy Obstructive Sleep Apnea on Cardiovascular Disease Outcomes: Post Hoc Machine Learning Analysis of the ISAACC Trial (ECSACT Study)
Date: 2024-2027
Award Organization: National Heart Lung and Blood Institute
[1R01HL168897-01]
PI: SHAH, NEOMI A; Co-PI: NADKARNI, GIRISH NITIN
Project Aim and Description
This project seeks to harness the power of machine learning and healthcare data to forecast outcomes during the perioperative period, with a primary focus on cardio-respiratory instability (CRI), encompassing hypotension and arrhythmia. This includes a comprehensive machine learning curriculum, enabling the use of ML techniques on a vast dataset of over 30,000 patients, combining electronic medical records, physiologic waveforms, and genomic data. Through collaboration with esteemed mentors and an expert advisory committee, this initiative aims to develop predictive models for CRI, addressing a critical gap in perioperative medicine. The potential impact lies in empowering clinicians to foresee and prevent perioperative instability, ultimately improving patient outcomes and paving the way for further research in this domain.
Related Publications:
- Laboratory Result Reference Ranges Stratified for Patient-Reported Sex and Ethnicity Are More Closely Associated With Postoperative Outcomes Than Currently Used Reference Ranges: A Retrospective Data Analysis
- Visual Analytics to Leverage Anesthesia Electronic Health Record
- Integration of feature vectors from raw laboratory, medication and procedure names improves the precision and recall of models to predict postoperative mortality and acute kidney injury
Date: 2020-2025
Supporter by: National Heart Lung and Blood Institute [5K01HL150318-04]
PI: HOFER, IRA
Project Aim and Description
The research aims to addresse the prevalent issues of intradialytic hypotension (IDH) and major adverse cardiovascular events (MACE) among hemodialysis (HD) patients. Although strategies exist to reduce these occurrences, the lack of precise predictive risk models hampers their effective utilization. Utilizing electrocardiography (ECG), this study aims to leverage deep learning to predict key outcomes. The study involves 1000 HD patients in New York City (derivation) and 150 patients in North Carolina (validation). Through ECG data and deep learning models, the study aims to predict IDH and MACE both in the short term (within 30 days) and the long term (1 year). Positive outcomes may lead to the deployment of predictive models in HD units for prevention and detection of IDH and MACE, as well as the utilization of novel wearables for risk prediction.
Related Publications:
Date:Â 2023-2028
Supported by: National Heart Lung and Blood Institute [1R01HL167050-01A1]
PI:Â NADKARNI, GIRISH NITIN;Â Co-PI: CHARYTAN, DAVID; M DIVERS, JASMIN
Project Aim and Description
This research aims to leverage data driven approaches to prevent the development of persistent acute kidney injury after cardiac surgery. Acute kidney injury is seen in over one third of patients after cardiac surgery. Patients who develop persistent acute kidney injury are especially at risk for poor outcomes. Management of acute kidney injury consists of interventions to reduce its further progression and management of its complications. These interventions when given as a bundle decrease the risk of clinically significant acute kidney injury, but the compliance with implementation of this bundle is very low. Our study addresses this challenge by developing a personalized strategy of interventions for patients using Reinforcement Learning. Our study also identifies individuals at high risk for persistent acute kidney injury, allowing targeted implementation of these interventions in this high-risk patient population.
Related Publications:
- Acceptance of Acute Kidney Injury Alert by Providers in Cardiac Surgery Intensive Care Unit
- Geographic and Temporal Trends in COVID-Associated Acute Kidney Injury in the National COVID Cohort Collaborative
- Contemporary Coronary Artery Bypass Grafting vs Multivessel Percutaneous Coronary Intervention
- Impact of Using Blood Warmer During Continuous Kidney Replacement Therapy in Patients With Acute Kidney Injury
- Clinical Informatics in Critical Care Medicine
- Heterogeneity in Acute Kidney Injury Management in Critically Ill Patients: National Survey
Date: 2022-2027
Supported by:Â National Institute of Diabetes and Digestive and Kidney Diseases [5K08DK131286]
PI: Sakhuja, Ankit
Project Aim and Description
The proposal seeks to counteract the existing racial disparities prevalent in kidney transplantation (KTx) by addressing the systemic biases and structural hurdles embedded within the transplant process. Acknowledging the far-reaching influence of bias and social determinants on patient evaluation and the waitlisting process, the initiative introduces a multifaceted intervention guided by a diverse stakeholder board. This innovative strategy involves patient navigators to assist individuals through the evaluation journey, focusing on addressing social needs and managing mental health challenges. Furthermore, an anti-bias initiative is proposed to refine language and improve education among KTx providers. Through a clinical trial, the project aims to assess the impact of these interventions on waitlisting among adults with kidney disease on hemodialysis. The overarching goal is to establish a national framework that mitigates inequities within the kidney transplantation process, emphasizing community involvement and collaborative stakeholder engagement.
Date: 2023-2028Â
Supported by: National Institute of Diabetes and Digestive and Kidney Diseases [3R01HL155915-03S1]
PI: CHAN, LILI; Co-PI: HOROWITZ, CAROL R; NADKARNI, GIRISH NITIN; MELAMED, MICHAL L
Project Aim and Description
The Mount Sinai Kidney Precision Medicine Project (KPMP) aims to revolutionize kidney disease understanding and treatment. With 40 million affected Americans and a growing number due to comorbid conditions, predicting and treating acute kidney injury (AKI) and chronic kidney disease (CKD) is challenging. This project establishes a patient-centered biopsy cohort to uncover AKI and CKD markers through molecular pathways in human kidney tissue. With robust patient recruitment strategies, the project focuses on diverse CKD patient cohorts while leveraging tools to identify high-risk patients for disease progression. Collaborating with KPMP, this initiative strives to unlock groundbreaking insights, transforming kidney disease management and treatment paradigms.
Related Publications:
- Introduction to Artificial Intelligence and Machine Learning in Nephrology
- Signal recovery in single cell batch integration
- Molecular Signatures of Glomerular Neovascularization in a Patient with Diabetic Kidney Disease
Date: 2022-2027
Supported by: National Institute of Diabetes and Digestive and Kidney Diseases [1U01DK137259-01]
PI: CAMPBELL, KIRK; Co-PI: COCA, STEVEN G; NADKARNI, GIRISH NITIN
Project Aim and Description
The research project has a dual objective. Firstly, it aims to enhance the identification of adverse social determinants of health (SDOH) in patients with a specific genetic mutation using natural language processing (NLP). NLP proves crucial as many social factors reside in unstructured medical narratives, enabling more accurate identification compared to the current code-based method. The second goal involves investigating the potential correlation between adverse SDOH and poor health outcomes in these patients. By examining medical records, researchers will compare health outcomes between those with and without adverse SDOH. The hypothesis suggests that adverse SDOH will correlate with worse health outcomes, with a probable incremental effect based on the severity of adverse SDOH. The study acknowledges limitations such as reliance on a single NLP tool and the binary definition of adverse SDOH, which future studies may refine by employing diverse methodologies and more detailed data.
Related Publications:
- Introduction to Artificial Intelligence and Machine Learning in Nephrology
- Proteomic characterization of acute kidney injury in patients hospitalized with SARS-CoV2 infection
- Federated Learning in Risk Prediction: A Primer and Application to COVID-19-Associated Acute Kidney Injury
- Phenome-wide Mendelian randomization study of plasma triglyceride levels and 2600 disease traits
- Causal effect of adiposity on the risk of 19 gastrointestinal diseases: a Mendelian randomization study
- Proteomic Characterization of Acute Kidney Injury in Patients Hospitalized with SARS-CoV2 Infection
Date: 2021-2025
Supported by: National Heart Lung and Blood Institute [5R01HL155915-03]
PI: DO, RON;Â Co-PI: NADKARNI, GIRISH NITIN
Project Aim and Description
The project aims to advance understanding and diagnosis of hereditary transthyretin amyloid cardiomyopathy (hATTR-CM) caused by mutations in the TTR gene. Despite the TTR V122I mutation significantly increasing heart failure risk, underdiagnosis prevails, particularly in African American and Hispanic populations. This research focuses on filling critical knowledge gaps to enhance targeted therapies. Leveraging polygenic risk scores, the study examines 6,609 African Americans and 9,006 Hispanics to grasp the interaction between genetic and clinical risk factors. Employing machine learning on electronic health records will identify V122I carriers in an 8-million patient cohort. Additionally, the project recalls asymptomatic carriers for in-depth heart imaging to detect subclinical effects. By utilizing diverse biobanks and cutting-edge methods, this initiative strives to enable precision medicine in heart failure management for minority populations.
Related Publications:
- Proteomic characterization of acute kidney injury in patients hospitalized with SARS-CoV2 infection
- Phenome-wide Mendelian randomization study of plasma triglyceride levels and 2600 disease traits
- Development of a machine learning model using electrocardiogram signals to improve acute pulmonary embolism screening
- Overcoming constraints on the detection of recessive selection in human genes from population frequency data
Date: 2021-2025
Supported by: National Heart Lung and Blood Institute [5R01HL155915-03]
PI: DO, RON;Â Co-PI: NADKARNI, GIRISH NITIN
Project Aim and Description The project aims to address the critical need for accurate predictive models in critically ill patients with acute kidney injury (AKI) and hemodynamic instability who require continuous renal replacement therapy (CRRT) in intensive care units (ICU). As no universally accepted tools exist for survival or kidney recovery prediction in this population, this research leverages artificial intelligence and deep learning models, using diverse clinical and CRRT machine data from multiple institutions. The study intends to develop and validate predictive models for survival and identify patient subtypes through the Feasible Solution Algorithm (FSA). Preliminary results show the superiority of deep learning models in predicting CRRT-free survival and mortality. The hypothesis is that time-series multi-modal data will generate accurate risk predictions, aiding clinical interventions and facilitating personalized treatments in CRRT. This research is innovative, aiming to develop decision support tools and identify patient subtypes for improved CRRT interventions and outcomes.
Related Publications:
Date: 2023-2027
Award Organization: National Institute of Diabetes and Digestive and Kidney Diseases [1R01DK133539-01A1]
PI: NEYRA, JAVIER A.; Co-PI: NADKARNI, GIRISH NITIN
Project Aim and DescriptionÂ
This research project focuses on deciphering the underlying causes of racial and ethnic disparities in kidney disease, particularly concerning the Apolipoprotein L1 (APOL1) gene variants prevalent in African ancestry populations. While these high-risk APOL1 genotypes are significantly associated with kidney disease, not all carriers manifest the condition, suggesting additional genetic or environmental factors, termed 'second hits.' By analyzing genetic and environmental modifiers in sizable datasets of minority individuals, this study aims to uncover these 'second hits' and their role in influencing kidney disease development among those with APOL1 high-risk genotypes. The investigation spans diverse cohorts, exploring SNPs, air pollution, neighbourhood factors, heavy metal exposure, and inflammatory biomarkers to better comprehend the APOL1-kidney disease link and address disparities in kidney health among different ethnicities.
Related Publications:
- A Primer in Precision Nephrology: Optimizing Outcomes in Kidney Health and Disease through Data-Driven Medicine
- The Disclosure of Personally Identifiable Information in Studies of Neighborhood Contexts and Patient Outcomes
Date: 2020-2025
Award Organization: National Institute of Diabetes and Digestive and Kidney Diseases [5R01DK127139-03]
PI: NADKARNI, GIRISH NITIN
Project Aim and DescriptionÂ
The primary objective of this project is to support Dr. Lili Chan’s development as an independent clinical investigator focused on enhancing risk prediction for adverse outcomes in hemodialysis (HD) patients by incorporating social determinants of health (SDOH) using electronic health records (EHR). Dr. Chan has assembled a multidisciplinary mentoring team led by Dr. Steven Coca and co-mentored by Dr. Peter Kotanko, along with experts in machine learning, psychosocial factors, and biostatistics. Her training plan encompasses advanced statistical methodologies, bioinformatics, patient-centered outcomes, and career development. The research aims to address the high morbidity and mortality rates among HD patients by improving risk stratification through a cohort study involving diverse patients receiving care from various hemodialysis units in New York City. The specific aims include assessing the impact of SDOH on hospitalizations, accurately identifying these determinants using natural language processing, and developing risk prediction models that integrate both standard measures and SDOH. This innovative approach seeks to identify high-risk HD patients for future intervention trials and lays the groundwork for subsequent R01 studies that will validate findings and test EHR-integrated clinical decision tools aimed at reducing hospitalizations, readmissions, and mortality.
Related Publications:
- Testing Interventions that Address Kidney Health Disparities
- Comparison of predicting cardiovascular disease hospitalization using individual, ZIP code-derived, and machine learning model-predicted educational attainment in New York City
Date: 2021-2025
Award Organization: National Institute of Diabetes and Digestive and Kidney Diseases [1K23DK124645-01A1 ]
PI: CHAN, LILI
Project Aim and DescriptionÂ
The K08 research project led by Dr. Ethan Abbott aims to establish him as an independent physician-scientist focused on improving survival rates for out-of-hospital cardiac arrest (OHCA) patients by addressing healthcare disparities. Supported by a multidisciplinary mentorship team, his project seeks to enhance predictions of 30-day survival and hospital discharge by identifying individual-level health-related social needs (HRSN) that current models overlook. Utilizing data science techniques such as natural language processing (NLP) and large language models (LLMs), Dr. Abbott plans to create a baseline predictive model, evaluate the extraction of HRSN, and assess the impact of including HRSN on model performance. His career development plan includes acquiring skills in mediation analysis, predictive analytics, and research independence. The results will provide preliminary data for a future R01 application aimed at validating the predictive model and improving OHCA clinical care and survival outcomes.
Related Publications:
Date: 2024-2025
Award Organization: National Heart Lung and Blood Institute
[1K08HL169980-01A1 ]
PI: ABBOTT, ETHAN