AMAnoop Mayampurath

Anoop Mayampurath

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Anoop Mayampurath has a total of 159 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Explainable multimodal deep learning models for variable-length sequences in critically ill patients

    by Jennifer Martin, Majid Afshar, Askar Safipour Afshar, John Caskey, Dmitriy Dligach, Yanjun Gao, Jifan Gao, Guanhua Chen, Anoop Mayampurath, Matthew M. Churpek


  • Open Access

    Designing a Substance Misuse Data Dashboard for Overdose Fatality Review Teams: User-Centered Design Approach

    by Marie Pisani, Madeline K Oguss, Julia Dickson-Gomez, Constance Kostelac, Amy Parry, Starr Moss, Elizabeth Salisbury-Afshar, Brian Patterson, Michael Spigner, Megan Gussick, et al.

    Background: Overdose fatality review (OFR) is a public health process in which cases of fatal overdose are carefully reviewed to identify prevention strategies. Current OFR requires review of multiple unconnected data sources, which is a manually intensive process. The Substance Misuse Data Commons (SMDC) was created to link electronic health record data with data from local and state agencies into a single, cloud-based e-platform but does not currently have a data visualization tool. Objective: We aimed to use human factors design principles to develop a comprehensive dashboard for the SMDC that could facilitate enhanced processes to support OFR. Methods: We first surveyed OFR leaders in Wisconsin using the National Aeronautics and Space Administration-Task Load Index to understand the cognitive workload of 3 tasks: (1) analysis of population-level overdose trends, (2) selection and preparation of individual cases for review, and (3) abstraction of data from individual causes. We then conducted semistructured interviews to identify targets for workflow optimization. Next, we developed a prototype dashboard for evaluation using a synthetic dataset built with GPT-4. We subsequently performed iterative design sessions with heuristic evaluations and collected end-user feedback on the final prototype via a second round of semistructured interviews and targeted surveys, including the Unified Theory of Acceptance and Use of Technology and the Perceived Usefulness Questionnaire. Results: The National Aeronautics and Space Administration-Task Load Index revealed a moderately high mental workload with the current workflow for all 3 tasks, with mean scores of 12.60 (SD 3.31), 11.90 (SD 3.57), and 12.43 (SD 5.41) for tasks 1, 2, and 3, respectively. Interviews pointed to causes including technological challenges and a reliance on manual processes. The prototype dashboard addressed these concerns by integrating multiple data sources to generate population-level visualizations and patient-level event timelines. End users reported the potential for improved efficiency and data accessibility compared to antecedent processes. The Unified Theory of Acceptance and Use of Technology results indicated the dashboard would likely be adopted if made available, with a mean of 4.07 out of 5.00 (SD 0.65). The Perceived Usefulness Questionnaire results suggested moderate usefulness for both the aggregate and individual-level data, with means of 3.61 (SD 0.82) and 3.64 (SD 0.85) out of 5.00, respectively. Conclusions: OFR is a data-intensive process that traditionally demands substantial cognitive and manual effort, and there are multiple barriers to efficiently collecting data and presenting them for review. The dashboard offers a user-centered, informatics-based approach to streamline data aggregation and presentation, potentially enhancing the efficiency of case reviews. Implementing a dashboard that consolidates and visualizes disparate data sources has the potential to alleviate the manual workload in OFR. Ultimately, our aim is to deliver a finalized data dashboard with real-world SMDC data, giving OFR leaders additional tools to aid in their rigorous work shaping interventions to reduce overdose fatalities.


  • Open Access

    Integrating multiple data sources to predict all-cause readmission or mortality in patients with substance misuse

    by Tim Gruenloh, Preeti Gupta, Askar Safipour Afshar, Madeline Oguss, Elizabeth Salisbury-Afshar, Marie Pisani, Ryan P. Westergaard, Michael Spigner, Megan Gussick, Matthew Churpek, et al.

    Patients with substance misuse who are admitted to the hospital are at heightened risk for adverse outcomes, such as readmission and death. This study aims to develop methods to identify at-risk patients to facilitate timely interventions that can improve outcomes and optimize healthcare resources. To accomplish this, we leveraged the Substance Misuse Data Commons to predict 30-day death or readmission from hospital discharge in patients with substance misuse. We explored several machine learning algorithms and approaches to integrate information from multiple data sources, such as structured features from a patient’s electronic health record (EHR), unstructured clinical notes, socioeconomic data, and emergency medical services (EMS) data. Our gradient-boosted machine model, which combined structured EHR data, socioeconomic status, and EMS data, was the best-performing model (c-statistic 0.746 [95% CI: 0.732-0.759]), outperforming other machine learning methods and structured data source combinations. The addition of unstructured text did not improve performance, suggesting a need for further exploration of how to leverage unstructured data effectively. Feature importance plots highlighted the importance of prior hospital and EMS encounters and discharge disposition in predicting our primary outcome. In conclusion, we integrated multiple data sources that offer complementary information from data sources beyond the typically used EHRs for risk assessment in patients with substance misuse.


  • Explaining alerts from a pediatric risk prediction model using clinical text

    by Samuel Nycklemoe, Sriharsha Devarapu, Yanjun Gao, Kyle Carey, Nicholas Kuehnel, Neil Munjal, Priti Jani, Matthew Churpek, Dmitriy Dligach, Majid Afshar, et al.

    Objective: Risk prediction models are used in hospitals to identify pediatric patients at risk of clinical deterioration, enabling timely interventions and rescue. The objective of this study was to develop a new explainer algorithm that uses a patient’s clinical notes to generate text-based explanations for risk prediction alerts. Materials and Methods: We conducted a retrospective study of 39 406 patient admissions to the American Family Children’s Hospital at the University of Wisconsin-Madison (2009-2020). The pediatric Calculated Assessment of Risk and Triage (pCART) validated risk prediction model was used to identify children at risk for deterioration. A transformer model was trained to use clinical notes from the 12-hour period preceding each pCART score to predict whether a patient was flagged as at risk. Then, label-aware attention highlighted text phrases most important to an at-risk alert. The study cohort was randomly split into derivation (60%) and validation (20%) data, and a separate test (20%) was used to evaluate the explainer’s performance. Results: Our pCART Explainer algorithm performed well in discriminating at-risk pCART alert vs no alert (c-statistic 0.805). Sample explanations from pCART Explainer revealed clinically important phrases such as “rapid breathing,” “fall risk,” “distension,” and “grunting,” thereby demonstrating excellent face validity. Discussion: The pCART Explainer could quickly orient clinicians to the patient’s condition by drawing attention to key phrases in notes, potentially enhancing situational awareness and guiding decision-making. Conclusion: We developed pCART Explainer, a novel algorithm that highlights text within clinical notes to provide medically relevant context about deterioration alerts, thereby improving the explainability of the pCART model.


  • Open Access

    Comparison of Multimodal Deep Learning Approaches for Predicting Clinical Deterioration in Ward Patients: Observational Cohort Study

    by Charles A Kotula, Jennie Martin, Kyle A Carey, Dana P Edelson, Dmitriy Dligach, Anoop Mayampurath, Majid Afshar, Matthew M Churpek

    Background: Implementing machine learning models to identify clinical deterioration in the wards is associated with decreased morbidity and mortality. However, these models have high false positive rates and only use structured data. Objective: We aimed to compare models with and without information from clinical notes for predicting deterioration. Methods: Adults admitted to the wards at the University of Chicago (development cohort) and University of Wisconsin-Madison (external validation cohort) were included. Predictors consisted of structured and unstructured variables extracted from notes as concept unique identifiers (CUIs). We parameterized CUIs in 5 ways: standard tokenization (ST), International Classification of Diseases rollup using tokenization (ICDR-T), International Classification of Diseases rollup using binary variables (ICDR-BV), concept unique identifiers as SapBERT embedding (SE), and concept unique identifier clustering using SapBERT embeddings (CC). Each parameterization method combined with structured data and each structured data-only method were compared for predicting intensive care unit transfer or death in the next 24 hours using deep recurrent neural networks. Results: The development (University of Chicago) cohort included 284,302 patients, while the external validation (University of Wisconsin-Madison) cohort included 248,055 patients. In total, 4.9% (n=26,281) of patients experienced the outcome. The SE model achieved the highest area under the precision-recall curve (0.208), followed by CC (0.199) and the structured-only model (0.199), ICDR-BV (0.194), ICDR-T (0.166), and ST (0.158). The CC and structured-only models achieved the highest area under the receiver operating characteristic (0.870), followed by ICDR-T (0.867), ICDR-BV (0.866), ST (0.860), and SE (0.859). Regarding sensitivity and positive predictive value, the CC model achieved the greatest positive predictive value (12.53%) and sensitivity (52.15%) at the cutoff that flagged 5% of the observations in the test set. At the 15% cutoff, the ICDR-T, CC, and ICDR-BV models tied for the highest positive predictive value at 5.67%, while their sensitivities were 70.95%, 70.92%, and 70.86%, respectively. All models were well calibrated, achieving Brier scores in the range of 0.011‐0.012. The modified integrated gradients method revealed that CUIs corresponding to terms such as “NPO – nothing by mouth,” “chemotherapy,” “transplanted tissue,” and “dialysis procedure” were most predictive of deterioration. Conclusions: A multimodal model combining structured data with embeddings using SapBERT had the highest area under the precision-recall curve, but performance was similar between models with and without CUIs. Although the addition of CUIs from notes to structured data did not meaningfully improve model performance for predicting clinical deterioration, models using CUIs could provide clinicians with relevant information and additional clinical context for supporting decision-making.


  • Open Access

    Age and Saving Lives in Crisis Standards of Care: A Multicenter Cohort Study of Triage Score Prognostic Accuracy

    by Michael Hermsen, Patrick G. Lyons, Govind Jd Persad, Alice F. Bewley, Chengsheng Mao, Kaveri Chhikara, Anoop Mayampurath, Matthew Churpek, Monica E. Md Peek, Yuan Luo, et al.

    IMPORTANCE: Current protocols to triage life support use scores that are biased and inaccurate. OBJECTIVES: To determine if adding age to triage protocols used in disaster scenarios improves the identification of critically ill patients likely to survive. DESIGN, SETTING, AND PARTICIPANTS: Observational cohort study from March 1, 2020, to March 1, 2022, at 22 hospitals in three networks, divided into derivation (12 hospitals) and validation cohorts (ten hospitals). Participants were critically ill adults (90% COVID-19 positive) who would have needed life support during an overwhelming case surge. Life support was defined as vasoactive medications for shock, invasive or noninvasive mechanical ventilation, or oxygen therapy with Pao 2/Fio 2 less than 200. MAIN OUTCOMES AND MEASURES: The primary outcome was death in the intensive care unit. We fit logistic regression models using a modified Sequential Organ Failure Assessment (SOFA) score with and without age in the derivation cohort and assessed predictive performance in the validation cohort using area under the receiver operating characteristic curves (AUCs) and compared observed and predicted mortality. RESULTS: The final analysis contained 7,660 patients with 16,711 life-support episodes. In the validation cohort, the AUC for age plus SOFA was significantly higher than the AUC for SOFA alone (0.66 vs. 0.54; p < 0.001). SOFA score substantially overpredicted mortality (13% predicted vs. 5% observed) for younger patients (< 40 yr) and underestimated mortality (14% predicted vs. 31% observed) for older patients (> 80 yr). In contrast, age plus SOFA had good calibration overall and across age groups. The addition of age improved but did not eliminate differences between observed and predicted mortality across racial-ethnic groups. CONCLUSIONS AND RELEVANCE: Age-inclusive triage better identifies ICU survivors than SOFA alone and is more equitable. Incorporating age into prioritization algorithms could save more lives in a crisis scenario.


  • Top Cited Papers
    Open Access

    Machine Learning for Predicting Critical Events Among Hospitalized Children

    by Sierra Strutz, Huan Liang, Kyle Carey, Fereshteh Bashiri, Priti Jani, Emily Gilbert, Julie L. Fitzgerald, Nicholas Kuehnel, Maya Dewan, L. Nelson Sanchez-Pinto, et al.

    Importance: Unrecognized deterioration among hospitalized children is associated with a high risk of mortality and morbidity. The current approach to pediatric risk stratification is fragmented, as each hospital unit (emergency, ward, or intensive care) uses different tools for predicting specific outcomes.Objective: To develop a machine learning model for the early detection of deterioration across all units, thereby enabling a unified risk assessment throughout the patient’s hospital stay.Design, Setting, and Participants: This retrospective cohort study used data from pediatric (age C statistic: XGB, 0.86; ward-focused models, 0.82 [P < .001] and 0.70 [P < .001]) and the number needed to alert (at an example 80% sensitivity: XGB, 6 ward-focused models: 9 and 11). The deep learning models did not exhibit improved performance. The XGB model performed better or equivalent to models trained for a specific hospital unit.Conclusions and Relevance: This retrospective cohort study describes the development of a novel hospitalwide model for continuously predicting the risk of critical events through the entirety of a child’s stay. The model facilitated a unified framework for risk assessment in a pediatric hospital.


  • Top Cited Papers
    Open Access

    Automating Evaluation of AI Text Generation in Healthcare with a Large Language Model (LLM)-as-a-Judge

    by Emma Croxford, Yanjun Gao, Elliot First, Nicholas Pellegrino, Miranda Schnier, John Caskey, Madeline Oguss, Graham Wills, Guanhua Chen, Dmitriy Dligach, et al.

    Electronic Health Records (EHRs) store vast amounts of clinical information that are difficult for healthcare providers to summarize and synthesize relevant details to their practice.12345 To reduce cognitive load on providers, generative AI with Large Language Models have emerged to automatically summarize patient records into clear, actionable insights and offload the cognitive burden for providers.678 However, LLM summaries need to be precise and free from errors, making evaluations on the quality of the summaries necessary.9101112 While human experts are the gold standard for evaluations, their involvement is time-consuming and costly.13 Therefore, we introduce and validate an automated method for evaluating real-world EHR multi-document summaries using an LLM as the evaluator, referred to as LLM-as-a-Judge. Benchmarking against the validated Provider Documentation Summarization Quality Instrument (PDSQI)-9 for human evaluation14, our LLM-as-a-Judge framework uses the PDSQI-9 rubric and demonstrated strong inter-rater reliability with human evaluators. GPT-o3-mini achieved the highest intraclass correlation coefficient of 0.818 (95% CI 0.772, 0.854), with a median score difference of 0 from human evaluators, and completes evaluations in just 22 seconds. Overall, the reasoning models excelled in inter-rater reliability, particularly in evaluations that require advanced reasoning and domain expertise, outperforming non-reasoning models, those trained on the task, and multi-agent workflows. Cross-task validation on the Problem Summarization15 task similarly confirmed high reliability. By automating high-quality evaluations, medical LLM-as-a-Judge offers a scalable, efficient solution to rapidly identify accurate and safe AI-generated summaries in healthcare settings.


  • Multicenter Development and Validation of a Multimodal Deep Learning Model to Predict Moderate to Severe AKI

    by Jay L. Koyner, Jennie Martin, Kyle A. Carey, John Caskey, Dana P. Edelson, Anoop Mayampurath, Dmitriy Dligach, Majid Afshar, Matthew M. Churpek

    Key Points: We developed and validated a multimodal (structured and unstructured data) model to predict moderate to severe AKI using multicenter data. This multimodal AKI risk score accurately identifies patients who will develop stage 2 AKI over 2 days earlier than serum creatinine alone. The multimodal model performed better than a model based solely on structured data and performed similarly during temporal and site-based validation. Background: Prior models for the early identification of AKI have used structured data (e.g., vital signs and laboratory values). We aimed to develop and validate a deep learning model to predict moderate to severe AKI by combining structured data and information from unstructured notes. Methods: Adults (18 years or older) admitted to the University of Wisconsin (2009–2020) and the University of Chicago Medicine (2016–2022) were eligible for inclusion. Patients were excluded if they had no documented serum creatinine (SCr), ESKD, an admission SCr ≥3.0 mg/dl, developed ≥stage 2 AKI before reaching the wards or intensive care unit, or required dialysis (KRT) within the first 48 hours. Text from unstructured notes was mapped to standardized concept unique identifiers to create predictor variables, and structured data variables were also included. An intermediate fusion deep learning recurrent neural network architecture was used to predict ≥stage 2 AKI within the next 48 hours. This multimodal model was developed in the first 80% of the data and temporally validated in the next 20%. Results: There were 339,998 admissions in the derivation cohort and 84,581 in the validation cohort, with 12,748 (3%) developing ≥stage 2 AKI. Patients with ≥stage 2 AKI were older, more likely to be male, had higher baseline SCr, and were more commonly in the intensive care unit (P < 0.001 for all). The multimodal model outperformed a model based only on structured data for all outcomes, with an area under the receiver operating characteristic curve (95% confidence interval) of 0.88 (0.88 to 0.88) for predicting ≥stage 2 AKI and 0.93 (0.93 to 0.94) for receiving KRT. The area under the precision-recall-curve for ≥stage 2 AKI was 0.20. The results were similar during external validation. Conclusions: We developed and validated a multimodal deep learning model using structured and unstructured data that predicts the development of severe AKI across the hospital stay for earlier intervention.


  • Open Access

    Comparison of Multimodal Deep Learning Approaches for Predicting Clinical Deterioration in Ward Patients: An Observational Cohort Study (Preprint)

    by Charles A Kotula, Jennie Martin, Kyle A Carey, Dana P Edelson, Dmitriy Dligach, Anoop Mayampurath, Majid Afshar, Matthew M Churpek

    BACKGROUND Implementing machine learning models to identify clinical deterioration on the wards is associated with improved outcomes. However, these models have high false positive rates and only use structured data. OBJECTIVE We aim to compare models with and without information from clinical notes for predicting deterioration. METHODS Adults admitted to the wards at the University of Chicago (development cohort) and University of Wisconsin-Madison (external validation cohort) were included. Predictors consisted of structured and unstructured variables extracted from notes as Concept Unique Identifiers (CUIs). We parameterized CUIs in five ways: Standard Tokenization (ST), ICD Rollup using Tokenization (ICDR-T), ICD Rollup using Binary Variables (ICDR-BV), CUIs as SapBERT Embeddings (SE), and CUI Clustering using SapBERT embeddings (CC). Each parameterization method combined with structured data and structured data-only were compared for predicting intensive care unit transfer or death in the next 24 hours using deep recurrent neural networks. RESULTS The study included 506,076 ward patients, 4.9% of whom experienced the outcome. The SE model achieved the highest AUPRC (0.208), followed by CC (0.199) and the structured-only model (0.199), ICDR-BV (0.194), ICDR-T (0.166), and ST (0.158). The CC and structured-only models achieved the highest AUROC (0.870), followed by ICDR-T (0.867), ICDR-BV (0.866), ST (0.860), and SE (0.859). CONCLUSIONS A multimodal model combining structured data with embeddings using SapBERT had the highest AUPRC, but performance was similar between models with and without CUIs. The addition of CUIs from notes to structured data did not meaningfully improve model performance for predicting clinical deterioration.