About

Nicholas Kuehnel, MD, is an Associate Professor in the Department of Emergency Medicine and Pediatrics at the University of Wisconsin–Madison. He serves as Vice Chair of Clinical Operations for the Department of Emergency Medicine and is the Division Chief of Pediatric Emergency Medicine, leading clinical and academic efforts at the Pediatric Emergency Department at American Family Children’s Hospital. Dr. Kuehnel has held multiple leadership roles since joining the department in 2017, including Medical Director of Pediatric Emergency Medicine and Performance Improvement, and was the inaugural Medical Director for Quality and Safety for UW Health Kids, where he led systemwide initiatives to improve pediatric quality, safety, and patient experience. His research includes clinical decision support for early detection of deterioration in hospitalized children, funded by NIH/NHLBI. Dr. Kuehnel completed his undergraduate studies at the University of Wisconsin–Madison, earned his medical degree from the Medical College of Wisconsin, and completed pediatrics residency at Ann & Robert H. Lurie Children’s Hospital of Chicago, followed by a pediatric emergency medicine fellowship at MCW and Children’s Hospital of Wisconsin.

Research topics

• Computer Science
• Medical emergency
• Medicine
• Political Science
• Telecommunications
• Artificial Intelligence
• Family medicine
• Data Mining
• Nursing
• Emergency medicine
• Psychology
• Internal medicine
• Pathology

Selected publications

• Machine Learning for Predicting Critical Events Among Hospitalized Children

  JAMA Network Open · 2025-05-30 · 20 citations

  articleOpen access

  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 <18 years) admissions to inpatient and intensive care units at 3 tertiary care academic hospitals. Data were analyzed from January 2024 to March 2025. Main Outcomes and Measures: The primary outcome was critical events, defined as invasive mechanical ventilation, administration of vasoactive medications, or death within 12 hours of an observation. Results: The cohort included 135 621 patients (mean [SD] age, 7 [6] years; 60 376 [44.5%] female). Patient age, hospital unit, vital signs, laboratory results, and prior comorbidities were used to derive a regression-based model, an extreme gradient-boosted machine (XGB) model, and 2 deep learning models. Data from 2 hospitals were used as a derivation cohort, while patients in the third hospital constituted the hold-out external test cohort. The XGB model was the best-performing machine learning model, outperforming 2 existing ward-focused models in terms of discrimination (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.

  PublisherDOI

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

  Journal of the American Medical Informatics Association · 2025-07-14 · 1 citations

  articleOpen access

  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.

  PublisherOA PDFDOI

• Disseminating child abuse clinical decision support among commercial electronic health records: Effects on clinical practice

  JAMIA Open · 2023 · 10 citations

  • Political Science
  • Computer Science
  • Psychology

  Objectives: The use of electronic health record (EHR)-embedded child abuse clinical decision support (CA-CDS) may help decrease morbidity from child maltreatment. We previously reported on the development of CA-CDS in Epic and Allscripts. The objective of this study was to implement CA-CDS into Epic and Allscripts and determine its effects on identification, evaluation, and reporting of suspected child maltreatment. Materials and Methods: After a preimplementation period, CA-CDS was implemented at University of Wisconsin (Epic) and Northwell Health (Allscripts). Providers were surveyed before the go-live and 4 months later. Outcomes included the proportion of children who triggered the CA-CDS system, had a positive Child Abuse Screen (CAS) and/or were reported to Child Protective Services (CPS). Results: At University of Wisconsin (UW), 3.5% of children in the implementation period triggered the system. The CAS was positive in 1.8% of children. The proportion of children reported to CPS increased from 0.6% to 0.9%. There was rapid uptake of the abuse order set.At Northwell Health (NW), 1.9% of children in the implementation period triggered the system. The CAS was positive in 1% of children. The child abuse order set was rarely used. Preimplementation, providers at both sites were similar in desire to have CA-CDS system and perception of CDS in general. After implementation, UW providers had a positive perception of the CA-CDS system, while NW providers had a negative perception. Discussion: CA-CDS was able to be implemented in 2 different EHRs with differing effects on clinical care and provider feedback. At UW, the site with higher uptake of the CA-CDS system, the proportion of children who triggered the system and the rate of positive CAS was similar to previous studies and there was an increase in the proportion of cases of suspected abuse identified as measured by reports to CPS. Our data demonstrate how local environment, end-users' opinions, and limitations in the EHR platform can impact the success of implementation. Conclusions: When disseminating CA-CDS into different hospital systems and different EHRs, it is critical to recognize how limitations in the functionality of the EHR can impact the success of implementation. The importance of collecting, interpreting, and responding to provider feedback is of critical importance particularly with CDS related to child maltreatment.

  PublisherOA PDFDOI

• Comprehensive Care Improvement for Oncologic Fever and Neutropenia from a Pediatric Emergency Department

  Pediatric Quality and Safety · 2021-02-12 · 2 citations

  articleOpen access1st authorCorresponding

  Introduction: Rapid time to antibiotics (TTA) for pediatric patients with fever and neutropenia in an emergency department decreases in-hospital mortality. Additionally, national guidelines recommend outpatient antibiotic management strategies for low-risk fever and neutropenia (LRFN). This study had two specific aims: (1) improve the percent of patients with suspected fever and neutropenia who receive antibiotics within 60 minutes of arrival from 55% to 90%, and (2) develop and operationalize a process for outpatient management of LRFN patients by October 2018. Methods: Using Lean methodologies, we implemented Plan-Do-Check-Act cycles focused on guideline development, electronic medical record reminders, order-set development, and a LRFN pathway as root causes for improvements. We used statistical process control charts to assess results. Results: The project conducted from July 2016 to October 2018 showed special cause improvement in December 2016 on a G-chart. Monthly Xbar-chart showed improvement in average TTA from 68.5 minutes to 42.5 minutes. A P-chart showed improvement in patients receiving antibiotics within 60 minutes, from 55% to 86.4%. A LRFN guideline and workflow was developed and implemented in October 2017. Conclusions: Implementation of guidelines, electronic medical record reminders, and order sets are useful tools to improve TTA for suspected fever and neutropenia. Utilizing more sensitive statistical process control charts early in projects with fewer patients can help recognize and guide process improvement. The development of workflows for outpatient management of LRFN may be possible, though it requires further study.

  PublisherDOI

• Electronic Health Record-Based Surveillance for Community Transmitted COVID-19 in the Emergency Department

  Western Journal of Emergency Medicine · 2020 · 8 citations

  • Medicine
  • Medical emergency
  • Emergency medicine

  INTRODUCTION: SARS-CoV-2, a novel coronavirus, manifests as a respiratory syndrome (COVID-19) and is the cause of an ongoing pandemic. The response to COVID-19 in the United States has been hampered by an overall lack of diagnostic testing capacity. To address uncertainty about ongoing levels of SARS-CoV-2 community transmission early in the pandemic, we aimed to develop a surveillance tool using readily available emergency department (ED) operations data extracted from the electronic health record (EHR). This involved optimizing the identification of acute respiratory infection (ARI)-related encounters and then comparing metrics for these encounters before and after the confirmation of SARS-CoV-2 community transmission. METHODS: We performed an observational study using operational EHR data from two Midwest EDs with a combined annual census of over 80,000. Data were collected three weeks before and after the first confirmed case of local SARS-CoV-2 community transmission. To optimize capture of ARI cases, we compared various metrics including chief complaint, discharge diagnoses, and ARI-related orders. Operational metrics for ARI cases, including volume, pathogen identification, and illness severity, were compared between the preand post-community transmission timeframes using chi-square tests of independence. RESULTS: Compared to our combined definition of ARI, chief complaint, discharge diagnoses, and isolation orders individually identified less than half of the cases. Respiratory pathogen testing was the top performing individual ARI definition but still only identified 72.2% of cases. From the pre to post periods, we observed significant increases in ED volumes due to ARI and ARI cases without identified pathogen. CONCLUSION: Certain methods for identifying ARI cases in the ED may be inadequate and multiple criteria should be used to optimize capture. In the absence of widely available SARS-CoV-2 testing, operational metrics for ARI-related encounters, especially the proportion of cases involving negative pathogen testing, are useful indicators for active surveillance of potential COVID-19 related ED visits.

  PublisherOA PDFDOI

• Dissemination of child abuse clinical decision support: Moving beyond a single electronic health record

  International Journal of Medical Informatics · 2020 · 18 citations

  • Computer Science
  • Artificial Intelligence
  • Medical emergency

  BACKGROUND: Child maltreatment is a leading cause of pediatric morbidity and mortality. We previously reported on development and implementation of a child abuse clinical decision support system (CA-CDSS) in the Cerner electronic health record (EHR). Our objective was to develop a CA-CDSS in two different EHRs. METHODS: Using the CA-CDSS in Cerner as a template, CA-CDSSs were developed for use in four hospitals in the Northwell Health system who use Allscripts and two hospitals in the University of Wisconsin health system who use Epic. Each system had a combination of triggers, alerts and child abuse-specific order sets. Usability evaluation was done prior to launch of the CA-CDSS. RESULTS: Over an 18-month period, a CA-CDSS was embedded into Epic and Allscripts at two hospital systems. The CA-CDSSs vary significantly from each other in terms of the type of triggers which were able to be used, the type of alert, the ability of the alert to link directly to child abuse-specific order sets and the order sets themselves. CONCLUSIONS: Dissemination of CA-CDSS from one EHR into the EHR in other health care systems is possible but time-consuming and needs to be adapted to the strengths and limitations of the specific EHR. Site-specific usability evaluation, buy-in of multiple stakeholder groups and significant information technology support are needed. These barriers limit scalability and widespread dissemination of CA-CDSS.

  PublisherDOI

• Dead in the air- The need to adapt to CoVID adaptations

  The American Journal of Emergency Medicine · 2020-07-23 · 1 citations

  articleOpen access1st authorCorresponding
  PublisherOA PDFDOI

• Rapid cycle testing drives improved communication and satisfaction using in-person survey

  BMJ Open Quality · 2019-09-01 · 1 citations

  articleOpen access1st authorCorresponding

  Background: Good communication with families improves safety and drives patient/family satisfaction. Rapid cycle improvement for the communication is difficult in our emergency department as current mailed surveys provide little and delayed data. We had two aims in this quality improvement study: (1) to increase proportion of families responding 'always' when asked if they received consistent communication from nurses and providers from 52% to 80% and (2) increase families reporting their visit as excellent, reflecting higher family satisfaction. Methods: Key drivers of the consistent communication were determined using the model for improvement. Interventions focused on interprovider communication and parental knowledge of communication processes. Eight Plan-Do-Study-Act ramps were conducted focusing on each of the key drivers, with 1-10 cycles per ramp. A five-question in-person survey was conducted at the time of disposition by the research assistants. Process and outcome measures were tracked on the statistical process control charts. Results: Mean percentage of families who reported always receiving consistent communication increased from 52% to 70% over 12 months. Additionally, families reporting their visit as 'excellent' increased from 62.5% to 75%. Using in-person surveys, weekly responses increased from 3 to 22. Conclusions: Iterative processes to improve interprovider communication and inform families about their care led to improvement in families' perceived communication consistency. Improved communication can lead to higher family satisfaction, most affecting those previously feeling neutral about their visit. In-person surveys can inform the real-time improvement efforts.

  PublisherOA PDFDOI

• Onychomadesis as a Late Complication of Hand-Foot-Mouth Disease

  Pediatric Emergency Care · 2017-11-01 · 9 citations

  article1st authorCorresponding

  Hand-foot-mouth disease is a viral illness frequently caused by enterovirus and coxsackievirus. Traditionally, this disease initially causes malaise, fever, and rash with vesicles in the mouth, as well as on the hands and feet. Occasionally, more severe presentations and late postinfectious sequelae occur, including onychomadesis, nail matrix arrest. We describe a series of 4 cases of onychomadesis and its evaluation following recent hand-foot-mouth disease during this current enteroviral season as a way to ensure appropriate clinician diagnosis and guidance.

  PublisherDOI

Frequent coauthors

• Joshua Ross

  University of Wisconsin–Madison

  4 shared

• Sundas Khan

  East Cheshire NHS Trust

  4 shared

• Thomas McGinn

  Baylor College of Medicine

  3 shared

• Isabel A. Barata

  2 shared

• Francesca Bullaro

  Cohen Children's Medical Center

  2 shared

• Rachel P. Berger

  University of Pittsburgh

  2 shared

• Emily Heineman

  University of Pittsburgh

  2 shared

• Brian Sharp

  University of Wisconsin–Madison

  2 shared

Awards & honors

• Safety Leadership Award, UW Health (2023)
• Physician Leadership Development Program Participant, UW Hea…
• Rising Star Clinical Excellence Award, UW Health (2021)
• Faculty Award for Excellence in Clinical Care, BerbeeWalsh D…
• Emergency Medicine Research Career Development, Medical Coll…