Phillip Joel Schulte

Background/Objective: To externally validate our previously published machine learning model for identifying hospitalized patients with community-acquired pneumonia (CAP) at low risk of respiratory deterioration or death using the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. Methods: This is a retrospective cohort study of adult patients (≥18 years) who were admitted with CAP and acute hypoxemic respiratory failure using the publicly available MIMIC-IV critical care dataset (Beth Israel Deaconess Medical Center, Boston, MA). We conducted an external validation study of a previously developed gradient boosting machine (GBM) model without recalibration using data available within the first 6 h of hospital admission. Results: For the primary composite outcome (need for advanced respiratory support [high flow nasal cannula (HFNC), non-invasive mechanical ventilation (NIMV), invasive mechanical ventilation (IMV)] or in-hospital death), the gradient boosting model demonstrated comparable performance in the derivation and external validation cohorts. The area under the receiver operating characteristic curve (AUC) was 0.713 in the Mayo cohort (n = 4379) and 0.689 in the MIMIC-IV cohort. Accuracy was 0.612 (95% confidence interval [CI], 0.595–0.628) versus 0.606 (95% CI 0.600–0.611), specificity 0.574 versus 0.523, sensitivity 0.723 versus 0.754, NPV 0.860 versus 0.842, and PPV 0.364 versus 0.401, respectively. For secondary outcomes, model discrimination was comparable between cohorts. The AUC for in-hospital mortality was 0.727 in the Mayo Cohort versus 0.733 in MIMIC-IV; for IMV, 0.736 versus 0.724; and for NIMV, 0.732 versus 0.708. Conclusions: Our machine learning algorithm demonstrated good discrimination and a high negative predictive value for identifying low-risk hospitalized CAP patients for respiratory deterioration or death in the external MIMIC-IV dataset, supporting its potential utility for prognostic enrichment in pneumonia clinical trials.

Background In-person interpreters are underused, contributing to health inequities among medically complex inpatients with non-English language preference (NELP). Objective To integrate a machine learning and informatics algorithm into clinical and language services workflows to increase the use of in-person interpreters among complex inpatients with NELP. Design We conducted a two-armed stepped-wedge cluster randomized trial from May 1, 2023, to June 21, 2024, at Mayo Clinic Rochester across 35 inpatient units, including hospital floors, ICUs, and emergency department. Units were randomized to transition between standard care and intervention phases every 60 days. Participants Adult inpatients ≥ 18 years with NELP in acute care settings. Intervention We used the algorithm to identify and order inpatients with medical complexity and NELP. Among patient admissions in the intervention arm, we shared the ordered list with language services, who did targeted outreach to bedside nurses via a secure chat message. The message stated that the patient would benefit from an in-person interpreter and provided contact details to reach language services. Main Measures Primary outcome was use of in-person interpreter services; secondary outcome assessed time from study entry to first interpreter visit. Key Results We enrolled 749 control and 672 intervention unique admissions. Time to receiving interpreter did not differ between groups (HR = 1.02, 95% C.I. (0.81, 1.29), p = 0.87). Conclusions We demonstrated an OR and C.I. close to > 1 for receiving an in-person interpreter among inpatients with NELP and complex care needs with the intervention. These findings are supportive of a future well-powered multicenter trial. Trial Registration ClinicalTrials.gov ID: NCT05860777. Graphical Abstract

Sepsis progression varies widely among intensive care unit (ICU) patients, making early and accurate trajectory prediction crucial. Traditional models focus on static assessments or binary outcomes, limiting their clinical utility. This study considers clinical trajectories of patients with sepsis for up-to 14 days hospital stay since ICU admission. Using data from 19,177 ICU patients, this study presents a hybrid CNN-GRU model that predicts four distinct clinical trajectories—Fast Decline, Fast Recovery, Slow Recovery, and Delayed Decline—by integrating sequential time-series data with static patient features. The model updates predictions every 2 hours, achieving AUCs of 0.90 (Fast Decline), 0.76 (Delayed Decline), 0.70 (Fast Recovery), and 0.55 (Slow Recovery) on validation data at hour 36. To enhance interpretability, we analyzed feature importance across both modalities. For sequential data, gradient-based saliency identified peripheral capillary oxygen saturation, lactate, positive end-expiratory pressure, temperature, Glasgow Coma Scale, respiratory rate, and mean arterial pressure as key drivers—aligned with known sepsis indicators. For static features, permutation-based analysis highlighted age, ICU stay duration before prediction, comorbidity burden, and pulmonary disease history as top contributors. These insights confirm the model's physiological relevance and support its potential for real-time, explainable sepsis trajectory prediction and risk stratification.

Background: Transfusion-related acute lung injury (TRALI) and transfusion-associated circulatory overload (TACO) are leading causes of transfusion-related morbidity and mortality. Soluble factors in erythrocyte supernatant may increase risk for these complications. The authors hypothesized that point-of-care allogeneic erythrocyte washing may be an effective intervention to mitigate elevations in soluble factors as well as physiologic responses associated with transfusion-associated respiratory complications in the setting of cardiac surgery. Methods: This is a two-center, nonblinded, randomized clinical trial evaluating point-of-care washed versus standard issue allogeneic erythrocyte transfusions administered during or on the day of cardiac surgery. The primary analysis was performed via modified intention to treat. The primary outcomes assessed were changes in intermediate markers of lung injury as well as cardiopulmonary physiologic responses to erythrocyte transfusion. Secondary outcomes included the duration of intensive care unit and hospital stay, durations of mechanical ventilation and oxygen supplementation, presence of TRALI or TACO, and mortality. Results: Among 154 analyzed patients (81 washed, 73 standard issue), the median age was 66 yr, and 77 (50.0%) were women. The median (interquartile range) number of allogeneic erythrocyte units transfused on the day of surgery was 3.0 (2.0 to 5.0) in the washed erythrocyte group and 3.0 (2.0 to 4.0) in the standard issue group (P = 0.13). No between-group differences were identified in any of the assessed recipient lung injury biomarkers (all P values > adjusted alpha). Durations of intensive care unit stay (median [interquartile range], 3.0 [2.0 to 5.0] vs. 3.0 [2.0 to 4.0] days; P = 0.117) and hospital length of stay (12.0 [9.0 to 17.0] vs. 12.0 [9.0 to 17.0] days; P = 0.801) were similar, as were the number of ventilator-free days at day 28 (27.0 [27.0 to 27.0] vs. 27.0 [26.0 to 27.0]; P = 0.699) and oxygen-free days at day 28 (24.0 [19.0 to 26.0] vs. 24.0 [22.0 to 26.0]; P = 0.400). No significant differences were noted in mortality rate or in incidence rates for TRALI, TACO, and acute kidney injury. Conclusions: Among patients undergoing cardiovascular surgery with high risk of erythrocyte transfusion, point-of-care washing of allogeneic erythrocyte transfusions did not mitigate changes in intermediate markers of lung injury or cardiopulmonary physiologic responses to erythrocyte transfusion and was not associated with improved clinical outcomes.

Importance: Emerging literature has associated the use of anticholinergic medications to cognitive decline. Objective: The aim of this study was to evaluate the association of overactive bladder medications on cognitive function with prospective longitudinal cognitive assessments. Study design: A population-based cohort of individuals 50 years and older who had serial validated cognitive assessment, in accordance with the Mayo Clinic Study of Aging, was evaluated from October 2004 through December 2021. Anticholinergic overactive bladder medications were grouped by traditional anticholinergic medications and central nervous system (CNS)- sparing anticholinergic medications and compared to no medication exposure. A linear mixed effects model with time-dependent exposures evaluated the association between overactive bladder anticholinergic medication exposure and subsequent trajectories of cognitive z-scores. Results: We included 5,872 participants with a median follow-up of 6.4 years. Four hundred forty-three were exposed to traditional anticholinergic medications, 60 to CNS-sparing medications, and 5,369 had no exposure. On multivariable analyses, exposure to any anticholinergic overactive bladder medication was significantly associated with deterioration in longitudinal cognitive scores in the language and attention assessments compared to the control cohort. Traditional anticholinergic medication exposure was associated with worse attention scores than nonexposed participants. Exposure to CNS-sparing anticholinergic medications was associated with a deterioration in the language domain compared to those unexposed. Among women, traditional anticholinergic medication exposure was associated with worse global and visuospatial scores than nonexposed participants, but this association was not identified in the CNS-sparing group. Conclusion: Exposure to anticholinergic overactive bladder medications was associated with small but significantly worse decline in cognitive scoring in the language and attention domains when compared to nonexposed individuals.

Current prognostic tools for pneumonia predominantly focus on mortality, often neglecting other crucial outcomes such as the need for advanced respiratory support. The objective of this study was to develop and validate a tool that predicts the early risk of non-occurrence of respiratory deterioration or mortality. We conducted a single-center, retrospective cohort study involving hospitalized adult patients with community-acquired pneumonia (CAP) and acute hypoxic respiratory failure from January 2009 to December 2019 (n = 4379). We employed the gradient boosting machine (GBM) learning to create a model that estimates the likelihood of patients requiring advanced respiratory support (high flow nasal cannula [HFNC], non-invasive mechanical ventilation [NIMV], and invasive mechanical ventilation [IMV]) or facing mortality during hospitalization. This model utilized readily available data including demographic, physiologic, and laboratory data, sourced from electronic health records and obtained within the first six hours of admission. Out of the cohort, 890 patients (25.2%) either required advanced respiratory support or died during their hospital stay. Our predictive model displayed superior discrimination and higher sensitivity (cross-validation C-statistic = 0.71; specificity = 0.56; sensitivity = 0.72) compared to the pneumonia severity index (PSI) (C-statistic = 0.65; specificity = 0.91; sensitivity = 0.24; P value < 0.001), while maintaining a negative predictive value (NPV) of approximately 0.85. These data demonstrate that our machine learning model predicted the non-occurrence of respiratory deterioration or mortality among hospitalized CAP patients more accurately than the PSI. The enhanced sensitivity of this model holds potential for reliably excluding low-risk patients from pneumonia clinical trials.

Research articles in the clinical and translational science literature commonly use quantitative data to inform evaluation of interventions, learn about the etiology of disease, or develop methods for diagnostic testing or risk prediction of future events. The peer review process must evaluate the methodology used therein, including use of quantitative statistical methods. In this manuscript, we provide guidance for peer reviewers tasked with assessing quantitative methodology, intended to complement guidelines and recommendations that exist for manuscript authors. We describe components of clinical and translational science research manuscripts that require assessment including study design and hypothesis evaluation, sampling and data acquisition, interventions (for studies that include an intervention), measurement of data, statistical analysis methods, presentation of the study results, and interpretation of the study results. For each component, we describe what reviewers should look for and assess; how reviewers should provide helpful comments for fixable errors or omissions; and how reviewers should communicate uncorrectable and irreparable errors. We then discuss the critical concepts of transparency and acceptance/revision guidelines when communicating with responsible journal editors.

Background: Innovative restructuring of cardiac rehabilitation (CR) delivery remains critical to reduce barriers and improve access to diverse populations. Destination Cardiac Rehab is a novel virtual world technology‐based CR program delivered through the virtual world platform, Second Life, which previously demonstrated high acceptability as an extension of traditional center‐based CR. This study aims to evaluate efficacy and adherence of the virtual world–based CR program compared with center‐based CR within a community‐informed, implementation science framework. Methods: Using a noninferiority, hybrid type 1 effectiveness‐implementation, randomized controlled trial, 150 patients with an eligible cardiovascular event will be recruited from 6 geographically diverse CR centers across the United States. Participants will be randomized 1:1 to either the 12‐week Destination Cardiac Rehab or the center‐based CR control groups. The primary efficacy outcome is a composite cardiovascular health score based on the American Heart Association Life's Essential 8 at 3 and 6 months. Adherence outcomes include CR session attendance and participation in exercise sessions. A diverse patient/caregiver/stakeholder advisory board was assembled to guide recruitment, implementation, and dissemination plans and to contextualize study findings. The institutional review board–approved randomized controlled trial will enroll and randomize patients to the intervention (or control group) in 3 consecutive waves/year over 3 years. The results will be published at data collection and analyses completion. Conclusions: The Destination Cardiac Rehab randomized controlled trial tests an innovative and potentially scalable model to enhance CR participation and advance health equity. Our findings will inform the use of effective virtual CR programs to expand equitable access to diverse patient populations. Registration: URL: https://www.clinicaltrials.gov ; Unique identifier: NCT05897710.

An abstract is unavailable. This article is available as a PDF only.