Challener, Douglas

Abstract Background Native vertebral osteomyelitis (NVO) is a life-threatening spinal infection with rising incidence and significant morbidity. Despite its growing burden, long-term data on clinical characteristics, management trends, and outcomes remain limited. Methods We conducted a 26-year multicenter retrospective cohort study of adults (≥18 years) diagnosed with NVO at Mayo Clinic sites between 1999 and 2024. Demographic, microbiologic, treatment, and outcome data were analyzed across five time periods. Predictors of treatment failure were assessed using a multivariable competing risk model. Results Among 1255 patients (median age 67; 66% male), lumbosacral involvement was most common (65%), and 21% had multilevel involvement. Pathogens were identified in 77%, most commonly Staphylococcus aureus (49%; Methicillin-susceptible S. aureus 37%, methicillin-resistant S. aureus 13%). Over time from 1999–2004 to 2020–2024, Gram-negative bacilli increased from 6% to 14% (P = .048). Comorbidities including chronic kidney disease (10% to 21%), active chemotherapy (6% to 11%), and immunosuppression (8% to 17%) increased significantly. Additionally, 1-year treatment failure declined (16% to 10%). In multivariable analysis, diabetes mellitus (subdistribution hazard ratio [sHR] 1.92, 95% CI 1.18–3.13) and multilevel involvement (sHR 1.67, 95% CI 1.17–2.38) were associated with increased incidence of treatment failure, while concurrent infections (sHR 0.57, 95% CI 0.37–0.87) and higher Charlson Comorbidity Index (sHR 0.62, 95% CI 0.43–0.90) were associated with lower failure. Conclusions This large multicenter cohort highlights increasing host complexity, shifting microbiology, and predictors of failure, emphasizing the importance of early risk stratification and tailored strategies, such as multidisciplinary evaluation and close follow-up of high-risk patients to improve outcomes.

Abstract Background Hospital discharge summaries are critical for ensuring continuity of care, but their quality often varies. Large language models (LLMs) have the potential to standardize and enhance the efficiency of this documentation process. Objectives To evaluate the quality of hospital discharge summaries created by an LLM‐based hospital course drafting tool created by Epic Systems compared with human‐written summaries. Methods Retrospective study at a single tertiary‐care institution in 2024. The cohort included 100 adult hospitalizations lasting >72 h across medical and surgical dismissing services. No interventions were performed. Summaries (LLM‐generated vs. human‐written) were independently reviewed using a standardized rubric covering nine domains (e.g., comprehensiveness, clarity, relevance). Scores were normalized and compared. Readability was assessed using Flesch Reading Ease. Results LLM‐generated summaries outperformed human‐written summaries across all criteria ( p < .05), with the greatest difference observed in comprehensiveness (LLM median 0.62 vs. human −0.23). Human‐written summaries from surgical services scored lower than those from medical services, but LLM performance was consistent across both. Human summaries had higher Flesch Reading Ease scores (33.11 vs. 26.2; p < .05), reflecting simpler language. Conclusions LLM‐generated summaries demonstrated superior quality, consistency, and clinical utility compared with human‐written summaries, highlighting their potential to improve documentation efficiency and standardization.

Abstract Background We examined how the COVID-19 pandemic influenced influenza vaccination, healthcare engagement, and patient satisfaction. Methods We retrospectively analyzed influenza vaccination records and patient experience surveys of adult primary care patients in southeastern Minnesota during two 3-year phases: Prepandemic (1 January 2017–28 February 2020) and Pandemic-Plus (1 March 2020–31 December 2023). Vaccination status was defined as “always vaccinated” (AV), “never vaccinated” (NV), or “intermittently vaccinated” (IV) for seasonal influenza. Results During the Pandemic-Plus phase, 7.0% (n = 3556) of the Prepandemic AV and 48.7% (n = 16 710) of the Prepandemic IV patients never received another influenza vaccine. Compared to AV, IV and NV patients were more likely to reside in areas with greater socioeconomic deprivation (odds ratio [OR], 1.58 [95% confidence interval {CI}, 1.53–1.62] and 1.99 [1.94–2.05], respectively), have a high school education or less (2.86 [2.74–2.98] and 3.38 [3.23–3.53]), and report healthcare disengagement (1.59 [1.55–1.64] and 4.21 [4.09–4.33]). After adjusting for Area Deprivation Index and medical comorbidities, healthcare disengagement increased among the NV versus AV between phases (3.33 [3.24–3.41] and 4.23 [4.10–4.35]). In a subgroup analysis those with severe comorbidities were less likely AV (NV vs AV: 1.21 [1.14–1.27]) and more dissatisfied with care (NV vs AV: 1.25 [1.18–1.33]). Conclusions The COVID-19 pandemic altered vaccination behaviors and healthcare satisfaction, especially among those at high risk of developing influenza-related complications. Medical providers and public health officials should be aware of factors associated with vaccine refusal to better target interventions.

Background Electronic health records (EHRs) serve as a critical tool for clinical documentation, yet their complexity has raised concerns about readability and usability for healthcare providers and patients. Readability scores, such as the Flesch-Kincaid (FK) Grade Level, offer a quantitative method to assess documentation clarity. In this study, the authors evaluated the impact of an EHR transition on the readability of clinical notes across multiple medical specialties. Methods All clinical notes recorded between 2016 and 2019 at a large healthcare organization were evaluated across specialties. The organization transitioned EHR systems during the study period, in May 2018. FK Grade Level scores were calculated for each clinical consultation note and stratified by specialty and EHR. Results Across 10 medical specialties, 630,246 clinical notes were evaluated. The median FK grade levels from notes composed in the new EHR were similar to those composed in the previous EHR (8.228 vs 8.240; p=0.58). The variation in FK grade level as measured by standard deviation was lower in the new EHR compared with the old EHR (1.405 vs 1.764; p <0.01). Conclusions Calculation of FK grade level of clinical notes is a feasible method to assess clinical note readability. Study results showed that clinical documentation readability may not be strongly associated with underlying EHR. Migrating to a new EHR alone should not be expected to improve readability of clinical notes.

Abstract Background We investigated the relationship between cumulative, historical antibiotic therapy and the presence of ceftriaxone-resistant (CRO-R) Enterobacterales bloodstream infections (BSI) from urinary sources. Methods Adult patients established with primary care at Mayo Clinic with a first episode of monomicrobial Escherichia coli, Klebsiella pneumoniae, Klebsiella oxytoca, or Proteus mirabilis BSI from a urinary source from May 2019 to May 2023 were included. Antibiotic therapy over the year prior to BSI was quantified as days of therapy (DOT), cumulative modified Antibiotic Spectrum Index (mASI), and cumulative Days of Antibiotic Spectrum Coverage (DASC) using prescription data. Patients with CRO-R urinary pathogens prior to antibiotic therapy assessment were excluded. Results In total, 721 adults had monomicrobial gram-negative BSI, including 70 patients (9.7%) with a CRO-R pathogen and 651 patients (90.3%) with a ceftriaxone-susceptible (CRO-S) pathogen. In the 12 months prior to BSI, levels of antibiotic DOT, mASI, and DASC were significantly greater among patients with CRO-R BSI than CRO-S BSI (P &lt; .001). The median time from blood culture collection to appropriate antibiotic therapy was significantly longer in the CRO-R group (24 hours, interquartile range [IQR] 16–48) than the CRO-S group (1 hour, IQR 0–2) (P &lt; .001). Increasing cumulative antibiotic therapy prior to BSI, as quantified by the mASI, was significantly associated with increased risk of 1-year CDI after BSI (P = .027). Conclusions Patients with first episode CRO-R Enterobacterales BSI had significantly greater prior antibiotic exposure than patients with CRO-S BSI and experienced significant delay of appropriate therapy. Routine quantitation and reporting of prior antibiotic therapy may improve empiric prescribing.

Artificial intelligence (AI) has emerged as a promising adjunct in the diagnosis and management of infective endocarditis (IE), a disease characterized by diagnostic complexity and significant morbidity. Machine learning (ML) models such as SABIER and SYSUPMIE have demonstrated strong predictive accuracy for early IE diagnosis, embolic risk stratification, and postoperative mortality, surpassing traditional clinical scoring systems. In imaging, AI-enhanced echocardiography and advanced modalities like FDG-PET/CT offer improved sensitivity, specificity, and reduced inter-observer variability, potentially transforming clinical decision making. Additionally, AI-powered microbiological techniques, including MALDI-TOF mass spectrometry combined with ML and neural network-based metagenomic classifiers, show promise in rapidly identifying pathogens and predicting antimicrobial resistance. Despite encouraging early results, widespread adoption faces barriers, including data limitations, interpretability issues, ethical concerns, and the need for robust validation. Future directions include leveraging generative AI as clinical consultative tools, provided their capabilities and limitations are carefully managed. Ultimately, collaborative efforts addressing these challenges could transform IE care, enhancing diagnostic accuracy, clinical outcomes, and patient safety.

Abstract Background There is a lack of comprehensive, large cohort studies investigating the predictors for severity and hospitalization in human granulocytic anaplasmosis (HGA). Methods We conducted a retrospective cohort study including all cases of HGA identified by positive blood Anaplasma polymerase chain reaction (PCR) at the Mayo Clinic Laboratory, Rochester between 2011 and 2021. Multivariable logistic regression was performed to evaluate risk factors associated with hospitalization. Results A total of 465 cases were identified. Of those, 67% (n = 312) were managed in the outpatient setting. Hospitalized patients (n = 153, 33%) were more likely to be older (median age of 71 vs 61; P ≤ .001) and immunocompromised (17% vs 7%; P ≤ .001). The triad of leukopenia, thrombocytopenia, and transaminitis was observed in 24% of the cases, but was associated with risk of hospitalization (odds ratio [OR] 1.78, 95% confidence interval [CI]: 1.05–3.03; P ≤ .033). The presence of a coinfection did not impact mortality or hospitalization. Conclusions The risk factors for hospitalization in patients with HGA include altered mental status, higher absolute neutrophil count (ANC), advanced age, comorbidities, and immunosuppression. The classic hematological and biochemical profile may be absent in the majority of cases. Co-testing may be of higher benefit in select cases.

Background Patients enrolling in hospice care are asked their preferences for antibiotic treatment. There is little information regarding which patients are more likely to receive antibiotics. To better inform discussions, we sought to characterize the use of antibiotics in the outpatient hospice setting. Methods We performed a retrospective review of patients enrolled in outpatient hospice within the Mayo Clinic Health System from 1/1/2017 through 1/1/2023. We calculated what percent of patients received antibiotics based on Hospice Qualifying Condition (HQC), age at enrollment, sex, survival time, and Charlson Comorbidity Index and calculated adjusted odds ratios (aOR). We documented which antibiotics were prescribed based on HQC. Results Of 6452 patients identified, 1259 (19.5%) received antibiotic prescriptions. Cephalosporins were the most common class of antibiotics prescribed (22.8% of antibiotics prescribed), followed by fluoroquinolones (20.3%) and penicillin derivatives (14.9%). Patients with Pulmonary HQCs were most likely to receive antibiotics (28.6% aOR 1.85 [1.51-2.25]), those with neurologic HQCs were least likely (14.9% aOR 0.66 [0.53-0.83]). There was no difference of age for those receiving antibiotics (80.7 yr) vs those not receiving (80.3 yr [ P = 0.25]), or for sex (male vs female aOR 1.10 [0.96-1.26). 4.6% of those surviving <7d received antibiotics (aOR .22 [0.17-0.29]), compared to 47.7% of those surviving >6mo (aOR 4.46 [3.50-5.69]). Discussion 19.5% of patients will receive antibiotics during their hospice course, with more frequent prescriptions in those surviving longer periods and those enrolled for pulmonary conditions. It is important to clarify patient preferences regarding antimicrobial utilization at the time of hospice enrollment.

Abstract Objective This study aimed to conduct a scoping review of machine learning (ML) techniques in outpatient parenteral antimicrobial therapy (OPAT) for predicting adverse outcomes and to evaluate their validation, implementation and potential barriers to adoption. Materials and methods This scoping review included studies applying ML algorithms to adult OPAT patients, covering techniques from logistic regression to neural networks. Outcomes considered were medication intolerance, toxicity, catheter complications, hospital readmission and patient deterioration. A comprehensive search was conducted across databases including Cochrane Central, Cochrane Reviews, Embase, Ovid MEDLINE and Scopus, from 1 January 2000 to 1 January 2024. Results Thirty-two studies met the inclusion criteria, with the majority being single-centre experiences primarily from North America. Most studies focused on developing new ML models to predict outcomes such as hospital readmissions and medication-related complications. However, there was very little reporting on the performance characteristics of these models, such as specificity, sensitivity and C-statistics. There was a lack of multi-centre or cross-centre validation, limiting generalizability. Few studies advanced beyond traditional logistic regression models, and integration into clinical practice remains limited. Discussion ML shows promise for enhancing OPAT outcomes by predicting adverse events and enabling pre-emptive interventions. Despite this potential, significant gaps exist in development, validation and practical implementation. Barriers include the need for representative data sets and broadly applicable, validated models. Conclusion Future research should address these barriers to fully leverage ML’s potential in optimizing OPAT care and patient safety. Models must deliver timely, accurate and actionable insights to improve adverse event prediction and prevention in OPAT settings.

Abstract Cardiac implantable electronic device infection in the context of corynebacterial bloodstream infection (BSI) remains poorly understood. From 2012 to 2023 at Mayo Clinic, 4 of 12 patients with corynebacterial BSI had cardiac implantable electronic device infection: 1 patient was diagnosed during a relapsing BSI episode. Undefined source, persistent BSI, and the presence of a prosthetic cardiac valve were common characteristics.