Machine learning for adverse event prediction in outpatient parenteral antimicrobial therapy: a scoping review.

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.

Rubella virus-associated necrotizing granulomatous inflammation with extensive eyelid, ocular, and orbital involvement.

We present a case of cutaneous granulomatous disease associated with rubella virus in a 4-year-old girl without an identifiable immunodeficiency. In this case, a combination of anti-inflammatory, anti-viral, and anti-neutrophil therapies successfully treated vision-threatening eyelid, conjunctival, scleral, and orbital inflammation.

Outcomes in Pediatric Central Line-associated Bloodstream Infections Treated With Antimicrobial Locks: A 14-Year Retrospective Analysis.

BACKGROUND: Central line-associated bloodstream infections (CLABSI) are the most common healthcare-associated infection in children. Antimicrobial lock therapy (ALT) is a recommended component of CLABSI treatment; however, studies characterizing pediatric ALT-based CLABSI therapy are few. METHODS: All CLABSI treated with ALT at a single tertiary-care pediatric hospital for years 2004-2018 were identified by retrospective chart review and demographic, clinical, microbiological and pharmacy data were extracted. Three clinical outcomes were defined: CLABSI clearance, central venous catheter (CVC) removal and infection recurrence. Factors associated with these outcomes were examined and patterns in pathogen occurrence were described. RESULTS: During the study period, 1188 CLABSI treated with ALT were identified. In all, 969 (85%) CLABSI initially cleared. In the CLABSI with initial clearance, CVCs were retained a median of 63 days after infection onset, and 20% recurred. Skin-colonizing microbes ( Staphylococcus aureus , coagulase-negative Staphylococci) were more likely to produce monomicrobial infections, whereas bowel-colonizing microbes ( Escherichia coli , Enterococci) were more associated with polymicrobial infections. In a multivariate model, infection history, resistant pathogens and CVC type were strongly correlated with infection recurrence, whereas bone marrow transplant status, neutrophil count and line type were correlated with CVC removal. CONCLUSIONS: Optimal therapeutic approaches to pediatric CLABSI remain uncertain, including the utility of ALT. Prior reports of ALT use in children have been smaller in size and largely described hematology-oncology or hemodialysis populations. We describe trends in pediatric CLABSI, including efficacy among solid organ transplant recipients and children with short-gut syndrome. These data provide guidance on selecting CLABSI for optimal ALT use.

Predicting outcomes in central venous catheter salvage in pediatric central line-associated bloodstream infection.

OBJECTIVE: Central line-associated bloodstream infections (CLABSIs) are a common, costly, and hazardous healthcare-associated infection in children. In children in whom continued access is critical, salvage of infected central venous catheters (CVCs) with antimicrobial lock therapy is an alternative to removal and replacement of the CVC. However, the success of CVC salvage is uncertain, and when it fails the catheter has to be removed and replaced. We describe a machine learning approach to predict individual outcomes in CVC salvage that can aid the clinician in the decision to attempt salvage. MATERIALS AND METHODS: Over a 14-year period, 969 pediatric CLABSIs were identified in electronic health records. We used 164 potential predictors to derive 4 types of machine learning models to predict 2 failed salvage outcomes, infection recurrence and CVC removal, at 10 time points between 7 days and 1 year from infection onset. RESULTS: The area under the receiver-operating characteristic curve varied from 0.56 to 0.83, and key predictors varied over time. The infection recurrence model performed better than the CVC removal model did. CONCLUSIONS: Machine learning-based outcome prediction can inform clinical decision making for children. We developed and evaluated several models to predict clinically relevant outcomes in the context of CVC salvage in pediatric CLABSI and illustrate the variability of predictors over time.

Increase in Invasive Group A Streptococcal Disease and Emergence of Mucoid Strains in a Pediatric Population: February-June 2017.

BACKGROUND: Infection with group A Streptococcus (GAS) can cause severe systemic and locally invasive disease. Invasive group A streptococcal (iGAS) disease incidence varies both seasonally and year-to-year, and it may exhibit clustered outbreaks. We observed an upswing in iGAS cases at a tertiary care Children's Hospital, prompting further characterization of local iGAS disease. METHODS: Cases of iGAS disease were abstracted from the medical record by manual chart review of all positive screening tests and cultures for GAS over a 4-year span. Incidence rates per 1000 hospital admissions and per 100 positive GAS tests were calculated and compared. Selected isolates were further characterized by whole-genome sequencing. RESULTS: Significant year-to-year differences in per-admission iGAS incidence rate were observed in February and June, although per-positive test incidence rates were not significantly different. Whole-genome sequencing revealed 2 dominant serotypes-emm3 and emm6-with high rates of mucoid phenotype and systemic bacteremia. CONCLUSIONS: We document a significant but transient increase in iGAS disease incidence in 2 months of 2017. Genome sequencing revealed 2 dominant serotypes associated with mucoid phenotypes and severe disease, highlighting the dynamic nature of iGAS disease pattern.

Serious Bacterial Infections in Neonates Presenting Afebrile With History of Fever.

BACKGROUND: Infants ≤28 days of age with fever are frequently hospitalized while undergoing infectious evaluation. We assessed differences in rates of serious bacterial infection (SBI; bacteremia, bacterial meningitis, urinary tract infection) and invasive bacterial infection (IBI; bacteremia, bacterial meningitis) among the following neonates: (1) febrile at presentation (FP), (2) afebrile with history of fever without subsequent fever during hospitalization, and (3) afebrile with history of fever with subsequent fever during hospitalization. METHODS: We performed a single-center retrospective study of neonates evaluated for SBI during emergency department evaluation between January 1, 2006, and December 31, 2017. Patients were categorized into FP, afebrile with no subsequent fever (ANF), and afebrile with subsequent fever (ASF) groups. We compared rates of SBI and IBI between groups using logistic regression and assessed time to fever development using time-to-event analysis. RESULTS: Of 931 neonates, 278 (29.9%) were in the ANF group, 93 (10.0%) were in the ASF group, and 560 (60.2%) were in the FP group. Odds of SBI in neonates ANF were 0.42 (95% confidence interval [CI] 0.23-0.79) compared with infants FP, although differences in IBI were not statistically significant (0.52, 95% CI 0.19-1.51). In infants ASF, median time to fever was 5.6 hours (interquartile range, 3.1-11.4). Infants ASF had higher odds of SBI compared to infants FP (odds ratio 1.93, 95% CI 1.07-3.50). CONCLUSIONS: Neonates with history of fever who remain afebrile during hospitalization may have lower odds for SBI and be candidates for early discharge after an observation period.

Immature neutrophils in young febrile infants.

BACKGROUND: To describe the diagnostic value of the absolute band count (ABC) and ratio of immature to total neutrophils (I:T) for invasive bacterial infections (IBIs; bacterial meningitis and bacteraemia) among young febrile infants. METHODS: We performed a cross-sectional study in a paediatric emergency department of febrile infants ≤60 days over 12 years to evaluate the accuracy of the ABC and I:T for IBI. RESULTS: Of 2930 included patients, 75 (2.6%) had IBIs. The area under the curve (AUC; 95% CI) for ABC was 0.69 (0.62 to 0.76) with sensitivity 0.27 (0.17 to 0.38) and specificity 0.94 (0.93 to 0.95) at cutoff ≥1500 cells/µL. The AUC for I:T was 0.65 (0.59 to 0.72) with sensitivity 0.29 (0.19 to 0.41) and specificity 0.88 (0.87 to 0.89) at cutoff ≥0.2. Only the ABC in infants 29-60 days was minimally accurate. CONCLUSION: The ABC and I:T were generally inaccurate for detecting IBI in febrile infants. Guidelines without these parameters may be better for risk assessment.

Factors associated with serious bacterial infections in infants ≤60 days with hypothermia in the emergency department.

BACKGROUND: We sought to investigate risk factors for serious bacterial infection (SBI: bacterial meningitis, bacteremia, and urinary tract infection [UTI]) among infants ≤60 days of age presenting to the emergency department (ED) with hypothermia (temperature < 36 °C). METHODS: We performed a single center study over a 12-year period including all patients ≤60 days old with hypothermia, excluding patients who did not receive a blood culture and patients who received antibiotics prior to culture acquisition. The primary outcome was SBI. Secondary outcomes were mortality and herpes simplex infection. We performed multivariable logistic regression to identify risk factors for primary outcomes reporting adjusted odds ratios with 95% confidence intervals (aOR, 95% CI). RESULTS: 360 infants were identified. 10/360 (2.8%) had an SBI. All episodes of SBI occurred in infants ≤28 days of age. Two patients had meningitis, two had meningitis with bacteremia, one had isolated bacteremia, and five had UTI. Associated diagnoses included prematurity (46.9%), hyperbilirubinemia (28.3%) and dehydration (14.7%). In multivariable analysis, presentation at 15-28 days (7.60, 1.81-31.86; p = 0.005) compared to 0-14 days, higher absolute neutrophil count (1.25, 1.04-1.50; p = 0.015) and lower platelet count (0.99, 0.99-1.00; p = 0.046) were associated with SBI. Three patients without SBI died during or soon after their hospitalization. One patient had positive testing for herpes simplex. CONCLUSION: In this cohort of hypothermic infants, 2.8% had a SBI. Age of presentation, ANC, and lower platelet count were associated with serious infections. Hypothermic infants presenting to the ED carry significant morbidity and require prospective study to better risk-stratify this population.

Experiential Communications Curriculum to Improve Resident Preparedness When Responding to Discriminatory Comments in the Workplace.

BACKGROUND: Patients and families can make discriminatory comments leading to physician distress. Residents receive little training in appropriate responses to such comments and may be ill equipped to respond to intolerance without alienating the individual(s) making the comments. OBJECTIVE: We assessed whether a simulated curriculum would enhance pediatrics residents' ability to effectively respond to discriminatory comments. METHODS: In the 2016-2017 academic year, we modified an existing communication skills curriculum for senior pediatrics residents. Residents engaged a simulated parent who used discriminatory speech in 4 scenarios, followed by a group debriefing. We conducted anonymous surveys to assess residents' preparedness to respond to these comments before and immediately following participation and examined their experience with discriminatory comments in the workplace. RESULTS: The majority of residents reported prior experience with discriminatory comments (32 of 45 [71%] witnessed such comments, and 27 of 48 [56%] were targeted by such comments), most often regarding age, race, and ethnicity. Mean precourse scores ranged from 2.1 to 3.1 (on a 5-point scale) regarding ability to engage in a firm yet respectful dialogue, to reference the hospital code of conduct, to coach a learner to respond, and to facilitate a team debrief. Mean postcourse scores improved significantly for these questions (range 3.8-4.1). The greatest improvement was in referencing the code of conduct (2.1 versus 4.0, P < .001). CONCLUSIONS: Immediately after participating in simulation, pediatrics residents reported a significant improvement in self-reported readiness to respond to discriminatory comments made by a parent and reported the simulation experience was beneficial.

Mapping the shapes of phylogenetic trees from human and zoonotic RNA viruses.

A phylogeny is a tree-based model of common ancestry that is an indispensable tool for studying biological variation. Phylogenies play a special role in the study of rapidly evolving populations such as viruses, where the proliferation of lineages is constantly being shaped by the mode of virus transmission, by adaptation to immune systems, and by patterns of human migration and contact. These processes may leave an imprint on the shapes of virus phylogenies that can be extracted for comparative study; however, tree shapes are intrinsically difficult to quantify. Here we present a comprehensive study of phylogenies reconstructed from 38 different RNA viruses from 12 taxonomic families that are associated with human pathologies. To accomplish this, we have developed a new procedure for studying phylogenetic tree shapes based on the 'kernel trick', a technique that maps complex objects into a statistically convenient space. We show that our kernel method outperforms nine different tree balance statistics at correctly classifying phylogenies that were simulated under different evolutionary scenarios. Using the kernel method, we observe patterns in the distribution of RNA virus phylogenies in this space that reflect modes of transmission and pathogenesis. For example, viruses that can establish persistent chronic infections (such as HIV and hepatitis C virus) form a distinct cluster. Although the visibly 'star-like' shape characteristic of trees from these viruses has been well-documented, we show that established methods for quantifying tree shape fail to distinguish these trees from those of other viruses. The kernel approach presented here potentially represents an important new tool for characterizing the evolution and epidemiology of RNA viruses.

T cells contain an RNase-insensitive inhibitor of APOBEC3G deaminase activity.

The deoxycytidine deaminase APOBEC3G (A3G) is expressed in human T cells and inhibits HIV-1 replication. When transfected into A3G-deficient epithelial cell lines, A3G induces catastrophic hypermutation by deaminating the HIV-1 genome. Interestingly, studies suggest that endogenous A3G in T cells induces less hypermutation than would be expected. However, to date, the specific deaminase activity of endogenous A3G in human CD4+ T cells has not been examined directly. Here, we compared deaminase activity of endogenous and exogenous A3G in various human cell lines using a standard assay and a novel, quantitative, high-throughput assay. Exogenous A3G in epithelial cell lysates displayed deaminase activity only following RNase treatment, as expected given that A3G is known to form an enzymatically inactive RNA-containing complex. Surprisingly, comparable amounts of endogenous A3G from T cell lines or from resting or activated primary CD4+ T cells exhibited minimal deaminase activity, despite RNase treatment. Specific deaminase activity of endogenous A3G in H9, CEM, and other T cell lines was up to 36-fold lower than specific activity of exogenous A3G in epithelial-derived cell lines. Furthermore, RNase-treated T cell lysates conferred a dose-dependent inhibition to epithelial cell lysates expressing enzymatically active A3G. These studies suggest that T cells, unlike epithelial-derived cell lines, express an unidentified RNase-resistant factor that inhibits A3G deaminase activity. This factor could be responsible for reduced levels of hypermutation in T cells, and its identification and blockade could offer a means for increasing antiretroviral intrinsic immunity of T cells.