Development and Validation of a Machine Learning Model for Early Detection of Untreated Infection.

BACKGROUND: Early diagnostic uncertainty for infection causes delays in antibiotic administration in infected patients and unnecessary antibiotic administration in noninfected patients. OBJECTIVE: To develop a machine learning model for the early detection of untreated infection (eDENTIFI), with the presence of infection determined by clinician chart review. DERIVATION COHORT: Three thousand three hundred fifty-seven adult patients hospitalized between 2006 and 2018 at two health systems in Illinois, United States. VALIDATION COHORT: We validated in 1632 patients in a third Illinois health system using area under the receiver operating characteristic curve (AUC). PREDICTION MODEL: Using a longitudinal discrete-time format, we trained a gradient boosted machine model to predict untreated infection in the next 6 hours using routinely available patient demographics, vital signs, and laboratory results. RESULTS: eDENTIFI had an AUC of 0.80 (95% CI, 0.79-0.81) in the validation cohort and outperformed the systemic inflammatory response syndrome criteria with an AUC of 0.64 (95% CI, 0.64-0.65; p < 0.001). The most important features were body mass index, age, temperature, and heart rate. Using a threshold with a 47.6% sensitivity, eDENTIFI detected infection a median 2.0 hours (interquartile range, 0.9-5.2 hr) before antimicrobial administration, with a negative predictive value of 93.6%. Antibiotic administration guided by eDENTIFI could have decreased unnecessary IV antibiotic administration in noninfected patients by 10.8% absolute or 46.4% relative percentage points compared with clinicians. CONCLUSION: eDENTIFI could both decrease the time to antimicrobial administration in infected patients and unnecessary antibiotic administration in noninfected patients. Further prospective validation is needed.

A Common Longitudinal Intensive Care Unit data Format (CLIF) to enable multi-institutional federated critical illness research.

Background: Critical illness, or acute organ failure requiring life support, threatens over five million American lives annually. Electronic health record (EHR) data are a source of granular information that could generate crucial insights into the nature and optimal treatment of critical illness. However, data management, security, and standardization are barriers to large-scale critical illness EHR studies. Methods: A consortium of critical care physicians and data scientists from eight US healthcare systems developed the Common Longitudinal Intensive Care Unit (ICU) data Format (CLIF), an open-source database format that harmonizes a minimum set of ICU Data Elements for use in critical illness research. We created a pipeline to process adult ICU EHR data at each site. After development and iteration, we conducted two proof-of-concept studies with a federated research architecture: 1) an external validation of an in-hospital mortality prediction model for critically ill patients and 2) an assessment of 72-hour temperature trajectories and their association with mechanical ventilation and in-hospital mortality using group-based trajectory models. Results: We converted longitudinal data from 94,356 critically ill patients treated in 2020-2021 (mean age 60.6 years [standard deviation 17.2], 30% Black, 7% Hispanic, 45% female) across 8 health systems and 33 hospitals into the CLIF format, The in-hospital mortality prediction model performed well in the health system where it was derived (0.81 AUC, 0.06 Brier score). Performance across CLIF consortium sites varied (AUCs: 0.74-0.83, Brier scores: 0.06-0.01), and demonstrated some degradation in predictive capability. Temperature trajectories were similar across health systems. Hypothermic and hyperthermic-slow-resolver patients consistently had the highest mortality. Conclusions: CLIF facilitates efficient, rigorous, and reproducible critical care research. Our federated case studies showcase CLIF's potential for disease sub-phenotyping and clinical decision-support evaluation. Future applications include pragmatic EHR-based trials, target trial emulations, foundational multi-modal AI models of critical illness, and real-time critical care quality dashboards.

Individualized Treatment Effects of Oxygen Targets in Mechanically Ventilated Critically Ill Adults.

Importance: Among critically ill adults, randomized trials have not found oxygenation targets to affect outcomes overall. Whether the effects of oxygenation targets differ based on an individual's characteristics is unknown. Objective: To determine whether an individual's characteristics modify the effect of lower vs higher peripheral oxygenation-saturation (Spo2) targets on mortality. Design, Setting, and Participants: A machine learning model to predict the effect of treatment with a lower vs higher Spo2 target on mortality for individual patients was derived in the Pragmatic Investigation of Optimal Oxygen Targets (PILOT) trial and externally validated in the Intensive Care Unit Randomized Trial Comparing Two Approaches to Oxygen Therapy (ICU-ROX) trial. Critically ill adults received invasive mechanical ventilation in an intensive care unit (ICU) in the United States between July 2018 and August 2021 for PILOT (n = 1682) and in 21 ICUs in Australia and New Zealand between September 2015 and May 2018 for ICU-ROX (n = 965). Exposures: Randomization to a lower vs higher Spo2 target group. Main Outcome and Measure: 28-Day mortality. Results: In the ICU-ROX validation cohort, the predicted effect of treatment with a lower vs higher Spo2 target for individual patients ranged from a 27.2% absolute reduction to a 34.4% absolute increase in 28-day mortality. For example, patients predicted to benefit from a lower Spo2 target had a higher prevalence of acute brain injury, whereas patients predicted to benefit from a higher Spo2 target had a higher prevalence of sepsis and abnormally elevated vital signs. Patients predicted to benefit from a lower Spo2 target experienced lower mortality when randomized to the lower Spo2 group, whereas patients predicted to benefit from a higher Spo2 target experienced lower mortality when randomized to the higher Spo2 group (likelihood ratio test for effect modification P = .02). The use of a Spo2 target predicted to be best for each patient, instead of the randomized Spo2 target, would have reduced the absolute overall mortality by 6.4% (95% CI, 1.9%-10.9%). Conclusion and relevance: Oxygenation targets that are individualized using machine learning analyses of randomized trials may reduce mortality for critically ill adults. A prospective trial evaluating the use of individualized oxygenation targets is needed.

Individualized Treatment Effects of Bougie versus Stylet for Tracheal Intubation in Critical Illness.

Rationale: A recent randomized trial found that using a bougie did not increase the incidence of successful intubation on first attempt in critically ill adults. The average effect of treatment in a trial population, however, may differ from effects for individuals. Objective: We hypothesized that application of a machine learning model to data from a clinical trial could estimate the effect of treatment (bougie vs. stylet) for individual patients based on their baseline characteristics ("individualized treatment effects"). Methods: This was a secondary analysis of the BOUGIE (Bougie or Stylet in Patients Undergoing Intubation Emergently) trial. A causal forest algorithm was used to model differences in outcome probabilities by randomized group assignment (bougie vs. stylet) for each patient in the first half of the trial (training cohort). This model was used to predict individualized treatment effects for each patient in the second half (validation cohort). Measurements and Main Results: Of 1,102 patients in the BOUGIE trial, 558 (50.6%) were the training cohort, and 544 (49.4%) were the validation cohort. In the validation cohort, individualized treatment effects predicted by the model significantly modified the effect of trial group assignment on the primary outcome (P value for interaction = 0.02; adjusted qini coefficient, 2.46). The most important model variables were difficult airway characteristics, body mass index, and Acute Physiology and Chronic Health Evaluation II score. Conclusions: In this hypothesis-generating secondary analysis of a randomized trial with no average treatment effect and no treatment effect in any prespecified subgroups, a causal forest machine learning algorithm identified patients who appeared to benefit from the use of a bougie over a stylet and from the use of a stylet over a bougie using complex interactions between baseline patient and operator characteristics.

Helmet noninvasive ventilation in acute hypoxic respiratory failure.

PURPOSE OF REVIEW: Invasive mechanical ventilation is a lifesaving intervention for patients with severe acute hypoxic respiratory failure (AHRF), but it is associated with neuromuscular, cognitive, and infectious complications. Noninvasive ventilation (NIV) may provide sufficient respiratory support without these complications. The helmet interface for NIV could address concerns raised for the use of NIV as first-line therapy in AHRF. This review will summarize and appraise the current evidence for helmet NIV in AHRF. RECENT FINDINGS: There are only six randomized controlled trials comparing helmet NIV to standard nasal cannula, facemask NIV, or high-flow nasal oxygen in patients with AHRF. Lower rates of endotracheal intubations and fewer days of mechanical ventilation were reported, with inconsistent findings on patient survival. Facemask NIV may worsen preexisting lung injury, delay intubations, and be inferior at delivering lung protective ventilation strategies compared with mechanical ventilation. The helmet interface could circumvent some of these concerns through the delivery of higher positive end expiratory pressure and more uniform distribution of negative pleural pressure. SUMMARY: There is limited evidence to support or refute the use of helmet NIV in AHRF. Further studies investigating the interface of helmet in NIV as a separate clinical entity are needed.

Oxygen-Saturation Targets for Critically Ill Adults Receiving Mechanical Ventilation.

BACKGROUND: Invasive mechanical ventilation in critically ill adults involves adjusting the fraction of inspired oxygen to maintain arterial oxygen saturation. The oxygen-saturation target that will optimize clinical outcomes in this patient population remains unknown. METHODS: In a pragmatic, cluster-randomized, cluster-crossover trial conducted in the emergency department and medical intensive care unit at an academic center, we assigned adults who were receiving mechanical ventilation to a lower target for oxygen saturation as measured by pulse oximetry (Spo2) (90%; goal range, 88 to 92%), an intermediate target (94%; goal range, 92 to 96%), or a higher target (98%; goal range, 96 to 100%). The primary outcome was the number of days alive and free of mechanical ventilation (ventilator-free days) through day 28. The secondary outcome was death by day 28, with data censored at hospital discharge. RESULTS: A total of 2541 patients were included in the primary analysis. The median number of ventilator-free days was 20 (interquartile range, 0 to 25) in the lower-target group, 21 (interquartile range, 0 to 25) in the intermediate-target group, and 21 (interquartile range, 0 to 26) in the higher-target group (P = 0.81). In-hospital death by day 28 occurred in 281 of the 808 patients (34.8%) in the lower-target group, 292 of the 859 patients (34.0%) in the intermediate-target group, and 290 of the 874 patients (33.2%) in the higher-target group. The incidences of cardiac arrest, arrhythmia, myocardial infarction, stroke, and pneumothorax were similar in the three groups. CONCLUSIONS: Among critically ill adults receiving invasive mechanical ventilation, the number of ventilator-free days did not differ among groups in which a lower, intermediate, or higher Spo2 target was used. (Supported by the National Heart, Lung, and Blood Institute and others; PILOT ClinicalTrials.gov number, NCT03537937.).

Creation of a medical procedure service in a tertiary medical center: Blueprint and procedural outcomes.

BACKGROUND: Medical Procedure Services (MPS) may represent a solution to the educational gap in procedural training among internal medicine residents and the unmet need for the clinical provision of non-urgent inpatient procedures. However, there is little guidance available to help launch an MPS. Furthermore, procedural outcomes from a newly initiated MPS, including those comparing trainees versus attending physicians, are lacking. OBJECTIVE: To describe the blueprint used in the design, implementation, and ongoing oversight of an MPS and to report its procedural outcomes. DESIGN, SETTINGS AND PARTICIPANTS: Vanderbilt University Medical Center (VUMC), Nashville, Tennessee. INTERVENTION: The launch of an MPS at a large tertiary academic hospital. MAIN OUTCOME AND MEASURES: 6,152 procedural consultations resulting in 5,320 attempted procedures over a four-and-a-half year period. RESULTS: The primary proceduralist was a supervised resident in 58.7% (3124 /5,320) and an attending in 41.3% (2,196/5,320) of procedures. The overall success rate was 91.1% (95% CI: 90.3-91.9%) and the major complication rate was 0.7% (95% CI: 0.5-1.0%). There was no difference in the mean number of attempts required to complete a procedure (1.6 vs 1.5 attempts, p=0.68) and the complication rates between supervised residents and attending proceduralists, respectively (20/3,124 vs 20/2,196, p=0.26). CONCLUSION: At a tertiary academic medical center, the implementation and maintenance of MPS is feasible, safe, and results in high rates of successful procedures performed by supervised residents. Procedures performed by supervised residents require comparable number of attempts for completion and carry similar risks as those performed alone by attendings.

Cryptococcal Meningitis Causing Refractory Hemichorea-Hemiballismus Treated With Pallidotomy.

We report a case of a 31-year-old immunocompetent male who presented with altered mental status and agitation requiring intubation. As sedation was weaned, he demonstrated choreiform movements with associated hemiballismus of the right upper and lower extremities, and he was ultimately diagnosed with cryptococcal meningitis. The patient's chorea did not terminate after the completion of induction antifungal therapy and all pharmacologic options for the management of chorea were ineffective. He underwent a successful unilateral pallidotomy using standard stereotactic methodology targeting the posterior-ventral pallidum, and his choreiform movements dramatically improved post-operatively within 48 hours.

Neurocysticercosis and HIV/AIDS co-infection: A scoping review.

OBJECTIVES: Neurocysticercosis (NCC) and human immunodeficiency virus (HIV) have a high disease burden and are prevalent in overlapping low- and middle-income areas. Yet, treatment guidance for people living with HIV/AIDS (PLWH/A) co-infected with NCC is currently lacking. This study aims to scope the available literature on HIV/AIDS and NCC co-infection, focusing on epidemiology, clinical characteristics, diagnostics and treatment outcomes. METHODS: The scoping literature review methodological framework, and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines were followed. A total of 16,969 records identified through database searching, and 45 additional records from other sources were reduced to 52 included studies after a standardised selection process. RESULTS: Two experimental studies, ten observational studies, 23 case series/case reports and 17 reviews or letters were identified. Observational studies demonstrated similar NCC seroprevalence in PLWH/A and their HIV-negative counterparts. Of 29 PLWH/A and NCC co-infection, 17 (59%) suffered from epileptic seizures, 15 (52%) from headaches and 15 (52%) had focal neurological deficits. Eighteen (62%) had viable vesicular cysts, and six (21%) had calcified cysts. Fifteen (52%) were treated with albendazole, of which 11 (73%) responded well to treatment. Five individuals potentially demonstrated an immune-reconstitution inflammatory syndrome after commencing antiretroviral therapy, although this was in the absence of immunological and neuroimaging confirmation. CONCLUSIONS: There is a paucity of evidence to guide treatment of PLWH/A and NCC co-infection. There is a pressing need for high-quality studies in this patient group to appropriately inform diagnostic and management guidelines for HIV-positive patients with NCC.

Protocolized Postextubation Respiratory Support to Prevent Reintubation: A Randomized Clinical Trial.

Rationale: Respiratory support (noninvasive ventilation or high-flow nasal cannula) applied at the time of extubation has been reported to reduce reintubation rates, but concerns regarding effectiveness have limited uptake into practice.Objectives: To determine if providing postextubation respiratory support to all patients undergoing extubation in a medical ICU would decrease the incidence of reintubation.Methods: We conducted a pragmatic, two-armed, cluster-crossover trial of adults undergoing extubation from invasive mechanical ventilation between October 1, 2017, and March 31, 2019, in the medical ICU of an academic medical center. Patients were assigned to either protocolized postextubation respiratory support (a respiratory therapist-driven protocol in which patients with suspected hypercapnia received noninvasive ventilation and patients without suspected hypercapnia received high-flow nasal cannula) or usual care (postextubation management at the discretion of treating clinicians). The primary outcome was reintubation within 96 hours of extubation.Measurements and Main Results: A total of 751 patients were enrolled. Of the 359 patients assigned to protocolized support, 331 (92.2%) received postextubation respiratory support compared with 66 of 392 patients (16.8%) assigned to usual care, a difference driven by differential use of high-flow nasal cannula (74.7% vs. 2.8%). A total of 57 patients (15.9%) in the protocolized support group experienced reintubation compared with 52 patients (13.3%) in the usual care group (odds ratio, 1.23; 95% confidence interval, 0.82 to 1.84; P = 0.32).Conclusions: Among a broad population of critically ill adults undergoing extubation from invasive mechanical ventilation at an academic medical center, protocolized postextubation respiratory support, primarily characterized by an increase in the use of high-flow nasal cannula, did not prevent reintubation compared with usual care.Clinical trial registered with www.clinicaltrials.gov (NCT0328831).

Medical Procedure Services in Internal Medicine Residencies in the US: a Systematic Review and Meta-Analysis.

BACKGROUND AND AIMS: The number of procedures performed by internal medicine residents in the United States (US) is declining. An increasing proportion of residents do not feel confident performing essential invasive bedside procedures and, upon graduation, desire additional training. Several residency programs have utilized the medical procedure service (MPS) to address this issue. We aim to summarize the current state of evidence by systematically evaluating the effect of the MPS on resident education, comfort, and training, as well as patient safety and procedural outcomes in the US. METHODS: We conducted a systematic review of all studies reporting the use of an MPS with supervision from a board-certified physician in internal medicine residencies in the US. Database search was performed on PubMed, Embase, ERIC, and Cochrane Library from January 2000 to November 2020 for relevant studies. Quality of evidence assessment and random-effects proportion meta-analyses were performed. RESULTS: A total of nine studies reporting on 3879 procedures performed by MPS were identified. Procedures were safely performed, with a pooled complication rate of 2.1% (95% CI: 1.0-3.5) and generally successful, with a pooled success rate of 94.7% (95% CI: 90.8-97.7). The range of procedures performed by residents under MPS was 6.7-72.8 procedures per month (n = 9) compared to 4.3-64.4 procedures (n = 4) without MPS. MPS significantly increased confidence, comfort, and use of appropriate safety measures among residents. CONCLUSION: There are a limited number of published studies on MPS supervised by a board-certified physician in US internal medicine residencies. Procedures performed by MPS are generally successfully completed and safe. MPS benefits internal medicine residents training by improving competency, comfort, and confidence.

Time to First Culture Positivity for Gram-Negative Rods Resistant to Ceftriaxone in Critically Ill Adults.

BACKGROUND: The optimal timing for the de-escalation of broad-spectrum antibiotics with activity against Pseudomonas aeruginosa and resistant Gram-negative rods (GNRs) in critically ill adults remains unknown. RESEARCH QUESTION: We tested the hypothesis that cultures will identify GNRs that ultimately demonstrate resistance to ceftriaxone within 48 hours, potentially allowing safe de-escalation at this time point. STUDY DESIGN AND METHODS: We conducted a secondary analysis of data from the Isotonic Solutions and Major Adverse Renal Events Trial: a pragmatic, cluster-randomized, multiple-crossover trial comparing balanced crystalloids versus saline for intravenous fluid administration in 15,802 critically ill adults at 5 intensive care units (ICUs) at Vanderbilt University Medical Center in Nashville, TN, USA. The primary endpoint was the time-to-positivity of respiratory and blood cultures that ultimately demonstrated growth of GNRs resistant to ceftriaxone. Multivariable logistic regression modeling was used to examine risk factors for the growth of cultures after 48 hours. RESULTS: A total of 524 respiratory cultures had growth of GNRs, of which 284 (54.2%) had resistance to ceftriaxone. A total of 376 blood cultures grew GNRs, of which 70 (18.6%) had resistance to ceftriaxone. At 48 hours, 87% of respiratory cultures and 85% of blood cultures that ultimately grew GNRs resistant to ceftriaxone had demonstrated growth. Age, gender, predicted risk of inpatient mortality and prior use of antibiotics did not predict the growth of cultures after 48 hours. INTERPRETATION: Among a cohort of critically ill adults, 13% of respiratory cultures and 15% of blood cultures that ultimately grew GNRs resistant to ceftriaxone did not demonstrate growth until at least 48 hours after collection. Further work is needed to determine the ideal time for critically ill adults to de-escalate from broad-spectrum antibiotics targeting Pseudomonas aeruginosa and extended-spectrum ß-lactamase-producing gram-negative pathogens.

Expediting the management of cauda equina syndrome in the emergency department through clinical pathway design.

Introduction: Cauda equina syndrome (CES) is a neurosurgical emergency. Early diagnosis with MRI and subsequent surgical decompression surgery can prevent permanent neurological dysfunction. Charing Cross Hospital (CXH) is a tertiary neurosurgical referral centre where in the emergency department (ED), current practice mandated a neurosurgery review prior to requesting MRI. Hypothesis: It was hypothesised that a new clinical pathway, with better coordination from the ED, radiology and neurosurgical teams could reduce the time of presentation to diagnosis or exclusion of CES. Method: Retrospective case-note analysis of patients presenting with back pain to CXH ED over a 3-month period was performed. The primary outcome was the time interval between the patient's arrival to the ED and the MRI preliminary report. Results: The baseline primary outcome was recorded at 8 hours and 16 min (n=30). A new clinical pathway was designed empowering ED senior decision makers to order MRIs prior to neurosurgical review. Two Plan-Do-Study-Act (PDSA) cycles were performed, each measured over a 2-month period. The first PDSA cycle was performed after the pathway was initially launched (n=17), while the second PDSA cycle measured the effect of staff education and active promotion of the pathway (n=17). MRI was requested earlier, waiting and reporting time for MRI were reduced. The exclusion or diagnosis of CES was reduced to 5 hours and 54 min in PDSA 1 and 5 hours 17 min in PDSA 2, a 29% and 36% reduction (p=0.048 and p=0.012, respectively). Conclusion: The clinical protocol was a cost-neutral and sustainable intervention that effectively reduced the time taken to diagnose or exclude CES and ED waiting times.

Atypical Clinical Manifestations of Loiasis and Their Relevance for Endemic Populations.

BACKGROUND: Loiasis is mostly considered a relatively benign infection when compared with other filarial and parasitic diseases, with Calabar swellings and eyeworm being the most common signs. Yet, there are numerous reports in the literature of more serious sequelae. Establishing the relationship between infection and disease is a crucial first step toward estimating the burden of loiasis. METHODS: We conducted a systematic review of case reports containing 329 individuals and detailing clinical manifestations of loiasis with a focus on nonclassical, atypical presentations. RESULTS: Results indicate a high proportion (47%) of atypical presentations in the case reports identified, encompassing a wide range of cardiac, respiratory, gastrointestinal, renal, neurological, ophthalmological, and dermatological pathologies. Individuals with high microfilarial densities and residing in an endemic country were at greater risk of suffering from atypical manifestations. CONCLUSIONS: Our findings have important implications for understanding the clinical spectrum of conditions associated with Loa loa infection, which extends well beyond the classical eyeworm and Calabar swellings. As case reports may overestimate the true rate of atypical manifestations in endemic populations, large-scale, longitudinal clinico-epidemiological studies will be required to refine our estimates and demonstrate causality between loiasis and the breadth of clinical manifestations reported. Even if the rates of atypical presentations were found to be lower, given that residents of loiasis-endemic areas are both numerous and the group most at risk of severe atypical manifestations, our conclusions support the recognition of loiasis as a significant public health burden across Central Africa.

Protocolized Post-Extubation Respiratory Support to prevent reintubation: protocol and statistical analysis plan for a clinical trial.

INTRODUCTION: Following extubation from invasive mechanical ventilation, nearly one in seven critically ill adults requires reintubation. Reintubation is independently associated with increased mortality. Postextubation respiratory support (non-invasive ventilation or high-flow nasal cannula applied at the time of extubation) has been reported in small-to-moderate-sized trials to reduce reintubation rates among hypercapnic patients, high-risk patients without hypercapnia and low-risk patients without hypercapnia. It is unknown whether protocolised provision of postextubation respiratory support to every patient undergoing extubation would reduce the overall reintubation rate, compared with usual care. METHODS AND ANALYSIS: The Protocolized Post-Extubation Respiratory Support (PROPER) trial is a pragmatic, cluster cross-over trial being conducted between 1 October 2017 and 31 March 2019 in the medical intensive care unit of Vanderbilt University Medical Center. PROPER compares usual care versus protocolized post-extubation respiratory support (a respiratory therapist-driven protocol that advises the provision of non-invasive ventilation or high-flow nasal cannula based on patient characteristics). For the duration of the trial, the unit is divided into two clusters. One cluster receives protocolised support and the other receives usual care. Each cluster crosses over between treatment group assignments every 3 months. All adults undergoing extubation from invasive mechanical ventilation are enrolled except those who received less than 12 hours of mechanical ventilation, have 'Do Not Intubate' orders, or have been previously reintubated during the hospitalisation. The anticipated enrolment is approximately 630 patients. The primary outcome is reintubation within 96 hours of extubation. ETHICS AND DISSEMINATION: The trial was approved by the Vanderbilt Institutional Review Board. The results will be submitted for publication in a peer-reviewed journal and presented at one or more scientific conferences. TRIAL REGISTRATION NUMBER: NCT03288311.