Explore the vast range of titles available.

The authors used a validated model with electronic-medical-record data to identify hospitalized patients at high risk for clinical deterioration. The intervention, which involved remote monitoring by nurses who reviewed records of high-risk patients and communicated with in-hospital rapid-response teams, was associated with decreased 30-day mortality.

Prior sepsis studies evaluating antibiotic timing have shown mixed results. To evaluate the association between antibiotic timing and mortality among patients with sepsis receiving antibiotics within 6 hours of emergency department registration. Retrospective study of 35,000 randomly selected inpatients with sepsis treated at 21 emergency departments between 2010 and 2013 in Northern California. The primary exposure was antibiotics given within 6 hours of emergency department registration. The primary outcome was adjusted in-hospital mortality. We used detailed physiologic data to quantify severity of illness within 1 hour of registration and logistic regression to estimate the odds of hospital mortality based on antibiotic timing and patient factors. The median time to antibiotic administration was 2.1 hours (interquartile range, 1.4-3.1 h). The adjusted odds ratio for hospital mortality based on each hour of delay in antibiotics after registration was 1.09 (95% confidence interval [CI], 1.05-1.13) for each elapsed hour between registration and antibiotic administration. The increase in absolute mortality associated with an hour's delay in antibiotic administration was 0.3% (95% CI, 0.01-0.6%; P = 0.04) for sepsis, 0.4% (95% CI, 0.1-0.8%; P = 0.02) for severe sepsis, and 1.8% (95% CI, 0.8-3.0%; P = 0.001) for shock. In a large, contemporary, and multicenter sample of patients with sepsis in the emergency department, hourly delays in antibiotic administration were associated with increased odds of hospital mortality even among patients who received antibiotics within 6 hours. The odds increased within each sepsis severity strata, and the increased odds of mortality were greatest in septic shock.

[...]there has also been enormous interest in applying AI to health care and, in particular, to data-rich environments like the intensive care unit. With its highly touted AI—the car’s technology deploys sensors, computer vision, and deep learning to drive under its own guidance—having logged billions of driving miles, I anticipated a seamless transition between myself and the vehicle. [...]algorithms are trained using existing data and, thus, encode prior decisions and biases within them. Rights and permissions Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.

Background Recent single-center reports have suggested that community-acquired bacteremic co-infection in the context of Coronavirus disease 2019 (COVID-19) may be an important driver of mortality; however, these reports have not been validated with a multicenter, demographically diverse, cohort study with data spanning the pandemic. Methods In this multicenter, retrospective cohort study, inpatient encounters were assessed for COVID-19 with community-acquired bacteremic co-infection using 48-h post-admission blood cultures and grouped by: (1) confirmed co-infection [recovery of bacterial pathogen], (2) suspected co-infection [negative culture with ≥ 2 antimicrobials administered], and (3) no evidence of co-infection [no culture]. The primary outcomes were in-hospital mortality, ICU admission, and mechanical ventilation. COVID-19 bacterial co-infection risk factors and impact on primary outcomes were determined using multivariate logistic regressions and expressed as adjusted odds ratios with 95% confidence intervals (Cohort, OR 95% CI, Wald test p value). Results The studied cohorts included 13,781 COVID-19 inpatient encounters from 2020 to 2022 in the University of Alabama at Birmingham (UAB, n  = 4075) and Ochsner Louisiana State University Health—Shreveport (OLHS, n  = 9706) cohorts with confirmed (2.5%), suspected (46%), or no community-acquired bacterial co-infection (51.5%) and a comparison cohort consisting of 99,170 inpatient encounters from 2010 to 2019 (UAB pre-COVID-19 pandemic cohort). Significantly increased likelihood of COVID-19 bacterial co-infection was observed in patients with elevated ≥ 15 neutrophil-to-lymphocyte ratio (UAB: 1.95 [1.21–3.07]; OLHS: 3.65 [2.66–5.05], p  < 0.001 for both) within 48-h of hospital admission. Bacterial co-infection was found to confer the greatest increased risk for in-hospital mortality (UAB: 3.07 [2.42–5.46]; OLHS: 4.05 [2.29–6.97], p  < 0.001 for both), ICU admission (UAB: 4.47 [2.87–7.09], OLHS: 2.65 [2.00–3.48], p  < 0.001 for both), and mechanical ventilation (UAB: 3.84 [2.21–6.12]; OLHS: 2.75 [1.87–3.92], p  < 0.001 for both) across both cohorts, as compared to other risk factors for severe disease. Observed mortality in COVID-19 bacterial co-infection (24%) dramatically exceeds the mortality rate associated with community-acquired bacteremia in pre-COVID-19 pandemic inpatients (5.9%) and was consistent across alpha, delta, and omicron SARS-CoV-2 variants. Conclusions Elevated neutrophil-to-lymphocyte ratio is a prognostic indicator of COVID-19 bacterial co-infection within 48-h of admission. Community-acquired bacterial co-infection, as defined by blood culture-positive results, confers greater increased risk of in-hospital mortality, ICU admission, and mechanical ventilation than previously described risk factors (advanced age, select comorbidities, male sex) for COVID-19 mortality, and is independent of SARS-CoV-2 variant.

Both quality improvement and clinical research efforts over the past few decades have focused on consensus definition of sepsis and acute respiratory distress syndrome (ARDS). Although clinical definitions based on readily available clinical data have advanced recognition and timely use of broad supportive treatments, they likely hinder the identification of more targeted therapies that manipulate select biological mechanisms underlying critical illness. Sepsis and ARDS are by definition heterogeneous, and patients vary in both their underlying biology and their severity of illness. We have long been able to identify subtypes of sepsis and ARDS that confer different prognoses. The key is that we are now on the verge of identifying subtypes that may confer different response to therapy. In this perspective, inspired by a 2015 American Thoracic Society International Conference Symposium entitled \"Lumpers and Splitters: Phenotyping in Critical Illness,\" we highlight promising approaches to uncovering patient subtypes that may predict treatment responsiveness and not just differences in prognosis. We then discuss how this information can be leveraged to improve the success and translatability of clinical trials by using predictive enrichment and other design strategies. Last, we discuss the challenges and limitations to identifying biomarkers and endotypes and incorporating them into routine clinical practice.

AbstractObjectiveTo understand the epidemiology and burden of severe coronavirus disease 2019 (covid-19) during the first epidemic wave on the west coast of the United States.DesignProspective cohort study.SettingKaiser Permanente integrated healthcare delivery systems serving populations in northern California, southern California, and Washington state.Participants1840 people with a first acute hospital admission for confirmed covid-19 by 22 April 2020, among 9 596 321 healthcare plan enrollees. Analyses of hospital length of stay and clinical outcomes included 1328 people admitted by 9 April 2020 (534 in northern California, 711 in southern California, and 83 in Washington).Main outcome measuresCumulative incidence of first acute hospital admission for confirmed covid-19, and subsequent probabilities of admission to an intensive care unit (ICU) and mortality, as well as duration of hospital stay and ICU stay. The effective reproduction number (RE) describing transmission dynamics was estimated for each region.ResultsAs of 22 April 2020, cumulative incidences of a first acute hospital admission for covid-19 were 15.6 per 100 000 cohort members in northern California, 23.3 per 100 000 in southern California, and 14.7 per 100 000 in Washington. Accounting for censoring of incomplete hospital stays among those admitted by 9 April 2020, the estimated median duration of stay among survivors was 9.3 days (with 95% staying 0.8 to 32.9 days) and among non-survivors was 12.7 days (1.6 to 37.7 days). The censoring adjusted probability of ICU admission for male patients was 48.5% (95% confidence interval 41.8% to 56.3%) and for female patients was 32.0% (26.6% to 38.4%). For patients requiring critical care, the median duration of ICU stay was 10.6 days (with 95% staying 1.3 to 30.8 days). The censoring adjusted case fatality ratio was 23.5% (95% confidence interval 19.6% to 28.2%) among male inpatients and 14.9% (11.8% to 18.6%) among female inpatients; mortality risk increased with age for both male and female patients. Reductions in RE were identified over the study period within each region.ConclusionsAmong residents of California and Washington state enrolled in Kaiser Permanente healthcare plans who were admitted to hospital with covid-19, the probabilities of ICU admission, of long hospital stay, and of mortality were identified to be high. Incidence rates of new hospital admissions have stabilized or declined in conjunction with implementation of social distancing interventions.

A significant minority of individuals develop trauma- and stressor-related disorders (TSRD) after surviving sepsis, a life-threatening immune response to infections. Accurate prediction of risk for TSRD can facilitate targeted early intervention strategies, but many existing models rely on research measures that are impractical to incorporate to standard emergency department workflows. To increase the feasibility of implementation, we developed models that predict TSRD in the year after survival from sepsis using only electronic health records from the hospitalization (n = 217,122 hospitalizations from 2012-2015). The optimal model was evaluated in a temporally independent prospective test sample (n = 128,783 hospitalizations from 2016-2017), where patients in the highest-risk decile accounted for nearly one-third of TSRD cases. Our approach demonstrates that risk for TSRD after sepsis can be stratified without additional assessment burden on clinicians and patients, which increases the likelihood of model implementation in hospital settings.

Interest in machine-learning applications within medicine has been growing, but few studies have progressed to deployment in patient care. We present a framework, context and ultimately guidelines for accelerating the translation of machine-learning-based interventions in health care. To be successful, translation will require a team of engaged stakeholders and a systematic process from beginning (problem formulation) to end (widespread deployment).

Background Increasing evidence suggests that social factors and problems with physical and cognitive function may contribute to patients’ rehospitalization risk. Understanding a patient’s readmission risk may help healthcare providers develop tailored treatment and post-discharge care plans to reduce readmission and mortality. This study aimed to evaluate whether including patient-reported data on social factors; cognitive status; and physical function improves on a predictive model based on electronic health record (EHR) data alone. Methods We conducted a prospective study of 1,547 hospitalized adult patients in 3 Kaiser Permanente Northern California hospitals. The main outcomes were non-elective rehospitalization or death within 30 days post-discharge. Exposures included patient-reported social factors and cognitive and physical function (obtained in a pre-discharge interview) and EHR–derived data for comorbidity burden, acute physiology, care directives, prior utilization, and hospital length of stay. We performed bivariate comparisons using Chi-square, t-tests, and Wilcoxon rank-sum tests and assessed correlations between continuous variables using Spearman’s rho statistic. For all models, the results reported were obtained after fivefold cross validation. Results The 1,547 adult patients interviewed were younger (age, p  = 0.03) and sicker (COPS2, p  < 0.0001) than the rest of the hospitalized population. Of the 6 patient-reported social factors measured, 3 (not living with a spouse/partner, transportation difficulties, health or disability-related limitations in daily activities) were significantly associated ( p  < 0.05) with the main outcomes, while 3 (living situation concerns, problems with food availability, financial problems) were not. Patient-reported cognitive ( p  = 0.027) and physical function ( p  = 0.01) were significantly lower in patients with the main outcomes. None of the patient-reported variables, singly or in combination, improved predictive performance of a model that included acute physiology and longitudinal comorbidity burden (area under the receiver operator characteristic curve was 0.716 for both the EHR model and maximal performance of a random forest model including all predictors). Conclusions In this insured population, incorporating patient-reported social factors and measures of cognitive and physical function did not improve performance of an EHR-based model predicting 30-day non-elective rehospitalization or mortality. While incorporating patient-reported social and functional status data did not improve ability to predict these outcomes, such data may still be important for improving patient outcomes.

Hospital readmission is a key quality metric, yet post-discharge interventions often yield variable results. In the first large-scale randomized evaluation of causal machine learning in a health system, we assessed whether a novel model (the Predicted Benefit Intervention (PBI) score) could identify lower-risk patients most likely to benefit from post-discharge care coordination within Kaiser Permanente Northern California (KPNC). From May to December 2022, 9959 low-risk patients at 19 KPNC hospitals were randomized to usual care or the Transitions Program, which included medication reconciliation, primary care follow-up scheduling, and weekly calls for 30 days. While 30-day readmissions declined in the intervention group (7.7% vs. 8.2%), the difference was not statistically significant. However, the observed-to-expected readmission ratio declined into randomization and remained low thereafter; this decline was statistically significant. This study demonstrates the feasibility of implementing causal machine learning at scale to improve targeting and resource allocation in care delivery.