Evaluating a sepsis prediction machine learning algorithm using minimal electronic health record data in the emergency department and intensive care unit

Burdick, H.; Pino, E.; Gabel-Comeau, D.; Gu, C.; Huang, H.; Lynn-Palevsky, A.; Das, R.

doi:10.1101/224014

Home/Archive/bioRxiv

bioRxiv · 10.1101/224014

Evaluating a sepsis prediction machine learning algorithm using minimal electronic health record data in the emergency department and intensive care unit

Burdick, H.·Pino, E.·Gabel-Comeau, D.·Gu, C.·Huang, H.·Lynn-Palevsky, A.·Das, R.

View original source ↗Export metadata ↓

Abstract

IntroductionSepsis is a major health crisis in US hospitals, and several clinical identification systems have been designed to help care providers with early diagnosis of sepsis. However, many of these systems demonstrate low specificity or sensitivity, which limits their clinical utility. We evaluate the effects of a machine learning algodiagnostic (MLA) sepsis prediction and detection system using a before-and-after clinical study performed at Cabell Huntington Hospital (CHH) in Huntington, West Virginia. Prior to this study, CHH utilized the St. Johns Sepsis Agent (SJSA) as a rules-based sepsis detection system.\n\nMethodsThe Predictive algoRithm for EValuation and Intervention in SEpsis (PREVISE) study was carried out between July 1, 2017 and August 30, 2017. All patients over the age of 18 who were admitted to the emergency department or intensive care units at CHH were monitored during the study. We assessed pre-implementation baseline metrics during the month of July, 2017, when the SJSA was active. During implementation in the month of August, 2017, SJSA and the MLA concurrently monitored patients for sepsis risk. At the conclusion of the study period, the primary outcome of sepsis-related in-hospital mortality and secondary outcome of sepsis-related hospital length of stay were compared between the two groups.\n\nResultsSepsis-related in-hospital mortality decreased from 3.97% to 2.64%, a 33.5% relative decrease (P = 0.038), and sepsis-related length of stay decreased from 2.99 days in the pre-implementation phase to 2.48 days in the post-implementation phase, a 17.1% relative reduction (P < 0.001).\n\nConclusionReductions in patient mortality and length-of-stay were observed with use of a machine learning algorithm for early sepsis detection in the emergency department and intensive care units at Cabell Huntington Hospital, and may present a method for improving patient outcomes.\n\nTrial RegistrationClinicalTrials.gov, NCT03235193, retrospectively registered on July 27th 2017.

Explore related subjects

clinical trials Preprint

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Burdick, H., Pino, E., Gabel-Comeau, D., Gu, C., Huang, H., Lynn-Palevsky, A., Das, R.. 2017-11-28. Evaluating a sepsis prediction machine learning algorithm using minimal electronic health record data in the emergency department and intensive care unit. https://doi.org/10.1101/224014

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2019

Nicotinamide riboside augments the human skeletal muscle NAD+ metabolome and induces transcriptomic and anti-inflammatory signatures in aged subjects: a placebo-controlled, randomized trial

NAD+ is modulated by conditions of metabolic stress and has been reported to decline with aging, but human data are sparse. Nicotinamide riboside (NR) supplementation ameliorates metabolic dysfunction in rodents. We aimed to establish whether oral NR supplementation in aged participants can increase the skeletal muscle NAD+ metabolome, and questioned if tissue NAD+ levels are depressed with aging. We supplemented 12 aged men with NR 1g per day for 21-days in a placebo-controlled, randomized, double-blind, crossover trial. Targeted metabolomics showed that NR elevated the muscle NAD+ metabolome, evident by increased nicotinic acid adenine dinucleotide and nicotinamide clearance products. Muscle RNA sequencing revealed NR-mediated downregulation of energy metabolism and mitochondria pathways. NR also depressed levels of circulating inflammatory cytokines. In an additional study, 31P magnetic resonance spectroscopy-based NAD+ measurement in muscle and brain showed no difference between young and aged individuals. Our data establish that oral NR is available to aged human muscle and identify anti-inflammatory effects of NR, while suggesting that NAD+ decline is not associated with chronological aging per se in human muscle or brain.

clinical trials↗

bioRxiv2019

Stanford Accelerated Intelligent Neuromodulation Therapy (SAINT) Induces Functional Connectivity Changes in Emotion Regulation Brain Areas for MDD Patients

BackgroundMajor depressive disorder (MDD) is prevalent and debilitating, and development of improved treatments is limited in part by insufficient understanding of the mechanism of disease remission. In turn, efforts to elucidate mechanisms have been challenging due to disease heterogeneity and limited effectiveness of treatments, which require weeks-to-months to induce remission. We recently developed a form of repetitive transcranial magnetic stimulation that induces remission in 90% of individuals with severe, treatment resistant MDD in 1-5 days (SAINT). This provides a new tool to begin exploring the network-level mechanisms of MDD remission. ObjectiveDetermine the functional connectivity (fc) changes that occur with SAINT in brain regions associated with emotion regulation. MethodsResting-state fMRI scans were performed just prior to (baseline), and following, open-label SAINT in 18 participants with severe, treatment-resistant MDD. Fc was determined between regions of interest (ROIs) defined a priori with well-described roles in emotion regulation. ResultsFollowing SAINT treatment fc was significantly decreased between the following ROI pairs: dorsolateral prefrontal cortex (DLPFC)-striatum, DLPFC-amygdala, default mode network (DMN)-subgenual cingulate cortex (sgACC), DMN-amygdala, DMN-striatum, amygdala-striatum, and between amygdala subregions. Greater clinical improvements were associated with larger decreases in fc between DLPFC-amygdala and DLPFC-insula. Greater clinical improvements were associated with smaller decreases in fc between sgACC-DMN. Significant increases and decreases in fc between insula-amygdala were observed depending on the subregions. Greater clinical improvements were associated with lower baseline fc between DMN-DLPFC, DMN-striatum, and DMN-ventrolateral prefrontal cortex. ConclusionSAINT-induced remission from depression is associated with fc changes that suggest improved regulation of emotion. Although preliminary, this leads us to hypothesize that interventions that augment top-down regulation of emotion may be effective depression treatments.

clinical trials↗

bioRxiv2019

Efficacy and safety of a novel vaginal medical device in Recurrent Bacterial Vaginosis: an international multicentre clinical trial

Several risk factors have been identified but the etiology and pathogenesis of Bacterial vaginosis (BV) are still not completely understood, and the recurrence rate of BV remains high despite adequate chemotherapy treatment. The primary objective of the study was to assess the effectiveness of a new vaginal medical device, which contains polycarbophil, 0.04 % lauryl glucoside, and glycerides (Polybactum(R) - Effik Italia), in reducing BV recurrence rate. This was a multicenter, open label, not comparative study performed in Italy and Romania. Female subjects over 18-years-old affected by recurrent BV were included. The latest episode was diagnosed by Amsel criteria 6-9 days before the start of the study and treated with vaginal metronidazole (gel 0.75% mg for 5 days or ovules 500 mg for 7 days). The recurrence was defined by at least 2 episodes in the previous 12 months. Polybactum(R) vaginal ovules, day 1-4-7, were started within the 12th and the 24th hr after the end of metronidazole therapy and repeated monthly for 3 cycles. The first 41 patients enrolled were evaluated for an interim analysis 6 months after the study started; 2 patients interrupted the trial, leaving 39 evaluable subjects. The recurrence rate was significantly reduced compared to previous published data (10.26% vs 40% p<0.001). In 35 patients without recurrence, the assessment of Lactobacillus vaginal flora performed by phase contrast microscopy evidenced a significant improvement form baseline (p=0.022) The Investigator global assessment of tolerability was excellent in 38 out of 39 cases. IMPORTANCEBacterial vaginosis (BV) is the most common vaginal disorder in women of childbearing age. In BV, Lactobacillus species, which are predominant in a healthy vaginal flora, are replaced by anaerobes, mainly Gardnerella vaginalis. BV is responsible for more than 60% of vulvovaginal infections and has been linked to serious, potentially life-threatening conditions, including: pelvic inflammatory disease, postoperative infections, acquisition and transmission of the human immunodeficiency virus, preterm birth, and several adverse pregnancy outcomes. Our research showed that 3 monthly cycles of Polybactum(R) ovules administered after one course of metronidazole vaginal therapy can reduce the rate of Bacterial vaginosis recurrence and improve the vaginal milieu, favouring the growth of vaginal lactobacillus species. Taken together our results confirm that Polibactum(R) is a safe and effective treatment to reduce BV recurrence rate after a first line therapy with metronidazole.

clinical trials↗