bioRxiv Science⌕ Search
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 trialsPreprint

Keep this discovery

BibTeXRIS

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

bioRxiv2018

Multi-Center Study of Resectable Lung Lesions by Ultra-Deep Sequencing of Targeted Genes in Plasma Cell-Free DNA to Assess Nodule Malignancy and Detect Lung Cancers

BACKGROUNDEarly detection of lung cancer to allow curative treatment remains challenging. Cell-free circulating tumor DNA (ctDNA) analysis may aid in malignancy assessment and early cancer diagnosis of lung nodules found in screening imagery.\n\nMETHODSThe multi-center clinical study enrolled 192 patients with operable occupying lung diseases. Plasma ctDNA, white blood cell genomic DNA (gDNA) and tumor tissue gDNA of each patient were analyzed by ultra-deep sequencing to an average of 35,000X of the coding regions of 65 lung cancer-related genes.\n\nRESULTSThe cohort consists of a quarter of benign lung diseases and three quarters of cancer patients with all histopathology subtypes. 64% of the cancer patients is at Stage I. Gene mutations detection in tissue gDNA and plasma ctDNA results in a sensitivity of 91% and specificity of 88%. When ctDNA assay was used as the test, the sensitivity was 69% and specificity 96%. As for the lung cancer patients, the assay detected 63%, 83%, 94% and 100%, for Stage I, II, III and IV, respectively. In a linear discriminant analysis, combination of ctDNA, patient age and a panel of serum biomarkers boosted the overall sensitivity to 80% at a specificity of 99%. 29 out of the 65 genes harbored mutations in the lung cancer patients with the largest number found in TP53 (30% plasma and 62% tumor tissue samples) and EGFR (20% and 40%, respectively).\n\nCONCLUSIONPlasma ctDNA was analyzed in lung nodule assessment and early cancer detection while an algorithm combining clinical information enhanced the test performance.

clinical trials↗
bioRxiv2018

Bacterial strain displacement in inflammatory bowel diseases after fecal microbiota transplantation

Fecal microbiota transplantation (FMT), which is thought to have the potential to correct dysbiosis of gut microbiota, has recently been used to treat inflammatory bowel disease (IBD). To elucidate the extent and principles of microbiota engraftment in IBD patients after FMT treatment, we conducted an interventional prospective cohort study. The cohort included two categories of patients: (1) patients with moderate to severe Crohns disease (CD) (Harvey-Bradshaw Index [&ge;] 7, n = 11, and (2) patients with ulcerative colitis (UC) (Montreal classification, S2 and S3, n = 4). All patients were treated with a single FMT (via mid-gut, from healthy donors) and follow-up visits were performed at baseline, 3 days, one week, and one month after FMT (missing time points included). At each follow-up time point, fecal samples of the participants were collected along with their clinical metadata. For comparative analysis, 10 fecal samples from 10 healthy people were included to represent the diversity level of normal gut microbiota. Additionally, the metagenomic data of 25 fecal samples from 5 individuals with metabolic syndrome who underwent autologous FMT treatment were downloaded from a previous published paper to represent natural microbiota shifts during FMT. All fecal samples underwent shotgun metagenomic sequencing. We found that 3 days after FMT, 11 out of 15 recipients were in remission (3 out of 4 UC recipients; 8 out of 11 CD recipients). Generally, bacterial colonization was observed to be lower in CD recipients than in UC recipients at both species and strain levels. Furthermore, across species, different strains displayed disease-specific displacement advantages under two-disease status. Finally, most post-FMT species (> 80%) could be properly predicted (AUC > 85%) using a random forest classification model, with the gut microbiota composition and clinical parameters of pre-FMT recipients acting as the most contributive factors for prediction accuracy.

clinical trials↗
bioRxiv2018

Dosage Matters: A Randomized Controlled Trial of Rehabilitation Dose in the Chronic Phase after Stroke

Background and PurposeFor stroke rehabilitation, task-specific training in animal models and human rehabilitation trials is considered important to trigger inherent neuroplasticity, promote motor learning, and functional recovery. Little is known, however, about what constitutes an effective dosage of therapy.\n\nMethodsThis is a parallel group, four arm, single blind, phase I, randomized control trial of four dosages of upper extremity therapy delivered in an outpatient setting during the chronic phase after stroke. Participants were randomized into groups that varied in total dosage of therapy (i.e., 0, 15, 30, or 60 hours). Seven hundred and four participants were assessed for eligibility, 50 were eligible to enroll, 45 were randomized, 44 participated and 41 completed the study. Planned primary analyses used linear mixed effects regression to model baseline to post-intervention changes in the Motor Activity Log-Quality of Movement rating (MALQ) and the Wolf Motor Function Test (WMFT) time score as a function of therapy dosage. A series of hierarchical models were constructed using the MALQ and WMFT.\n\nResultsWe observed a significant dose response curve: the greater the dosage of training, the greater the change in MALQ, with the dose by week slope parameter of 0.0045 ({Delta}MAL/hour/week; p = 0.0011; 95% CI = [0.0019; 0.0071]). Over the 3 weeks of therapy, this corresponds to a gain of 0.81 in MALQ for the 60 hour dose.\n\nConclusionsFor mild-to-moderately impaired stroke survivors, the dosage of a patient-centered, task specific motor therapy was shown to systematically influence the gain in quality of arm use in the natural environment, but not functional capacity as measured in the laboratory. We highlight the importance of recovery outcomes that capture arm use vs. functional capacity.\n\nClinical Trial RegistrationURL: http://www.clinicaltrials.gov. Unique identifier: NCT 01749358

clinical trials↗