Dataset Information

Critical Transitions in Intensive Care Units: A Sepsis Case Study.

ABSTRACT: The progression of complex human diseases is associated with critical transitions across dynamical regimes. These transitions often spawn early-warning signals and provide insights into the underlying disease-driving mechanisms. In this paper, we propose a computational method based on surprise loss (SL) to discover data-driven indicators of such transitions in a multivariate time series dataset of septic shock and non-sepsis patient cohorts (MIMIC-III database). The core idea of SL is to train a mathematical model on time series in an unsupervised fashion and to quantify the deterioration of the model's forecast (out-of-sample) performance relative to its past (in-sample) performance. Considering the highest value of the moving average of SL as a critical transition, our retrospective analysis revealed that critical transitions occurred at a median of over 35?hours before the onset of septic shock, which suggests the applicability of our method as an early-warning indicator. Furthermore, we show that clinical variables at critical-transition regions are significantly different between septic shock and non-sepsis cohorts. Therefore, our paper contributes a critical-transition-based data-sampling strategy that can be utilized for further analysis, such as patient classification. Moreover, our method outperformed other indicators of critical transition in complex systems, such as temporal autocorrelation and variance.

SUBMITTER: Ghalati PF

PROVIDER: S-EPMC6733794 | biostudies-literature | 2019 Sep

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Critical Transitions in Intensive Care Units: A Sepsis Case Study.

Ghalati Pejman F PF Samal Satya S SS Bhat Jayesh S JS Deisz Robert R Marx Gernot G Schuppert Andreas A

Scientific reports 20190909 1

The progression of complex human diseases is associated with critical transitions across dynamical regimes. These transitions often spawn early-warning signals and provide insights into the underlying disease-driving mechanisms. In this paper, we propose a computational method based on surprise loss (SL) to discover data-driven indicators of such transitions in a multivariate time series dataset of septic shock and non-sepsis patient cohorts (MIMIC-III database). The core idea of SL is to train ...[more]

PMID: 31501451

Similar Datasets

Project description:BackgroundThe prevalence and mortality of sepsis are largely unknown in Turkey, a country with high antibiotic resistance. A national, multicenter, point-prevalence study was conducted to determine the prevalence, causative microorganisms, and outcome of sepsis in intensive care units (ICUs) in Turkey.MethodsA total of 132 ICUs from 94 hospitals participated. All patients (aged > 18 years) present at the participating ICUs or admitted for any duration within a 24-h period (08:00 on January 27, 2016 to 08:00 on January 28, 2016) were included. The presence of systemic inflammatory response syndrome (SIRS), severe sepsis, and septic shock were assessed and documented based on the consensus criteria of the American College of Chest Physicians and Society of Critical Care Medicine (SEPSIS-I) in infected patients. Patients with septic shock were also assessed using the SEPSIS-III definitions. Data regarding demographics, illness severity, comorbidities, microbiology, therapies, length of stay, and outcomes (dead/alive during 30 days) were recorded.ResultsOf the 1499 patients included in the analysis, 237 (15.8%) had infection without SIRS, 163 (10.8%) had infection with SIRS, 260 (17.3%) had severe sepsis without shock, and 203 (13.5%) had septic shock. The mortality rates were higher in patients with severe sepsis (55.7%) and septic shock (70.4%) than those with infection alone (24.8%) and infection + SIRS (31.2%) (p < 0.001). According to SEPSIS-III, 104 (6.9%) patients had septic shock (mortality rate, 75.9%). The respiratory system (71.6%) was the most common site of infection, and Acinetobacter spp. (33.7%) were the most common isolated pathogen. Approximately, 74.9%, 39.1%, and 26.5% of Acinetobacter, Klebsiella, and Pseudomonas spp. isolates, respectively, were carbapenem-resistant, which was not associated with a higher mortality risk. Age, acute physiology and chronic health evaluation II score at ICU admission, sequential organ failure assessment score on study day, solid organ malignancy, presence of severe sepsis or shock, Candida spp. infection, renal replacement treatment, and a nurse-to-patient ratio of 1:4 (compared with a nurse-to-patient ratio of 1:2) were independent predictors of mortality in infected patients.ConclusionsA high prevalence of sepsis and an unacceptably high mortality rate were observed in Turkish ICUs. Although the prevalence of carbapenem resistance was high in Turkish ICUs, it was not associated with a higher risk for mortality.Trial registrationClinicalTrials.gov ID NCT03249246 . Date: August 15, 2017. Retrospectively registered.

Project description:INTRODUCTION:Many individual Intensive Care Unit (ICU) characteristics have been associated with patient outcomes, including staffing, expertise, continuity and team structure. Separately, many aspects of clinical care in ICUs have been operationalized through the development of complex treatment protocols. The United State Critical Illness and Injury Trials Group-Critical Illness Outcomes Study (USCIITG-CIOS) was designed to determine whether the extent of protocol availability and use in ICUs is associated with hospital survival in a large cohort of United States ICUs. Here, we describe the study protocol and analysis plan approved by the USCIITG-CIOS Steering Committee. METHODS:USCIITG-CIOS is a prospective, observational, ecological multi-centered "cohort" study of mixed ICUs in the U.S. The data collected include organizational information for the ICU (e.g., protocol availability and utilization, multi-disciplinary staffing assessment) and patient level information (e.g. demographics, acute and chronic medical conditions). The primary outcome is all-cause hospital mortality, with the objective being to determine whether there is an association between protocol number and hospital mortality for ICU patients. USCIITG-CIOS is powered to detect a 3% difference in crude hospital mortality between high and low protocol use ICUs, dichotomized according to protocol number at the median. The analysis will utilize regression modeling to adjust for outcome clustering by ICU, with secondary linear analysis of protocol number and mortality and a variety of a priori planned ancillary studies. There are presently 60 ICUs participating in USCIITG-CIOS to enroll approximately 6,000 study subjects. CONCLUSIONS:USCIITG-CIOS is a large multicentric study examining the effect of ICU protocol use on patient outcomes. The primary results of this study will inform our understanding of the relationship between protocol availability, use, and patient outcomes in the ICU. Moreover, given the shortage of intensivists worldwide, the results of USCIITG-CIOS can be used to promote more effective ICU and care team design and will impact the delivery of intensive care services beyond individual practitioners. TRIAL REGISTRATION:ClinicalTrials.gov Identifier NCT01109719.

Project description:ObjectivesTo assess the compliance of Asian intensive care units and hospitals to the Surviving Sepsis Campaign's resuscitation and management bundles. Secondary objectives were to evaluate the impact of compliance on mortality and the organisational characteristics of hospitals that were associated with higher compliance.DesignProspective cohort study.Setting150 intensive care units in 16 Asian countries.Participants1285 adult patients with severe sepsis admitted to these intensive care units in July 2009. The organisational characteristics of participating centres, the patients' baseline characteristics, the achievement of targets within the resuscitation and management bundles, and outcome data were recorded.Main outcome measureCompliance with the Surviving Sepsis Campaign's resuscitation (six hours) and management (24 hours) bundles.ResultsHospital mortality was 44.5% (572/1285). Compliance rates for the resuscitation and management bundles were 7.6% (98/1285) and 3.5% (45/1285), respectively. On logistic regression analysis, compliance with the following bundle targets independently predicted decreased mortality: blood cultures (achieved in 803/1285; 62.5%, 95% confidence interval 59.8% to 65.1%), broad spectrum antibiotics (achieved in 821/1285; 63.9%, 61.3% to 66.5%), and central venous pressure (achieved in 345/870; 39.7%, 36.4% to 42.9%). High income countries, university hospitals, intensive care units with an accredited fellowship programme, and surgical intensive care units were more likely to be compliant with the resuscitation bundle.ConclusionsWhile mortality from severe sepsis is high, compliance with resuscitation and management bundles is generally poor in much of Asia. As the centres included in this study might not be fully representative, achievement rates reported might overestimate the true degree of compliance with recommended care and should be interpreted with caution. Achievement of targets for blood cultures, antibiotics, and central venous pressure was independently associated with improved survival.

Dataset Information

Critical Transitions in Intensive Care Units: A Sepsis Case Study.

Publications

Critical Transitions in Intensive Care Units: A Sepsis Case Study.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets