Dataset Information

Sample size implications of mortality definitions in sepsis: a retrospective cohort study.

ABSTRACT:

Background

Many randomized controlled trials (RCTs) employ mortality at a given time as a primary outcome. There are at least three common ways to measure 90-day mortality: first, all-location mortality, that is, all-cause mortality within 90 days of randomization at any location. Second, ARDSnet mortality is death in a healthcare facility of greater intensity than the patient was in prior to the hospitalization during which they were randomized. Finally, in-hospital mortality is death prior to discharge from the primary hospitalization of randomization. Data comparing the impact of these different measurements on sample size are lacking. We evaluated the extent to which event rates vary by mortality definition.

Methods

This was a retrospective cohort study of 30,691 patients hospitalized at Veterans Affairs (VA) hospitals for sepsis during 2009. 12,727 (41.5%) received care in an ICU setting. For each patient, we measured event rates for three different 90-day mortality outcomes: all-location mortality, ARDSnet mortality, and in-hospital mortality. We also calculated sample sizes necessary to power an example RCT given those event rates.

Results

At 90 days, all-location mortality was 26.4% (95% CI 25.9-26.9%), ARDSnet mortality was 19.2% (95% CI 18.8-19.7%), and in-hospital mortality was 13.4% (95% CI 13.0-13.8%) (p < 0.01 all comparisons). These respective event rates result in different required sample sizes to achieve a 20% relative reduction in mortality with 80% power and a 5% false positive rate. Such a trial of VA sepsis patients would require 2080 patients for all-location mortality, 3080 for ARDSnet mortality, and 4796 for in-hospital mortality. Among sepsis patients mechanically ventilated in an ICU, 2438 experienced all-location mortality (46.2% [95% CI 44.8-47.5%]), 2181 experienced ARDSnet mortality (41.3% [95% CI 40.0-42.6%]), and 1894 experienced in-hospital mortality (36.0% [95% CI 34.7-37.3%]).

Conclusions

Event rates vary substantially in sepsis patients based on the chosen 90-day mortality definition. This could have important implications for RCT design trade-offs.

SUBMITTER: Govindan S

PROVIDER: S-EPMC5870299 | biostudies-literature | 2018 Mar

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Sample size implications of mortality definitions in sepsis: a retrospective cohort study.

Govindan Sushant S Prescott Hallie C HC Chopra Vineet V Iwashyna Theodore J TJ

Trials 20180327 1

<h4>Background</h4>Many randomized controlled trials (RCTs) employ mortality at a given time as a primary outcome. There are at least three common ways to measure 90-day mortality: first, all-location mortality, that is, all-cause mortality within 90 days of randomization at any location. Second, ARDSnet mortality is death in a healthcare facility of greater intensity than the patient was in prior to the hospitalization during which they were randomized. Finally, in-hospital mortality is death p ...[more]

PMID: 29587803

Similar Datasets

Project description:The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3) present clinical criteria for the classification of patients with sepsis. We investigated incidence and long-term outcomes of patients diagnosed with these classifications, which are currently unknown.We did a retrospective analysis using data from 30?239 participants from the USA who were aged at least 45 years and enrolled in the Reasons for Geographic and Racial Differences in Stroke (REGARDS) cohort. Patients were enrolled between Jan 25, 2003, and Oct 30, 2007, and we identified hospital admissions from Feb 5, 2003, to Dec 31, 2012, and applied three classifications: infection and systemic inflammatory response syndrome (SIRS) criteria, elevated sepsis-related organ failure assessment (SOFA) score from Sepsis-3, and elevated quick SOFA (qSOFA) score from Sepsis-3. We estimated incidence during the study period, in-hospital mortality, and 1-year mortality.Of 2593 first infection events, 1526 met SIRS criteria, 1080 met SOFA criteria, and 378 met qSOFA criteria. Incidence was 8·2 events (95% CI 7·8-8·7) per 1000 person-years for SIRS, 5·8 events (5·4-6·1) per 1000 person-years for SOFA, and 2·0 events (1·8-2·2) per 1000 person-years for qSOFA. In-hospital mortality was higher for patients with an elevated qSOFA score (67 [23%] of 295 patients died) than for those with an elevated SOFA score (125 [13%] of 960 patients died) or who met SIRS criteria (128 [9%] of 1392 patients died). Mortality at 1 year after discharge was also highest for patients with an elevated qSOFA score (29·4 deaths [95% CI 22·3-38·7] per 100 person-years) compared with those with an elevated SOFA score (22·6 deaths [19·2-26·6] per 100 person-years) or those who met SIRS criteria (14·7 deaths [12·5-17·2] per 100 person-years).SIRS, SOFA, and qSOFA classifications identified different incidences and mortality. Our findings support the use of the SOFA and qSOFA classifications to identify patients with infection who are at elevated risk of poor outcomes. These classifications could be used in future epidemiological assessments and studies of patients with infection.National Institute for Nursing Research, National Center for Research Resources, and National Institute of Neurological Disorders and Stroke.

Project description:BackgroundEarly and reliable identification of patients with sepsis who are at high risk of mortality is important to improve clinical outcomes. However, 3 major barriers to artificial intelligence (AI) models, including the lack of interpretability, the difficulty in generalizability, and the risk of automation bias, hinder the widespread adoption of AI models for use in clinical practice.ObjectiveThis study aimed to develop and validate (internally and externally) a conformal predictor of sepsis mortality risk in patients who are critically ill, leveraging AI-assisted prediction modeling. The proposed approach enables explaining the model output and assessing its confidence level.MethodsWe retrospectively extracted data on adult patients with sepsis from a database collected in a teaching hospital at Beth Israel Deaconess Medical Center for model training and internal validation. A large multicenter critical care database from the Philips eICU Research Institute was used for external validation. A total of 103 clinical features were extracted from the first day after admission. We developed an AI model using gradient-boosting machines to predict the mortality risk of sepsis and used Mondrian conformal prediction to estimate the prediction uncertainty. The Shapley additive explanation method was used to explain the model.ResultsA total of 16,746 (80%) patients from Beth Israel Deaconess Medical Center were used to train the model. When tested on the internal validation population of 4187 (20%) patients, the model achieved an area under the receiver operating characteristic curve of 0.858 (95% CI 0.845-0.871), which was reduced to 0.800 (95% CI 0.789-0.811) when externally validated on 10,362 patients from the Philips eICU database. At a specified confidence level of 90% for the internal validation cohort the percentage of error predictions (n=438) out of all predictions (n=4187) was 10.5%, with 1229 (29.4%) predictions requiring clinician review. In contrast, the AI model without conformal prediction made 1449 (34.6%) errors. When externally validated, more predictions (n=4004, 38.6%) were flagged for clinician review due to interdatabase heterogeneity. Nevertheless, the model still produced significantly lower error rates compared to the point predictions by AI (n=1221, 11.8% vs n=4540, 43.8%). The most important predictors identified in this predictive model were Acute Physiology Score III, age, urine output, vasopressors, and pulmonary infection. Clinically relevant risk factors contributing to a single patient were also examined to show how the risk arose.ConclusionsBy combining model explanation and conformal prediction, AI-based systems can be better translated into medical practice for clinical decision-making.

Dataset Information

Sample size implications of mortality definitions in sepsis: a retrospective cohort study.

Background

Methods

Results

Conclusions

Publications

Sample size implications of mortality definitions in sepsis: a retrospective cohort study.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets