Latent class analysis of pediatric patients with sepsis at community hospitals
CCCF ePoster library. Evans I. 11/13/19; 285184; EP104
Dr. Idris Evans
Dr. Idris Evans
Login now to access Regular content available to all registered users.

You may also access this content "anytime, anywhere" with the Free MULTILEARNING App for iOS and Android
Abstract
Rate & Comment (0)
ePoster
Topic: Retrospective or Prospective Cohort Study or Case Series

Evans, Idris VR1,2; Kennedy, Jason N1,2; Carcillo, Joseph A1; Angus, Derek C,2; Seymour, Christopher W1,2
1 Department of Critical Care Medicine, University of Pittsburgh, Pittsburgh, PA, USA
2 Clinical Research, Investigation, and Systems Modeling of Acute Illness (CRISMA) Center, Pittsburgh, PA, USA


Introduction:  Despite compliance with resuscitation guidelines, 1 out of 10 pediatric patients with sepsis die. The heterogeneity of pediatric sepsis may contribute to this mortality rate; identification of subclasses may allow more precise therapy and improve outcomes. Prior work identifying subclasses of pediatric sepsis used gene-based expression analysis of patients in the intensive care unit (ICU) who have already received treatment. However, it is unknown whether subclasses can be identified earlier in care using readily available clinical data alone.
Objective: To identify subclasses of pediatric sepsis with latent class analysis of clinical data.
Methods: We performed a retrospective analysis of electronic health record (EHR) data from pediatric patients (age <18 years) with sepsis that presented to 12 community hospitals in southwestern Pennsylvania from 2010 to 2015. We derived clinical subclasses using latent class analysis of 14 clinical and laboratory variables obtained in the 24 hours after patients met criteria for sepsis. We describe the subclasses' frequency of clinical characteristics, and correlation with a composite outcome of admission to the ICU, transfer to another acute care facility, or in-hospital death.
Results: Among 1,784 patients with sepsis, we determined that a four-class model (α1, α2, β1, and β2) best fit the data (Vuong-Lo-Mendell-Rubin likelihood ratio test, p-value <0.001). The α1-type (N = 514, 28.8%) and α2-type (N = 303, 17.0%) were young (mean age 3.9 years and 2.4 years, respectively) and the α2-type had more tachycardia and tachypnea (Figure, Panel A). The β1-type (N = 832, 46.6%) and β2-type (N = 135, 7.6%) were older (mean age 15.3 years and 15.2 years, respectively) and the β2-type had more tachycardia and lower platelet count (Figure, Panel B). The composite outcome of ICU admission, transfer to another acute care facility, or in-hospital death was most common among the β2-type (N = 94, 69.6%) compared to the α1-type (N = 142, 27.6%), α2-type (N = 119, 39.3%), and β1-type (N = 169, 20.3%).
Conclusion: At presentation, pediatric sepsis subclasses are feasible to identify using readily available clinical data in a multicenter EHR. The subclasses are notable for differences in age and patterns of physiologic response to infection. Subclasses may inform future evaluation of directed sepsis treatments.
 


Image

No references

ePoster
Topic: Retrospective or Prospective Cohort Study or Case Series

Evans, Idris VR1,2; Kennedy, Jason N1,2; Carcillo, Joseph A1; Angus, Derek C,2; Seymour, Christopher W1,2
1 Department of Critical Care Medicine, University of Pittsburgh, Pittsburgh, PA, USA
2 Clinical Research, Investigation, and Systems Modeling of Acute Illness (CRISMA) Center, Pittsburgh, PA, USA


Introduction:  Despite compliance with resuscitation guidelines, 1 out of 10 pediatric patients with sepsis die. The heterogeneity of pediatric sepsis may contribute to this mortality rate; identification of subclasses may allow more precise therapy and improve outcomes. Prior work identifying subclasses of pediatric sepsis used gene-based expression analysis of patients in the intensive care unit (ICU) who have already received treatment. However, it is unknown whether subclasses can be identified earlier in care using readily available clinical data alone.
Objective: To identify subclasses of pediatric sepsis with latent class analysis of clinical data.
Methods: We performed a retrospective analysis of electronic health record (EHR) data from pediatric patients (age <18 years) with sepsis that presented to 12 community hospitals in southwestern Pennsylvania from 2010 to 2015. We derived clinical subclasses using latent class analysis of 14 clinical and laboratory variables obtained in the 24 hours after patients met criteria for sepsis. We describe the subclasses' frequency of clinical characteristics, and correlation with a composite outcome of admission to the ICU, transfer to another acute care facility, or in-hospital death.
Results: Among 1,784 patients with sepsis, we determined that a four-class model (α1, α2, β1, and β2) best fit the data (Vuong-Lo-Mendell-Rubin likelihood ratio test, p-value <0.001). The α1-type (N = 514, 28.8%) and α2-type (N = 303, 17.0%) were young (mean age 3.9 years and 2.4 years, respectively) and the α2-type had more tachycardia and tachypnea (Figure, Panel A). The β1-type (N = 832, 46.6%) and β2-type (N = 135, 7.6%) were older (mean age 15.3 years and 15.2 years, respectively) and the β2-type had more tachycardia and lower platelet count (Figure, Panel B). The composite outcome of ICU admission, transfer to another acute care facility, or in-hospital death was most common among the β2-type (N = 94, 69.6%) compared to the α1-type (N = 142, 27.6%), α2-type (N = 119, 39.3%), and β1-type (N = 169, 20.3%).
Conclusion: At presentation, pediatric sepsis subclasses are feasible to identify using readily available clinical data in a multicenter EHR. The subclasses are notable for differences in age and patterns of physiologic response to infection. Subclasses may inform future evaluation of directed sepsis treatments.
 


Image

No references

    This eLearning portal is powered by:
    This eLearning portal is powered by MULTIEPORTAL
Anonymous User Privacy Preferences

Strictly Necessary Cookies (Always Active)

MULTILEARNING platforms and tools hereinafter referred as “MLG SOFTWARE” are provided to you as pure educational platforms/services requiring cookies to operate. In the case of the MLG SOFTWARE, cookies are essential for the Platform to function properly for the provision of education. If these cookies are disabled, a large subset of the functionality provided by the Platform will either be unavailable or cease to work as expected. The MLG SOFTWARE do not capture non-essential activities such as menu items and listings you click on or pages viewed.


Performance Cookies

Performance cookies are used to analyse how visitors use a website in order to provide a better user experience.


Save Settings