Pediatric Severe Sepsis: Current Trends and Outcomes From the Pediatric Health Information Systems Database

Amanda Ruth, MD; Courtney E. McCracken, PhD; James D. Fortenberry, MD, MCCM; Matthew Hall, PhD; Harold K. Simon, MD, MBA; Kiran B. Hebbar, MD, FCCM


Pediatr Crit Care Med. 2014;15(9):828-838. 

In This Article

Abstract and Introduction


Objective To 1) describe the characteristics and outcomes over time of PICU patients with severe sepsis within the dedicated U.S. children's hospitals, 2) identify patient subgroups at risk for mortality from pediatric severe sepsis, and 3) describe overall pediatric severe sepsis resource utilization.

Design Retrospective review of a prospectively collected multi-institutional children's hospital database.

Setting PICUs in 43 U.S. children's hospitals.

Patients PICU patients from birth to younger than 19 years were identified with severe sepsis by modified Angus criteria and International Classification of Diseases, 9th Revision, codes for severe sepsis and septic shock.

Interventions None.

Measurements and Main Results Data from the Pediatric Health Information System database collected by the Children's Hospital Association from 2004 to 2012. Pediatric severe sepsis was defined by 1) International Classification of Diseases, 9th Revision, codes reflecting severe sepsis and septic shock and 2) International Classification of Diseases, 9th Revision, codes of infection and organ dysfunction as defined by modified Angus criteria. From 2004 to 2012, 636,842 patients were identified from 43 hospitals. Pediatric severe sepsis prevalence was 7.7% (49,153) with an associated mortality rate of 14.4%. Age less than 1 year (vs age 10 to < 19) (odds ratio, 1.4), underlying cardiovascular condition (odds ratio, 1.4) and multiple organ dysfunction, conferred higher odds of mortality. Resource burden was significant with median hospital length of stay of 17 days (interquartile range, 8–36 d) and PICU length of stay of 7 days (interquartile range, 2–17 d), with median cost/day of $4,516 and median total hospitalization cost of $77,446. There was a significant increase in the severe sepsis prevalence rate from 6.2% to 7.7% from 2004 to 2012 (p < 0.001) and a significant decrease in mortality from 18.9% to 12.0% (p < 0.001). Center mortality was negatively correlated with prevalence (r s = –0.48) and volume (r s = –0.39) and positively correlated with cost (r s = 0.36).

Conclusions In this largest reported pediatric severe sepsis cohort to date, prevalence increased from 2004 to 2012 while associated mortality decreased. Age, cardiovascular comorbidity, and organ dysfunction were significant prognostic factors. Pediatric severe sepsis remains an important cause for PICU admission and mortality and leads to a substantial burden in healthcare costs. Individual center's prevalence and volume are associated with improved outcomes.


Pediatric severe sepsis (PSS) has been a leading cause of morbidity and mortality for infants and children in the United States. Bacterial sepsis of the newborn and septicemia have remained among the top 10 leading causes of death in children 0–14 years old.[1] Previous epidemiologic studies have evaluated severe sepsis in neonates and children in broad state-level databases.[2–4] Utilizing population-based Washington state discharge data not limited to PICU patients, Czaja et al[2] found a 6.8% early mortality rate in 7,183 patients with PSS between 1990 and 2004. In 2003, Watson et al[3] reported analysis of a larger database derived from hospital discharge registries from seven states in 1995, noting a neonatal/pediatric sepsis incidence of 0.56 cases/1,000 children. An updated review of the same databases in 2012 with data reported for 2000 and 2005 found increased PSS prevalence and decreased mortality.[4] Although valuable, data from these studies were not exclusive to children's hospitals or PICUs, and the data collection at 5-year intervals from Weiss et al[5] excluded the ability to perform annual trend analysis.

The last decade has seen major refinements in therapy of severe sepsis. Strategies such as early administration of antibiotics[6] and goal-directed therapy[7–10] have been accepted as standards of practice, although use in children has generally been extrapolated from adult studies. In 2002, the Surviving Sepsis Campaign was launched with the goal to reduce mortality from sepsis by 25% in 5 years.[11] An expert panel of the American College of Critical Care Medicine (ACCM) formulated standard guidelines in 2003 for the management of pediatric and neonatal severe sepsis and septic shock,[12] with subsequent revision in 2013.[13] The impact of these interventions and guidelines on patient outcomes in recent years remains uncertain.

One major difficulty in evaluating sepsis prevalence and outcomes across institutions is the problem of establishing a reliable case definition of severe sepsis. The use of International Classification of Diseases, 9th Revision (ICD-9) codes for severe sepsis have been shown to lack sensitivity in comparison with other methods of capturing cases of severe sepsis in adults.[14] A number of studies of adult and pediatric sepsis have employed the American College of Chest Physicians/Society of Critical Care Medicine definition of an ICD-9 infection code plus organ dysfunction, as validated by Angus et al,[15] and subsequently modified for pediatric patients.[5] Inconsistencies between various methods for identifying sepsis in administrative data remain.[14,16] Therefore, evaluation of pediatric sepsis data could benefit from using several definitions of PSS.

The primary objective of this study is to describe the characteristics of a large cohort of PICU patients with PSS from 2004 to 2012, utilizing several administrative definitions of severe sepsis. We used the Pediatric Healthcare Information Systems (PHIS) database, the nation's largest pediatric hospital-specific administrative data registry maintained by the Children's Hospital Association (CHA). The PHIS registry provides clinical and administrative data from dedicated children's hospitals, which generally provide state-of-the-art pediatric care for the most complex pediatric patients. Although the data are not population based, as state discharge data are, PHIS offers other advantages of providing detailed information about a select but informative sample of the nature of severe illness and care for children. PHIS data have been used in several studies evaluating epidemiologic characteristics and resource utilization in other pediatric conditions.[17,18] Further study objectives were to describe trends in outcomes of patients with PSS, to identify patient subgroups at risk for mortality from PSS, and to report overall PSS resource utilization. We hypothesized that PSS in children's hospital PICUs occurred with increasing prevalence and with increasing associated comorbidities, resource burden, and mortality over the time period studied.