Improving Pediatric Sepsis Outcomes

Improving Pediatric Sepsis Outcomes

Change Package

Measurement recommendations

Measuring sepsis epidemiology, key processes, and outcomes is critical to improving care and sustaining improved outcomes. Data allows teams to readily identify gaps and target improvement efforts.

This section provides recommendations for:

  • Defining sepsis for quality improvement (QI)
  • Choosing key performance measures
  • Approximating sepsis onset
  • Tracking and benchmarking

Defining sepsis for quality improvement

The Surviving Sepsis campaign defines septic shock as “severe infection leading to cardiovascular dysfunction (including hypotension, need for treatment with a vasoactive medication, or impaired perfusion)” and sepsis-associated organ dysfunction as “severe infection leading to cardiovascular and/or non-cardiovascular organ dysfunction” (Weiss et al, 2020b). These definitions are used broadly in real-time clinical practice to inform care decisions but are difficult to operationalize in data collection, limiting their utility in quality improvement.

Furthermore, there is no widely accepted standard definition for pediatric sepsis appropriate for quality improvement, making it difficult to track epidemiology and measure improvement. Work in this area continues to evolve rapidly. To evaluate progress over time, hospitals must reach local consensus on a standard definition. In addition, hospitals should consider data abstraction capabilities, areas of QI focus, and local or state reporting requirements.

Several sepsis definitions have been used in pediatric QI work and research. The IPSO collaborative used a retrospective, intention-to-treat criteria set to define the cohort (Scott et al, 2020). This definition was successfully standardized across the IPSO collaborating hospitals and is the cohort in which the IPSO bundle of care demonstrated improved outcomes. Therefore, if local capacity and infrastructure allow, we recommend adopting the IPSO sepsis definition.

The table below includes multiple proposed sepsis definitions along with their pros and cons for quality improvement work. This list is not exhaustive and continues to evolve. New pediatric sepsis definitions are expected in the next few years.

Note: With any definition of sepsis, hospitals will encounter episodes which originated from an outside hospital. (IPSO specifically defined these as episodes with a time zero within 24 hours of transfer to your hospital.) These can complicate performance analysis given your institution’s limited ability to impact timeliness of recognition and interventions. Hospitals may consider excluding this subset of patients from patient identification and analysis or implementing filters to evaluate this cohort independently.

Table 2
Pediatric sepsis definitions
Sepsis Definition Description Pros Cons

IPSO Sepsis
(Scott et al, 2020)

Intention-to-treat based plus International Classification of Diseases (ICD) codes
  • Developed for QI
  • Adapted and used across multiple institutions
  • Demonstrates strong content, criterion, and convergent construct validity (Scott et al, 2020)
  • Automatable from most EHRs
  • Feasible for large-scale data abstraction
  • Weakness in reliability (Scott et al, 2020): captures some patients who do not go on to develop sepsis
  • Significant initial burden to implement; however, reliable once established

IPSO Critical (subgroup of IPSO Sepsis) (Scott et al, 2020)

Intention-to-treat based plus ICD codes (IPSO Sepsis criteria plus third bolus or vasoactive medication)
  • Developed for QI
  • Adapted and used across multiple institutions
  • Demonstrates strong content, criterion, and convergent construct validity (Scott et al, 2020)
  • Automatable from most EHRs
  • Captures a sicker population
  • Weakness in reliability (Scott et al, 2020): may still capture some patients without sepsis
  • Significant initial burden to implement; however, reliable once established

ICD codes (Balamuth et al, 2015)

ICD codes (for septic shock, sepsis, or infection plus organ dysfunction)
  • Easily abstractable from the EHR
  • May under-capture sepsis, especially if septic shock codes (instead of infection plus organ dysfunction) are used

Pediatric Sequential Organ Failure Assessment (pSOFA)
(Matics & Sanchez-Pinto, 2017)
Organ dysfunction based
  • Validated internationally
  • Developed in the ICU setting only
  • May ultimately exclude patients who present with sepsis, were recognized early, were treated appropriately, and did not progress to a shock state; however, given the impact of early interventions on outcomes, it is important to capture these suspected cases of sepsis in QI work
  • Requires ability to capture organ dysfunctions that are complex and may not be automatable

Children’s Hospital of Philadelphia surveillance definition (Weiss et al, 2020a)
Organ dysfunction based
  • Developed across hospital settings (ED, inpatient, ICU)
  • Developed in a single center
  • May ultimately exclude patients who present with sepsis, were recognized early, treated appropriately, and did not progress to a shock state; however, given the impact of early interventions on outcomes, it is important to capture these suspected cases of sepsis in QI work
  • Requires ability to capture organ dysfunctions that are complex and may not be automatable

International Consensus Criteria for Pediatric Sepsis and Septic Shock, 2024 (“Phoenix criteria”) (Schlapbach et al, 2024)

Organ dysfunction based
  • Derived and validated internationally
  • Not intended (and should not be used) for screening or early identification
  • May ultimately exclude patients who present with sepsis, were recognized early, treated appropriately, and did not progress to a shock state; however, given the impact of early interventions on outcomes, it is important to capture these suspected cases of sepsis in QI work
  • Requires ability to capture organ dysfunctions that are complex and may not be automatable

International pediatric sepsis consensus conference: Definitions for sepsis and organ dysfunction in pediatrics, 2005 (“Goldstein criteria”) (Goldstein et al, 2005)

Vital sign and organ dysfunction based

Intended for clinical trials
  • Well-known/familiar to clinicians
  • Used in previous pediatric sepsis studies
  • May be useful for consistency
  • Consensus based
  • Frequently used in various modified forms due to challenges with shock definition (requires prerequisite of 40ml/kg bolus)
  • Requires ability to capture organ dysfunctions that are complex and may not be automatable
  • May become obsolete given newer consensus definitions

Surviving Sepsis (Weiss et al, 2020b)

Clinical definition emphasizing altered perfusion and organ dysfunction
  • Useful in clinical practice
  • Based on clinical variables with subjectivity (perfusion)
  • May not be automatically abstractable from EHRs

View chart.

Key performance measures

Due to the complex nature of pediatric sepsis care, measuring performance is challenging. Evaluating the timeliness of interventions requires a standardized approach to approximating the time of sepsis onset (see Approximating Sepsis Onset). In addition, sepsis outcome data may be scarce due to the relatively low incidence of mortality among children within an individual institution.

Despite the complexities, tracking performance data over time is vital to driving and sustaining improvement. When choosing key performance measures to track, hospitals must consider local factors that influence data availability, such as automatability from the EHR, local reporting requirements, and hospital strategic priorities, as well as external factors such as standardized definitions to facilitate comparative benchmarking.

The table below suggests outcome, process, and balancing measures for sepsis improvement work, as well as recommendations by both IPSO and the Centers for Disease Control and Prevention (CDC) for operationalizing each measure. The table lists top priorities and additional measures, but there are many more measures hospitals could track. Note that some states may have additional regulatory requirements to consider. Additionally, a standardized, system-wide approach to collecting social drivers of health data on children with sepsis is imperative to evaluating the equity of sepsis improvement initiatives.

Find comprehensive lists of sepsis performance measures from the IPSO collaborative and the CDC Hospital Sepsis Program Core Elements.

Table 3
Pediatric sepsis performance measures (priority and additional measures)
Outcome measures
IPSO Collaborative CDC Sepsis Score Elements

Mortality

  • Sepsis-attributable (SA)
  • Overall, in-hospital
  • 3-day and 30-day SA mortality
  • In-hospital mortality, overall
  • In-hospital mortality, subgroup (community-onset, hospital-onset, septic shock, etc)

Sepsis Epidemiology

  • Count
  • Severity
  • Community-onset vs hospital acquired
  • Sepsis pathogens
  • Hospital-onset IPSO critical sepsis
  • Incidence of IPSO critical sepsis/1000 hospital admissions
  • Rate of hospital-onset sepsis
  • Rate of sepsis episodes
  • Rate of sepsis episodes w/ and w/out shock

View chart.


Process measures (bundle compliance*)

IPSO Collaborative CDC Sepsis Score Elements

Recognition/Identification (examples)

  • Compliance with screens (sepsis triggers or alert system)
  • Compliance with huddles
  • Utilization of standard pathways or order sets
  • Percent trigger activations
  • Percent huddle activations
  • Percent order set utilization
  • Use of sepsis order sets

Management

  • Fluid bolus timeliness
  • Antibiotic timeliness
  • Blood culture prior to antibiotics (additional)
  • Check lactate (additional)
  • Vasopressor timeliness (additional)
  • Time to first fluid bolus
  • Time to first antibiotic
  • Time to first vasopressor
  • Time to antibiotics in community-onset sepsis with hypotension
  • Time from antibiotic order to administration
  • Proportion receiving fluid resuscitation in sepsis episodes with shock
  • Fluid bolus type and timeliness (additional)

All-or-none bundle compliance
  • Bundle compliance

N/A

View chart.


Balancing measures
IPSO Collaborative CDC Sepsis Score Elements

Antibiotic stewardship
  • Total IV antibiotic days
  • Antibiotic choice (additional)
    Days to narrowing (additional)

*bundle compliance = recognition + bolus in 60 minutes + antibiotic in 180 minutes

View chart.

Approximating sepsis onset

To accurately assess time-bound sepsis metrics (such as time to first fluid bolus or antibiotic), teams must standardize how they approximate the beginning of a sepsis episode. For this, IPSO implemented a functional time zero definition. This prospective approach leveraged EHR surrogates used in measure calculations.

IPSO functional time zero

Due to lessons learned during IPSO—including that huddles and order sets sometimes occur before a sepsis screen and that it may be difficult to determine functional time zero for outside hospital transfers—we propose the revised functional time zero logic below.

Functional time zero is:

  1. The earliest time of screen, huddle, or order set (if any is reported)
  2. Otherwise, the earlier of first antibiotic time or first bolus time (if either is reported)
  3. Else emergency department or hospital arrival time (if community onset case)
  4. Else cannot be determined or you may use some other proxy for time zero (such as transfer to ICU time)

View IPSO’s time zero cheat sheet.

Functional time zero becomes the basis of comparison for other reported values. If functional time zero cannot be determined, exclude these episodes from measure calculations where functional time zero is used.

Optional: Clinically derived time zero

IPSO also tracked the time of physiological sepsis onset. This process was optional and required manual chart review. In IPSO, we recommended using the Goldstein criteria for this determination. However, alternative sepsis definitions developed since then may be considered as long as they are used consistently. Because clinically derived time zero is retrospective, it is not a potential functional time zero. Clinically derived time zero can be compared to functional time zero to measure the gap between physiological onset of sepsis and recognition of sepsis.

The CDC Hospital Sepsis Program Core Elements recommends time of emergency department or hospital arrival for community onset cases.

Tracking and benchmarking

Hospitals should establish a standardized local process of abstracting data and tracking identified measures. For examples of sepsis dashboards from IPSO hospitals, see Dashboards.

Implementing IPSO’s standardized patient identification and measure definitions allows for benchmarking against peer hospitals. Members of Children’s Hospital Association can participate in sepsis data tracking through the Pediatric Health Information System® (PHIS®) and Inpatient Essentials (IE) databases.

Download the Change Package

For citations, see our references page. This change package was created in March 2025 by Children’s Hospital Association quality improvement consultants and Improving Pediatric Sepsis Outcomes thought leaders and reflects best evidence to date at the time of publication. Pediatric sepsis evidence is always evolving, and readers should make every effort to ensure incorporation of the latest best evidence during implementation of sepsis improvement projects.