Hospital-Onset Sepsis Systematic Detection and Structured Interventions (CATCH SEPSIS)

CCAT Hospital-Onset Sepsis Systematic Detection and Structured Interventions

Sepsis is a life-threatening reaction to an infection. It can begin while a person is already in the hospital for some other reason. Sepsis that starts in the hospital is often recognized late, and treatment can be delayed. On the general wards at the study hospital, patients who develop sepsis may benefit from receiving earlier treatments. Patients who are getting sicker but do not have sepsis also might benefit.

The hospital's electronic health record already runs an automated sepsis screening tool in the background. The tool gives each patient a score that rises when the record shows possible signs of sepsis. Today that score can be seen by staff, but it does not prompt any particular action.

This study tests whether connecting the score to a rapid bedside response improves how patients do. When a patient's score crosses a set level, the electronic health record sends an alert to that patient's bedside nurse. The nurse can bring in the Critical Care Activation Team, a group of critical care clinicians available at all hours. If the nurse does not respond within a short window, the team is paged automatically. The team comes to the bedside, examines the patient, and can order tests and treatments directly, including antibiotics. The patient's regular medical team is told at the same time and continues taking care of the patient with this additional assistance.

The study compares this alert-and-response pathway against the care the hospital provides today. Every adult on the participating units is included automatically when their score crosses the set level. A computer assigns each patient to one of two groups. In one group the alert is sent and the response pathway begins. In the other group the alert is not sent, and care goes on exactly as it does at the hospital now.

The main measure is each patient's condition five days after the score crossed the set level, sorted into five levels: discharged from the hospital, in the hospital without organ support, in the hospital with organ support, in intensive care, or died. The study also looks at how quickly antibiotics were started, how many days patients spend alive and out of the hospital, and whether the pathway leads to antibiotic use that was not needed.

Patients are not asked to do anything and are not seen by the study team. No extra visits, tests, or procedures are added beyond the tests the response team may order as part of ordinary care. All study information comes from the medical record. Both groups receive care that is within accepted practice, and asking each patient for permission ahead of time would change the very thing being studied. For those reasons the study is conducted under a waiver of informed consent granted by the hospital's institutional review board, and the hospital notifies patients about its use of these tools through its general admission materials.

Study Overview

• Status: Not yet recruiting
• Conditions: Sepsis; Critical Illness; Early Warning Score; Deterioration, Clinical
• Intervention / Treatment: Other: Algorithm-triggered sepsis alert with structured critical care team activation

Detailed Description

BACKGROUND AND RATIONALE

Hospital-onset sepsis carries higher mortality than sepsis present on admission, and recognition on general wards depends on individual clinician vigilance. At the study institution, approximately 300 adult sepsis events occur annually on the target units. Observed outcomes in this population include >20% inpatient mortality and >50% unplanned transfer to intensive care within 72 hours. The median interval from the sepsis reference time to a new antibiotic order is 15 hours (interquartile range 6 to 23 hours), and fewer than 5 percent of patients receive a new antibiotic within 3 hours. Approximately 55 percent of these patients are already receiving antibiotics at the reference time, so the potential failure for that majority is delayed reassessment and escalation rather than delayed initiation.

The institution already runs Epic Sepsis Model version 2 (ESM2) passively. Scores are visible on configurable patient lists but are not wired to any structured response. The trial therefore does not evaluate the algorithm, whose discrimination has been characterized externally and locally. It evaluates whether linking an existing detection signal to a structured response pathway with independent ordering authority changes patient outcomes. Prior evaluations of machine-learning early-warning systems have generally demonstrated process improvement without consistent outcome benefit, which establishes equipoise for the pathway rather than for the algorithm.

SETTING AND POPULATION

Single-center pragmatic trial at an academic medical center. Target units are non-cancer medicine (hospitalist, resident, and family medicine services), general surgery, and emergency department observation. These units account for approximately 60% of hospital-onset adult sepsis events at the institution. At an ESM2 threshold of 30, approximately 2,500 threshold crossings occur per year across these units, corresponding to roughly 8 alerts per day with an approximately flat distribution across hours of the day. The positive predictive value for sepsis at this threshold is approximately 5 percent, which is anticipated and is addressed analytically rather than by threshold restriction.

INTERVENTION PATHWAY

For patients assigned to the intervention condition, crossing the ESM2 threshold triggers a Best Practice Advisory delivered to the assigned bedside nurse. The advisory presents a small set of structured response options, including declining activation with a captured reason. The architecture is default-to-action: if the nurse does not respond within a pre-specified timeout window, the Critical Care Activation Team is paged automatically. Inaction therefore produces help rather than silence, and the nurse's discretion is exercised by declining rather than by initiating.

The Critical Care Activation Team is an existing 24-hour rapid response service staffed by critical care clinicians credentialed for independent ordering. On activation, the team performs a bedside evaluation, may place diagnostic and therapeutic orders directly, and documents the encounter in a structured note. Each activation is classified into one of five categories: sepsis unrecognized before evaluation; sepsis recognized with bundle incomplete or delayed; sepsis recognized with care already appropriate; not sepsis but acute illness with an alternative diagnosis pursued; and not sepsis with the patient clinically stable. This classification replaces binary true-positive and false-positive framing and is used for fidelity and threshold-refinement analyses.

The primary team is notified simultaneously with activation and retains authority to modify or discontinue any team-initiated order at any point. Overrides occurring before the four-hour mark are documented in the structured note with a reason, and override rate is analyzed as an implementation outcome. A four-hour reassessment by the responding team is required for activations classified into the first two categories and constitutes the formal handoff back to the primary team.

CONTROL CONDITION

For patients assigned to the control condition, the study advisory is suppressed silently. No advisory is displayed, no automatic orders are generated, and no page is sent. Suppression is limited to the study's own advisory and its two downstream actions. No pre-existing alert, alarm, order set, score display, or item of clinical information is altered or withheld in either condition. ESM2 scores remain visible on patient lists exactly as they are today, and the separately deployed deterioration index continues to trigger the existing rapid response pathway at its established thresholds. The care a control-condition patient receives is therefore identical to current standard practice at the institution.

RANDOMIZATION, CONCEALMENT, AND MASKING

Assignment is computed deterministically within the electronic health record at the first qualifying threshold crossing of an admission, with 1:1 random allocation (with block size 2). Subsequent threshold crossings within the same admission inherit the original assignment. The trial is unstratified; unit-level and temporal subgroup effects are addressed analytically.

Concealment holds until the threshold crossing by the determinism of the rule and the silence of the suppression mechanism. At crossing, assignment resolves operationally on the intervention side only. On the control side no human observes the suppressed advisory, and the assignment never becomes operationally visible during care. Patients are unmasked by design. Clinicians are unmasked on the intervention side and effectively masked on the control side, having never been alerted. The analytic dataset is prepared with condition replaced by a study-side code, and the analyst conducting the pre-specified primary analysis remains masked until the locked analysis.

RESPONSE FIDELITY INSTRUMENTATION

An ordered chain of timestamped checkpoints is captured natively in the electronic health record: advisory fired, nurse acknowledgment, team page, team bedside evaluation, per-element order placement with ordering-clinician attribution, and override capture. These support a completion funnel across the pathway and a decomposition of time-to-antibiotic into a detection-and-decision interval and a system-execution interval. An interpretive rule is pre-specified in advance of analysis to separate failure of the concept from failure of implementation on the basis of observed fidelity, so that a null result can be attributed rather than left ambiguous.

IMPLEMENTATION STAGING

The advisory, automatic orders, and paging workflow are first deployed in a non-randomized configuration on pilot units for a minimum of 1 week. Randomization begins only after that period completes without major workflow incident. Nurse education is delivered identically to staff caring for patients in both conditions and is treated as a study artifact with pre-specified scripts, materials, and a post-training assessment. Because education is necessarily uniform across conditions, a secular-trend analysis of time-to-antibiotic among non-alerted threshold-crossing patients is pre-specified as a probe for broader attention effects attributable to education rather than to the pathway.

SAFETY MONITORING

An independent safety monitor who is not a member of the study team reviews unmasked safety data at pre-specified intervals. Pre-specified operational triggers prompt immediate pause and review: a death within 24 hours of an activation classified as clinically stable; a death of a control-condition patient within 24 hours of a suppressed advisory; a cluster of adverse events associated with team-initiated antibiotics; team bedside response exceeding 30 minutes for more than 20 percent of activations in a week; nurse decline rate above 60 percent sustained over two weeks; and a stable-patient classification rate above 50 percent sustained over two weeks.

STATISTICAL APPROACH

The primary analysis is by intention to treat and estimates a common odds ratio on the day-5 clinical-state ordinal using proportional odds logistic regression, with all threshold-crossing patients analyzed by assigned condition regardless of advisory response or team arrival. Sample size was determined by Monte Carlo simulation against the empirically observed control-condition ordinal distribution rather than by analytic approximation, which was found to be optimistic in this setting. Eighteen months of randomization at the observed accrual rate yields approximately 1,840 patients per condition and approximately 86 percent power to detect a common odds ratio of 0.80. Extension to 24 months is reserved as a contingency if realized accrual undershoots projection by more than 10 percent.

A per-protocol sensitivity analysis restricted to patients who received a team evaluation is pre-specified alongside the primary analysis, as is an effect-modification analysis by actionable lead-time window, defined from the distribution of threshold-crossing time relative to first antibiotic order in a retrospective cohort.

REGULATORY FRAMING

The trial is structured as a single integrated protocol with layered determinations. The algorithm and advisory workflow are implemented through routine clinical decision support governance. Patient-level randomization and outcome ascertainment proceed as minimal-risk research under a waiver of informed consent and waiver of documentation of consent. Accompanying clinician interviews and surveys proceed as exempt research with verbal consent.

• Study Type: Interventional
• Enrollment (Estimated): 3700
• Phase: Not Applicable

Contacts and Locations

• Study Contact: Name: Program Team; Phone Number: 503 494-7680; Email: catchsepsis@ohsu.edu

Study Locations

• United States
  • Oregon
    • Portland, Oregon, United States, 97239
      • OHSU Hospital

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: No

Description

Inclusion Criteria:

• Age 18 years or older
• Admitted to a participating study unit at the time of the qualifying event. Participating units are non-cancer medicine wards (hospitalist, resident, and family medicine), general surgery.
• Sepsis prediction score computed by the electronic health record crosses the pre-specified study threshold at any point during the admission

Exclusion Criteria:

• Comfort-focused or hospice goals of care documented at the time of threshold crossing. Do-not-resuscitate or do-not-intubate status alone is not an exclusion; participants with these orders are eligible, and the responding team tailors its response to documented goals of care
• Already receiving intensive-care-level treatment while boarding on a ward unit at the time of threshold crossing
• Pre-existing order declining rapid response team activation (expected to be exceedingly rare)

Participants are identified automatically by the electronic health record at the time of threshold crossing. There is no active recruitment, no study visit, and no contact between participants and the study team. Eligibility is evaluated by rule from electronic health record data, identically in both arms.

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Health Services Research
• Allocation: Randomized
• Interventional Model: Parallel Assignment
• Masking: Single

Number of Arms

2

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
Experimental: Sepsis alert with critical care team activation; Crossing the sepsis algorithm threshold triggers an electronic health record alert (Best Practice Advisory) to the assigned bedside nurse, presenting structured response options including declining activation with a captured reason. Orders for serum lactate and complete blood count with differential are generated automatically. If the nurse does not respond within a pre-specified timeout window, the Critical Care Activation Team is paged automatically. The team performs a bedside evaluation, may place diagnostic and therapeutic orders directly including antibiotics, and documents the encounter in a structured note. The primary team is notified simultaneously and retains authority to modify or discontinue any team-initiated order at any time.	Other: Algorithm-triggered sepsis alert with structured critical care team activation; When a commercially deployed sepsis prediction score crosses a pre-specified threshold, an electronic health record alert is delivered to the assigned bedside nurse, presenting structured response options including declining activation with a captured reason. Serum lactate and complete blood count are ordered automatically. If the nurse does not respond within a pre-specified timeout window, a 24-hour critical care rapid response team is paged automatically. The team evaluates the patient at the bedside, may place diagnostic and therapeutic orders directly including antibiotics, and documents the encounter in a structured note. The primary team is notified simultaneously and retains clinical authority. The prediction score runs identically in both arms and is not modified by the study.; Other Names: CATCH SEPSIS detection-to-response pathway
No Intervention: Usual care with study alert suppressed; Current standard care. The study alert is suppressed silently: no advisory is displayed, no automatic orders are generated, and no page is sent. Suppression is limited to the study's own alert and its two downstream actions. No pre-existing alert, alarm, order set, score display, or item of clinical information is altered or withheld. Algorithm scores remain visible on patient lists exactly as they are today, and the separately deployed deterioration index continues to trigger the existing rapid response pathway at its established thresholds. Clinicians caring for these patients are not informed that a threshold crossing occurred, and assignment is never operationally visible during care.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Clinical status at day 5 on a five-level ordinal scale	Each participant is assigned one ordinal clinical state at 120 hours after the qualifying algorithm threshold crossing. The scale has five mutually exclusive, ordered levels: 1 = discharged alive; 2 = hospitalized on a ward without organ support; 3 = hospitalized on a ward with organ support; 4 = intensive care unit; 5 = dead. Lower values indicate better status. Organ support is any of: vasopressor infusion, invasive mechanical ventilation, non-invasive positive pressure ventilation for acute respiratory failure, high-flow nasal cannula at FiO2 of 0.4 or greater, or renal replacement therapy started after threshold crossing. State is derived by rule from electronic health record data, identically in both arms, using admission-discharge-transfer records for location, an organ-support pipeline, and hospitalization records for mortality. State is ascertained at the closest clean data snapshot to the 120-hour mark. Discharge to hospice is coded as state 5.	120 hours (Day 5) after the qualifying algorithm threshold crossing

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Clinical status at day 14 on a five-level ordinal scale	Each participant is assigned one ordinal clinical state at 336 hours after the qualifying algorithm threshold crossing, using the same five-level scale as the primary outcome: 1 = discharged alive; 2 = hospitalized on a ward without organ support; 3 = hospitalized on a ward with organ support; 4 = intensive care unit; 5 = dead. Lower values indicate better status. Organ support is any of: vasopressor infusion, invasive mechanical ventilation, non-invasive positive pressure ventilation for acute respiratory failure, high-flow nasal cannula at FiO2 of 0.4 or greater, or renal replacement therapy started after threshold crossing. State is derived by rule from electronic health record data, identically in both arms. Discharge to hospice is coded as state 5. This measure assesses persistence of any effect observed at day 5.	336 hours (day 14) after the qualifying algorithm threshold crossing
Days alive and free of organ support through day 28	Number of days within the first 28 days after the qualifying algorithm threshold crossing on which the participant was alive and free of organ support, defined as being in the "ward without organ support" state or better (levels 1 or 2 of the primary ordinal scale). Organ support is any of: vasopressor infusion, invasive mechanical ventilation, non-invasive positive pressure ventilation for acute respiratory failure, high-flow nasal cannula at FiO2 of 0.4 or greater, or renal replacement therapy started after threshold crossing. Death at any point within 28 days scores 0. Range 0 to 28 days; higher values indicate better outcome. Derived by rule from electronic health record data, identically in both arms.	28 days after the qualifying algorithm threshold crossing
Time from threshold crossing to first new antibiotic administration	Hours from the qualifying algorithm threshold crossing to administration of the first new antibiotic, defined as an antibiotic not already being administered at the time of crossing. Analyzed as time to event, with death and hospital discharge without a new antibiotic treated as competing events. Participants are followed for this measure until first new antibiotic administration, hospital discharge, or death, through day 28. All participants who cross the threshold are included in the denominator regardless of whether an antibiotic was ever ordered, so that participants who die or are discharged early are retained rather than excluded. Derived from the medication administration record and order entry timestamps, identically in both arms.	Up to 28 days after the qualifying algorithm threshold crossing
Completion of all one- and three-hour sepsis bundle elements within one hour	Proportion of participants for whom all elements of the Surviving Sepsis Campaign one-hour bundle were completed within one and three hours of the qualifying algorithm threshold crossing. Elements are ascertained by rule from electronic health record data, identically in both arms.	1 and 3 hours after the qualifying algorithm threshold crossing
All-cause mortality at 30 days	Proportion of participants who died from any cause within 30 days of the qualifying algorithm threshold crossing. Ascertained from electronic health record and hospitalization records, identically in both arms. Reported for clinical completeness; the trial is not powered for hypothesis testing on this measure.	30 days after the qualifying algorithm threshold crossing
Transfer to intensive care within 72 hours	Proportion of participants transferred from a ward to an intensive care unit within 72 hours of the qualifying algorithm threshold crossing, derived from admission-discharge-transfer records identically in both arms.	72 hours after the qualifying algorithm threshold crossing
Hospital length of stay among participants meeting sepsis criteria	Days from hospital admission to hospital discharge among the subgroup of participants meeting rule-based sepsis criteria at or within 48 hours of the qualifying algorithm threshold crossing. Length of stay is ascertained through hospital day 90. Sepsis criteria follow the Centers for Disease Control and Prevention Adult Sepsis Event definition, based on blood culture collection, qualifying antibiotic days, and organ dysfunction signals, and are computed by rule from electronic health record data identically in both arms.	From hospital admission to hospital discharge, assessed through day 90
New multidrug-resistant organism through hospital day 90	Proportion of participants with a first multidrug-resistant organism flagged for isolation precautions more than 48 hours after the qualifying algorithm threshold crossing and through hospital day 90. Ascertained from the institutional infection control isolation log, which operates identically for both arms. This measure tracks whether the intervention changes the incidence of resistant organisms; the direction of effect is not hypothesized in advance.	From 48 hours after the qualifying algorithm threshold crossing through hospital day 90
New Clostridioides difficile infection through hospital day 90	Proportion of participants with a first positive Clostridioides difficile test result, or first initiation of Clostridioides difficile treatment, more than 48 hours after the qualifying algorithm threshold crossing and through hospital day 90. Ascertained from laboratory and pharmacy records identically in both arms. This measure tracks a potential harm of increased antibiotic exposure.	From 48 hours after the qualifying algorithm threshold crossing through hospital day 90

Other Outcome Measures

Outcome Measure	Measure Description	Time Frame
Time from alert to critical care activation team bedside evaluation	Minutes from delivery of the electronic health record alert to documented bedside evaluation by the critical care team, derived from paging records and the structured team note. This measure is computable only for participants in the intervention arm, because the alert, the page, and the team note exist only when an alert is generated. It is reported as an implementation fidelity measure for that arm rather than as a between-arm comparison.	Up to 24 hours after the qualifying algorithm threshold crossing
Distribution of bedside nurse responses to the alert	Distribution of bedside nurse responses to the alert across the available structured options, including declining activation with a captured reason, requesting team activation, and automatic activation following an unanswered alert. Reported as proportions per calendar month across the 18-month randomization period. Computable only for the intervention arm, because alerts are not generated in the control arm. Interpreted as an indicator of how nurses weigh alert salience against the clinical picture; a high decline rate may reflect clinical judgment, alert fatigue, or both.	Up to 18 months
Proportion of team-initiated orders modified by the primary team	Proportion of orders placed by the critical care rapid response team that were modified or discontinued by the primary team before the required four-hour reassessment, with reason captured in the structured team note. Computable only for the intervention arm. Reported as a governance and implementation fidelity measure.	Within 4 hours of team activation
Time to first new antibiotic among control-arm participants, compared with pre-launch baseline	Median hours from the qualifying algorithm threshold crossing to first new antibiotic administration among control-arm participants, summarized per calendar quarter across the 18-month randomization period and compared with the institutional pre-launch baseline for the same units. This measure probes whether uniform staff education and general awareness of the study shift care for participants whose alert was suppressed. A secular trend toward faster antibiotics in the control arm would indicate an attention effect not attributable to the alert-and-response pathway itself.	Up to 18 months
Incremental hospital charges per alert-triggered activation	Total hospital charges attributable to the activation encounter, summarized by activation category, with a sensitivity analysis applying cost-to-charge ratios. Derived from the hospital billing extract. Computable only for the intervention arm, because charges are keyed to an activation encounter that occurs only when an alert is generated. Reported as an input to budget impact analysis.	Through hospital discharge, up to 90 days
Antibiotic escalation within 24 hours among participants already receiving antibiotics	Proportion of participants who were already receiving antibiotics at the qualifying algorithm threshold crossing and for whom an antibiotic was added, or coverage changed to broader spectrum, within 24 hours of crossing. Keyed to threshold crossing rather than to team activation so that the measure is computable identically in both arms. Derived from pharmacy and order records. The hypothesized direction is an increase, reflecting the value of structured reassessment for participants who may be inadequately covered.	24 hours after the qualifying algorithm threshold crossing
Antibiotic discontinuation or de-escalation within 24 hours among participants already receiving antibiotics	Proportion of participants who were already receiving antibiotics at the qualifying algorithm threshold crossing and for whom antibiotic coverage was discontinued or narrowed within 24 hours of crossing. Keyed to threshold crossing rather than to team activation so that the measure is computable identically in both arms. Derived from pharmacy and order records. The hypothesized direction is an increase among participants without rule-based evidence of confirmed infection. Analyzed separately from escalation so that the two directions are not netted against each other.	24 hours after the qualifying algorithm threshold crossing
Distribution of critical care team activation categories	Distribution of team activations across five mutually exclusive categories: sepsis unrecognized before evaluation; sepsis recognized with bundle incomplete or delayed; sepsis recognized with care already appropriate; not sepsis but acute illness with an alternative diagnosis pursued; and not sepsis with the participant clinically stable. Summarized per calendar month across the 18-month randomization period. Computable only for the intervention arm. The proportion falling in the final category is monitored as a signal for threshold refinement.	Up to 18 months
Antibiotics discontinued within 24 hours following a non-sepsis activation category	Proportion of team-initiated antibiotics discontinued within 24 hours among activations classified as not sepsis, whether an alternative diagnosis was pursued or the participant was clinically stable. Computable only for the intervention arm, since the classification derives from the structured team note. Reported as an antimicrobial stewardship monitoring measure.	24 hours after team activation

Collaborators and Investigators

• Sponsor: Patricks Lyons

Study record dates

Study Major Dates

• Study Start (Estimated): October 1, 2026
• Primary Completion (Estimated): December 1, 2027
• Study Completion (Estimated): December 1, 2027

Study Registration Dates

• First Submitted: August 23, 2026
• First Submitted That Met QC Criteria: September 2, 2026
• First Posted (Actual): September 4, 2026

Study Record Updates

• Last Update Posted (Actual): September 4, 2026
• Last Update Submitted That Met QC Criteria: September 2, 2026
• Last Verified: September 1, 2026

More Information

Terms related to this study

• Keywords: sepsis; clinical deterioration; early warning systems
• Additional Relevant MeSH Terms: Pathologic Processes; Disease Attributes; Infections; Systemic Inflammatory Response Syndrome; Inflammation; Disease Progression; Pathological Conditions, Signs and Symptoms; Clinical Deterioration; Critical Illness; Sepsis
• Other Study ID Numbers: STUDY00030267

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO

IPD Plan Description

Individual participant data will not be shared. Participants are enrolled under a waiver of informed consent and a waiver of HIPAA authorization, are never contacted by the study team, and have had no opportunity to agree to secondary distribution of their data. The approved waivers authorize use of these data for the research described in this record, not redistribution.

The dataset also carries elevated re-identification risk. It is drawn from a single institution and built from timestamped electronic health record events, including hospital unit, transfer sequences, medication administration times, and date of death. Within the source institution that combination identifies many participants even after direct identifiers are removed, and coarsening it enough to prevent re-identification would destroy the temporal structure on which the analysis depends.

The protocol and statistical analysis plan will be made available.

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No