SUPPLEMENTAL MATERIALS
Quality Database and Abstraction Methodology:
As part of the underlying quality improvement (QI) initiative, a centralized data collection process began with initial data collection in 2010. The local protocol and three-hour bundle was introduced in 2010 and refined in 2011. An initial run-in phase of analysis was intended to control the initial impacts of Hawthorne effect and denominator expansion as the system adapted to increased awareness and practice change. In January 2012 bundle elements were finalized. Data capture has continued since, but has been expanded at several points to include more rigorous and comprehensive data on sepsis and septic shock patients.
A dedicated team of data abstractors at each site screened all patients with known or suspected infection and 2 SIRS criteria for database capture.All abstractors received standardized training at the beginning of data-collection involvement. Abstractors only recorded information for patients meeting database inclusion criteria: 1) a suspected or confirmed infection, 2) ≥2 SIRS criteria, and 3) lactate ≥2.2mmol/L or hypotension (systolic blood pressure <90mmHg or mean arterial pressure <65mmHg) or ≥1 sepsis organ-dysfunction criteria (outlined below). Relevant data were abstracted into the QI database using a standardized data collection form, which abstractors submitted to a centralized data collection unit. Abstractors excluded patients: < 18 years; with advance directives precluding bundle interventions; who declined interventions; who were admitted from the ED directly to palliative care or hospice; or who were enrolled in an IRB-approved clinical trial that precluded standard application of the bundle. Interrater reliability was not assessed. Because abstractors did not document when a potential patient was excluded from the QI database, we are unable to determine how many total patients were screened to produce the final cohort in this investigation.
Demographic and clinical data obtained included patient age, sex, primary payer, initial lactate level, signs of hypoperfusion or organ dysfunction, and comorbidities at baseline, as well as treatment and laboratory data-points. Abstractors documented intravenous fluid resuscitation initiation time as the time that a 0.9% normal saline bolus of at least 30mL/kg was hung – i.e. the time administration began, not the order time. In the case of patients receiving multiple boluses, the time of the first bolus was used. Fluid completion times were not reliably documented and subsequently not recorded. A bolus was defined as volume of at least 500mL, ordered to be administered at a rate of at least 500mL per 15 minutes. Patients presenting with comorbid congestive heart failure and/or chronic renal dysfunction were NOT excluded from the fluid resuscitation requirement of the bundle, and the intention was to administer fully bundle compliant care. The only allowed physiologic exceptions to the administration of the 30mL/kg bolus within 30 minutes were 1) “acute decompensated heart failurewith use of a left ventricular assist device” or 2) “dialysis dependent with overt evidence of volume overload”, at time-zero. Antibiotic administration was the time at which the first antibiotic for the sepsis episode was administered – not the order time. Blood culture draw times were documented as the time of collection. Lactate result times were the time the laboratory result became available for review, not the time a provider actually viewed the result.
Database managers monitored database quality monthly. Quality control was applied by monthly gap-analyses between QI database records and a database of all hospital discharges administered by New York State. Records for patients with a discharge diagnosis of sepsis who were missing from the database were reviewed to determine if database inclusion criteria were met, and subsequently entered if appropriate. Patients missing from the database who had a discharge diagnosis of sepsis but who were determined to not to have met study inclusion criteria upon medical record review were not entered into the database. Patients who were “missed” by the treating clinician would have had their initial sepsis episode entered as the first time all objective inclusion criteria were met based on manual review of the medical record. Outcomes data (i.e., mortality, ICU admission and length of stay, etc…) were validated against the health system’s accounting/billing database.
Supplemental Table 1. Organ Dysfunction Criteria for Study InclusionOrgan Dysfunction Criteria a / Definition
1. Hyperlactermia / Serum lactate ≥2.2 mmol/L
2. Hypotension / Systolic blood pressure < 90 mmHg or a mean arterial pressure < 65 mmHg.
3. Acute Kidney Injury (AKI) / Serum creatinine >2.0 mg/dL or 50% increase from known baseline in the absence of chronic kidney disease
4. Thrombocytopenia / Platelet count < 150,000 cells/μm3
5. Coagulopathy / International normalized ratio (INR) >1.5, activated partial thromboplastin time >30 seconds, or partial thromboplastin time >60 seconds, not otherwise explained by medical history
6. Elevated Bilirubin / Serum bilirubin > 2.0 mg/dL in the absence of pre-existing liver failure
7. Acute Altered Mental Status (AMS) / New altered mentation unrelated to the patient’s prior medical history
8. Altered Gas Exchange / New increased O2 requirement to maintain SaO2 > 90% or a PaO2/FiO2 ratio < 300
7. ≥2 “Super-SIRS” Criteria at Triage / Locally developed consensus-criteria, where meeting ≥2 criteria at triage was a “time-zero” entry point for 3-hour bundle care. “Super-SIRS” criteria were:
1)Heart rate greater ≥ 120
2)Respiratory rate ≥ 24
3)Systolic blood pressure < 90 mmHg or mean arterial pressure < 65 mmHg
4)Temperature ≥ 38.0° C (101.0° F) or ≤ 36.0° C (96.8° F)
5)Acutely altered mental status
a All patients included in this study had a source of infection, met at least 2 SIRS criteria, and met at least one of the above organ dysfunction criteria. All organ dysfunction criteria was new, acute, and not explainable by the patients past medical history alone.
Supplemental Table 2 – Individual Site Contributions to Study Cohort
Variable / Entire Cohort / ≤ 30 Minutes / 31-120 Minutes / >120 Minutes or No Fluids in First 6 Hours
N / 11,182 / 5,336 / 2,388 / 3,458
Tertiary Hospitals / 8,069 (72.2%) / 3,740 (70.1%) / 1,735 (72.7%) / 2,594 (75%)
Hospital A / 2,349 (21.0%) / 1,086 (20.4%) / 500 (20.9%) / 763 (22.1%)
Hospital B / 1,410 (12.6%) / 587 (11.0%) / 369 (15.5%) / 454 (13.1%)
Hospital C / 1,450 (13.0%) / 777 (14.6%) / 391 (16.4%) / 282 (8.2%)
Hospital D / 2,860 (25.6%) / 1,290 (24.2%) / 475 (19.9%) / 1,095 (31.7%)
Hospital E / 830 (7.4%) / 389 (7.3%) / 186 (7.8%) / 255 (7.4%)
Community Hospitals / 3,113 (27.8%) / 1,596 (29.9%) / 653 (27.3%) / 864 (25.0%)
Hospital F / 1,038 (9.3%) / 458 (8.6%) / 261 (10.9%) / 319 (9.2%)
Hospital G / 255 (2.0%) / 168 (3.1%) / 32 (1.3%) / 25 (0.7%)
Hospital H / 790 (7.1%) / 438 (8.2%) / 142 (5.9%) / 210 (6.1%)
Hospital I / 230 (2.1%) / 143 (2.7%) / 32 (1.3%) / 55 (1.6%)
Study hospitals see an aggregate of 475,000 Emergency Department visits annually. All tertiary hospitals were academic medical centers. There were no “hybrid” centers. Hospital A is a Level 1 trauma center. Hospitals B, C, D, and F are urban hospitals within the 5-boroughs of New York City. Hospitals A, E, G, H, and I are sub-urban hospitals in Nassau and Suffolk Counties of Long Island, NY (classified as urban, i.e., not rural, by the Centers for Medicare and Medicaid Services).
Supplemental Table 3 – Distribution of Patient Characteristics Across Fluid Volume Groups
Variable / All Patients / < 5 mL/kg / 5-19 mL/kg / 20-35 mL/kg / > 35 mL/kg
N / 11,182 / 2,169 / 3,224 / 3,254 / 2,535
Male sex / 5,740 (51.3%) / 1,150 (53.0%) / 1,750 (54.3%) / 1,710 (52.6%) / 1,130 (44.6%)
Age – median (IQR) / 74 (62, 85) / 76 (65, 85) / 75 (63, 85) / 74 (61,84) / 74 (60,85)
BMI – median (IQR) / 26.0 (22.3,30.5) / 27.0 (23.0, 32.0) / 27.3 (23.5, 32.7) / 26.6 (23.2, 30.6) / 23.0 (19.8, 26.5)
Tertiary Care Hospital / 8,069 (72.2%) / 1,720 (79.3%) / 2,423 (75.2%) / 2,200 (67.6%) / 1,726 (68.1%)
Initial Sepsis Presentation in the ED / 8,673 (77.6%) / 1,117 (51.5%) / 2,501 (77.6%) / 2,810 (86.4%) / 2,245 (88.6%)
Comorbidity Burden a
Congestive Heart Failure / 1,647 (14.7%) / 575 (26.5%) / 510 (15.8%) / 341 (10.5%) / 221 (8.7%)
Chronic Renal Failure / 1,161 (10.4%) / 415 (19.1%) / 332 (10.3%) / 275 (8.5%) / 139 (5.5%)
Chronic Obstructive Pulmonary Disease / 793 (7.1%) / 200 (9.2%) / 240 (7.4%) / 188 (5.8%) / 165 (6.5%)
Diabetes / 3,631 (32.5%) / 803 (37.0%) / 1,127 (35.0%) / 1,031 (31.7%) / 670 (26.4%)
Immune Modifying Medications / 739 (6.6%) / 136 (6.3%) / 225 (7.0%) / 222 (6.8%) / 156 (6.2%)
Liver Failure / 173 (1.5%) / 49 (2.3%) / 47 (1.5%) / 46 (1.4%) / 31 (1.2%)
Metastatic Disease / 908 (8.1%) / 187 (8.6%) / 260 (8.1%) / 245 (7.5%) / 216 (8.5%)
Leukemia/Lymphoma/Multiple Myeloma / 431 (3.9%) / 95 (4.4%) / 123 (3.8%) / 131 (4.0%) / 82 (3.2%)
Organ Transplant / 120 (1.1%) / 29 (1.3%) / 30 (0.9%) / 41 (1.3%) / 20 (0.8%)
HIV/AIDS / 109 (1.0%) / 18 (0.8%) / 24 (0.7%) / 28 (0.9%) / 39 (1.5%)
Presentation and Acuity
Primary Source of Infection
Respiratory / 4,460 (39.9%) / 1,061 (48.9%) / 1,241 (38.5%) / 1,176 (36.1%) / 982 (38.7%)
Urinary / 2,802 (25.1%) / 367 (16.9%) / 800 (24.8%) / 926 (28.5%) / 709 (28.0%)
Gastrointestinal / 1,071 (9.6%) / 168 (7.7%) / 296 (9.2%) / 334 (10.3%) / 273 (10.8%)
Skin / 778 (7.0%) / 147 (6.8%) / 284 (8.8%) / 230 (7.1%) / 117 (4.6%)
Other / 2,071 (18.5%) / 426 (19.6%) / 603 (18.7%) / 588 (18.1%) / 454 (17.9%)
Nosocomial Infection / 1,213 (10.9%) / 381 (17.6%) / 368 (11.4%) / 275 (8.5%) / 189 (7.5%)
Positive Chest Radiography at Presentation / 4,818 (43.1%) / 1,073 (49.5%) / 1,363 (42.3%) / 1,260 (38.7%) / 1,122 (44.3%)
Systolic BP < 90mmHg or 40% Decrease from Baseline or MAP < 65mmHg / 3,714 (33.2%) / 593 (27.3%) / 868 (26.9%) / 1,057 (32.5%) / 1,196 (47.2%)
Initial Serum Lactate Level – mean (SD) / 3.2 mmol/L (2.4) / 2.8 mmol/L (2.3) / 2.9 mmol/L (2.0) / 3.3 mmol/L (2.4) / 3.7 mmol/L (2.8)
Initial Lactate ≥ 2mmol/L / 7,136 (63.8%) / 1,092 (50.3%) / 2,024 (62.8%) / 2,237 (68.7%) / 1,783 (70.3%)
Initial Lactate ≥ 4mmol/L / 6,687 (59.8%) / 325 (15.0%) / 539 (16.7%) / 784 (24.1%) / 802 (31.6%)
Septic Shock / 1,701 (15.2%) / 346 (16.0%) / 449 (13.9%) / 428 (13.2%) / 478 (18.9%)
Altered Mental Status / 2,675 (23.9%) / 537 (24.8%) / 683 (21.2%) / 741 (22.8%) / 714 (28.2%)
Altered Gas Exchange b / 2,412 (21.6%) / 659 (30.4%) / 634 (19.7%) / 619 (19.0%) / 500 (19.7%)
Bilirubin > 2.0 mg/dL / 621 (5.6%) / 108 (5.0%) / 161 (5.0%) / 204 (6.3%) / 148 (5.8%)
Acute Kidney Injury c / 2,328 (20.8%) / 417 (19.2%) / 663 (20.6%) / 672 (20.7%) / 576 (22.7%)
Coagulopathy d / 1,730 (15.5%) / 390 (18.0%) / 482 (15.0%) / 452 (13.9%) / 406 (16.0%)
Immunocompromised at Presentation / 2,346 (21.0%) / 462 (21.3%) / 668 (20.7%) / 671 (20.6%) / 545 (21.5%)
Interventions e, f
Crystalloid Volume in First 6 hrs – mean (SD) / 1.6 L (1.2) / 0 L (0.1) / 1.0 L (0.4) / 2.1 L (0.6) / 3.1 L (1.0)
Time to Crystalloid Initiation – mean (SD) / 105 minutes (160) / 331 minutes (94) / 66 minutes (128) / 43 minutes (118) / 38 minutes (119)
Fluids Initiated in ≤ 30 minutes / 5,336 (47.7%) / 101 (4.7%) / 1,612 (50.0%) / 1,990 (61.2%) / 1,633 (64.4%)
Fluids Initiated in 31-120 minutes / 2,338 (21.4%) / 59 (2.7%) / 923 (28.6%) / 819 (25.2%) / 587 (23.2%)
Fluids Not Initiated within 120 minutes / 3458 (30.9%) / 2009 (92.6%) / 689 (21.4%) / 445 (13.7%) / 315 (12.4%)
Lactate Order to Result Time – median (IQR) / 38 minutes (21,63) / 32 minutes (13,59) / 37 minutes (20,61) / 40 minutes (24,62) / 41 minutes (24,67)
Blood Cultures Drawn Before Antibiotics / 7,530 (67.3%) / 1,274 (58.7%) / 2,068 (64.1%) / 2,331 (71.6%) / 1,857 (73.3%)
Time to Antibiotics – median (IQR) / 53minutes(10,108) / 58 minutes (4,130) / 58 minutes (7,118) / 52 minutes (13,102) / 47 minutes (13,90)
Antibiotics ≤60 min (time-zero) and ≤180 min(2 SIRS) / 6,416 (57.4%) / 1,262 (58.2%) / 1,790 (55.5%) / 1,841 (56.6%) / 1,523 (60.1%)
Antibiotics ≤180 min (time-zero) / 9,040 (90.5%) / 1,488 (85.9%) / 2,504 (87.9%) / 2,504 (92.4%) / 2,249 (94.5%)
Repeat Lactate Available Within 24 Hours / 5,963 (53.3%) / 942 (43.4%) / 1,575 (48.9%) / 1,867 (57.4%) / 1,579 (62.3%)
a Comorbidities reflect status at time-zero, and would not reflect conditions that developed subsequently during hospital stay.
bAltered Gas Exchange defined as PaO2 /FiO2 < 300 or an increased O2 requirement to maintain SaO2 >90%.
c Acute Kidney Injury defined as creatinine >2.0 or 50% increase from a known baseline.
d Coagulopathy defined as platelet count < 150,000 cells/µm3, international normalized ratio > 1.5, or partial thromboplastin time > 60 seconds.
eExcept where otherwise specified, all times are in minutes, and reflect the time elapsed from time-zero unless otherwise indicated. A negative time indicates an intervention performed before time-zero. E.g., a patient whose fluid resuscitation began 20 minutes before laboratory results indicating organ dysfunction became available have a fluid initiation time of -20.
f All crystalloid was 0.9% normal saline solution.