- Email: [email protected]
In-Hospital Sepsis Mortality Rates Comparing Tertiary and Non-Tertiary Hospitals in Washington State
The Journal of Emergency Medicine, Vol. -, No. -, pp. 1–8, 2018 Ó 2018 Elsevier Inc. All rights reserved. 0736-4679/$ - see front matter
https://doi.org/10.1016/j.jemermed.2018.01.027
Original Contributions
IN-HOSPITAL SEPSIS MORTALITY RATES COMPARING TERTIARY AND NON-TERTIARY HOSPITALS IN WASHINGTON STATE Gail G. Salvatierra, RN, PHD,* Bernice G. Gulek, ACNP-BC, MS, RN,† Baran Erdik, MD, MHPA,† Deborah Bennett, RN, PHD,* and Kenn B. Daratha, PHD†‡§ *School of Nursing, California State University San Marcos, San Marcos, California, †College of Nursing, Washington State University, Spokane, Washington, ‡Providence Medical Research Center, Providence Sacred Heart Medical Center, Spokane, Washington, and §Department of Medical Education and Biomedical Informatics, University of Washington, Spokane and Seattle, Washington Reprint Address: Gail G. Salvatierra, RN, PHD, School of Nursing, California State University San Marcos, 333 S. Twin Oaks Valley Rd., San Marcos, CA 92096
, Keywords—in-hospital mortality; sepsis; severe sepsis; SIRS; tertiary hospitals; non-tertiary hospitals
, Abstract—Background: More than a million people a year in the United States experience sepsis or sepsisrelated complications, and sepsis remains the leading cause of in-hospital deaths. Unlike many other leading causes of inhospital mortality, sepsis detection and treatment are not dependent on the presence of any technology or services that differ between tertiary and non-tertiary hospitals. Objective: To compare sepsis mortality rates between tertiary and non-tertiary hospitals in Washington State. Methods: A retrospective longitudinal, observational cohort study of 73 Washington State hospitals for 2010–2015 using data from a standardized state database of hospital abstracts. Abstract records on adult patients (n = 86,378) admitted through the emergency department (ED) from 2010 through 2015 in all tertiary (n = 7) and non-tertiary (n = 66) hospitals in Washington State. Results: The overall mortality rate for all hospitals was 6.5%. In the fully adjusted model, the odds ratio for in-hospital death was higher in non-tertiary hospitals compared with tertiary hospitals (odds ratio 1.25; 95% confidence interval 1.17–1.35; p < 0.001). Conclusions: We observed higher sepsis mortality rates in non-tertiary hospitals, compared with tertiary hospitals. Because most patients who are treated for sepsis are treated outside of tertiary hospitals, and the number of patients treated for sepsis in non-tertiary hospitals seems to be rising, a better understanding of the cause or causes for this differential is crucial. Ó 2018 Elsevier Inc. All rights reserved.
INTRODUCTION Sepsis is ‘‘a life-threatening organ dysfunction due to a dysregulated host response to infection’’ (1). More than a million people a year in the United States experience sepsis (or sepsis-related complications), which is the leading cause of in-hospital deaths, and treatment costs are estimated to exceed $24 billion annually (2–7). Improving the quality of care for sepsis has gained widespread attention for more than a decade. In 2004, the Surviving Sepsis Campaign (SSC) published the first bundled guidelines, which provided a framework for widespread standardization of early treatment and management of sepsis (8,9). Since the first guidelines were released, studies have consistently reported increases in sepsis awareness and incidence (10). Studies have also reported a decrease in mortality and length of stay, despite the increased incidence of sepsis (8,9,11–13). Most sepsis studies, however, have been conducted in larger urban centers and teaching hospitals (14,15). Whether patients with sepsis or sepsis-related diagnoses treated in tertiary and non-tertiary hospitals are
RECEIVED: 21 October 2017; FINAL SUBMISSION RECEIVED: 6 January 2018; ACCEPTED: 19 January 2018 1
2
G. G. Salvatierra et al.
achieving comparable benefits is uncertain. Much of the research to date that includes sepsis comparisons, such as that of Villapiano et al. (2017), has made a rural/urban distinction that is not fully comparable with tertiary/nontertiary (16). Several studies reporting mortality rates between non-tertiary and tertiary hospitals found higher mortality rates for non-tertiary hospitals for conditions with specific treatment guidelines, such as acute myocardial infarction and stroke (17–19). Sepsis is unique among the leading causes of inhospital mortality because its diagnosis and treatment, as defined by the SSC, requires neither technology nor other resources that are differentially distributed between tertiary (mostly urban) and non-tertiary (mostly rural) hospitals (9). The diagnosis of sepsis, per guidelines, is by careful clinical observation and common laboratory tests. Treatment is through standard antibiotics and supportive therapies. The purpose of our study was to compare in-hospital mortality among multiple geographically contiguous tertiary and non-tertiary care hospitals in the state of Washington. MATERIALS AND METHODS Our retrospective longitudinal cohort study included 86,378 adult patients aged 18 years or older who were hospitalized with a diagnosis of sepsis between January 2010 and September 2015 in civilian, acute care hospitals in Washington State. The study period terminated with the changeover from International Classification of Diseases, 9th Revision (ICD-9) to ICD-10 coding in October 2015. Our project was determined to be exempt from review by the institutional review board at Washington State University (institutional review board exempt application #15607). We utilized de-identified patient data from the Washington State Comprehensive Hospital Abstract Reporting System (CHARS). The CHARS database contains abstracted information on all discharges from civilian hospitals in Washington State. The CHARS database includes data from all payers for hospitalization. The CHARS database does not include information on hospitalizations from federal, Veterans Affairs, or military institutions. Each of the selected hospitals was categorized as a tertiary care or a non-tertiary hospital based on services provided. Tertiary care institutions provide 24-h emergency services and access to all medical and surgical specialty services (20).
through September 30, 2015. Study participants included patients with an ICD-9-Clinical Modification (CM) primary diagnosis of septicemia (038*) or systemic inflammatory response syndrome (995.9*). An ICD-9-CM secondary diagnosis further classified systemic inflammatory response syndrome as not otherwise specified (995.90), sepsis due to infectious process without acute organ dysfunction (995.91), severe sepsis with acute organ dysfunction (995.92), systemic inflammatory response syndrome due to noninfectious process without acute organ dysfunction (995.93), and systemic inflammatory response syndrome due to noninfectious process with acute organ dysfunction (995.94). Study participants were identified as admitted through the ED by the presence of a revenue code of 450–459. To control for varying severity of illness at entry into the study, we calculated the number of previous hospitalizations within the past 12 months for each participant included in the study. We identified comorbidity variables based on the hospitalization and any other previous hospitalizations (past 12 months) using a set of 30 comprehensive comorbidity definitions employing the Elixhauser method, which allows for the inclusion of 7000 diagnostic codes (ICD-9-CM) that have been categorized into broader diagnosis groupings and assigned to each case according to conditions at hospitalization that worsen patient outcomes (21). Statistical Analysis We compared rates of in-hospital mortality in tertiary care hospitals to rates of mortality in non-tertiary care hospitals, and calculated a relative risk for mortality. To examine the risk for mortality between the two cohorts, we used binary logistic regression to control for comorbidities and severity of illness. We calculated the odds ratio (OR) and confidence interval (95% CI) for adjusted and unadjusted hospital mortality for each of the hospitals. After performing unadjusted analyses for both tertiary and non-tertiary centers, we adjusted for potential confounders in the following order: index in-hospital mortality, age, sex, intensive care unit (ICU) admissions, transfers, and comorbidity. Tests were two-tailed, and statistical significance set to 0.05. SPSS software package (Version 20.1; IBM, Armonk, NY) was used for all analyses. The internal validation of main study findings was completed using split-file validation in which the analysis was repeated after cases were randomly assigned to derivation and validation cohorts.
Inclusion Criteria and Hospitalization Characteristics RESULTS We included in the analysis adult patients (>18 years of age) admitted through the emergency department (ED) and hospitalized in Washington from January 1, 2010
Among 86,378 adult patients who were hospitalized with sepsis in Washington State hospitals from 2010 through
Sepsis Mortality in Tertiary vs. Non-Tertiary Hospitals
3
September 2015, there were demographic and clinical differences (Table 1). Care for 76% of the sample was provided in 66 non-tertiary hospitals, compared with patients admitted to seven tertiary hospitals (24% of the sample). Patients who were admitted to tertiary care hospitals were younger, on average, by almost 5 years. The proportion of patients covered by Medicare insurance was thus higher in non-tertiary hospitals, compared with tertiary hospitals. In addition, slight statistically significant differences were observed in other demographic characteristics. Transfers from skilled nursing facilities, other hospitals, and admission from outside the hospital referral region were statistically higher in tertiary hospitals compared with non-tertiary hospitals (p < 0.001). Clinical characteristics for patients admitted to both hospital types were similar. No statistical differences were detected in number of comorbidities, proportion of patients with diabetes, length of stay, or classification of sepsis. Although rates of in-hospital deaths have declined for both non-tertiary and tertiary hospitals (Figure 1), abso-
lute counts differed. Comparing the years 2010–2014, the number of discharges from non-tertiary hospitals for sepsis has nearly doubled (from 7317 patients in 2010 to 14,481 patients in 2014). By contrast, the number of discharges from tertiary hospitals for sepsis has increased by 39% (from 2951 patients in 2010 to 4113 patients in 2014). Within non-tertiary hospitals, the number of deaths has changed little over time, ranging from a low of 758 in-hospital deaths in 2010 to 889 in-hospital deaths in 2012. Comparing 2010 with 2014 in-hospital deaths, the absolute number has declined 2%, with the biggest reduction occurring in 2014, where the number of discharges were the highest (14,481) and the number of deaths the lowest (758). In contrast, the number of deaths within tertiary hospitals has declined each full year of the study, from 232 in-hospital deaths in 2010 to 134 inhospital deaths in 2014. Comparing 2010 with 2014 inhospital deaths, the absolute number has declined 42%. A multivariable model that controlled for differences in patient populations served (Table 2) confirmed that there was increased risk for in-hospital death in non-
Table 1. Characteristics of Adults Hospitalized for Sepsis by Type of Hospital (N = 86,378) Non-Tertiary 66 Hospitals (n = 66,021)
Patient Characteristics
Patient age (years)
p-Value
Median
SD
Median
SD
65.2
19.0
60.3
19.5
Median
IQR
Median
IQR
2 4
1–2 3–6
2 4
1–2 3–6
Number of comorbidities Length of stay (days)
Females Insurance payer Medicare Medicaid Commercial Self-pay and charitable Dual enrolled Medicare and Medicaid Transfers from SNF Transfers from other hospital ICU admission Admissions outside HRR Comorbidities Heart failure Hypertension w/o complications Hypertension w/complications Chronic pulmonary disease Diabetes w/o complications Diabetes w/complications Kidney failure Drug abuse Sepsis disease classification 995.91 995.92 995.90 or 995.93 or 995.94
Tertiary 7 Hospitals (n = 20,357)
<0.001
0.86 0.39
n
%
n
%
34,965
53.0
10,381
51.0
<0.001
28,914 7897 16,041 3062 10,107 1737 778 21,833 4410
43.8 12.0 24.3 4.6 15.3 2.6 1.2 33.1 6.7
6077 3656 5977 1402 3245 826 714 6982 1606
29.9 18.0 29.4 6.9 15.9 4.1 3.5 34.3 7.9
<0.001
<0.001 <0.001 0.001 <0.001
8824 12,648 5004 12,204 9118 3028 6396 4062
13.4 19.2 7.6 18.5 13.8 4.6 9.7 6.2
2305 3456 1321 3918 2794 981 1876 1794
11.3 17.0 6.5 19.2 13.7 4.8 9.2 8.8
<0.001 <0.001 < 0.001 0.02 0.76 0.17 0.05 <0.001
45,753 20,053 215
69.3 30.4 0.3
14,009 6264 84
68.8 30.8 0.4
0.09
SNF = skilled nursing facilities; ICU = intensive care unit; HRR = hospital referral region as defined by the Dartmouth Atlas.
4
G. G. Salvatierra et al.
Figure 1. In-hospital mortality rate of adults hospitalized for sepsis by hospital type (N = 86,378). ICD10 = International Classification of Diseases, 10th Revision.
tertiary hospitals compared with tertiary hospitals. The overall mortality rate for all hospitals was 6.5% (5645 in-hospital deaths divided by 86,378 discharges). The unadjusted OR for in-hospital death was higher in nontertiary hospitals compared with tertiary hospitals (OR 1.37; 95% CI 1.28–1.47; p < 0.001). The fully adjusted
1.37 1.28–1.47 <0.001 1.43 1.34–1.53 <0.001 1.28 1.19–1.37 <0.001
model, controlling for year of discharge, age, sex, insurance payer, admission to the ICU, transfers from skilled nursing facilities, transfers from other hospitals, admissions outside the hospital referral region, and number of comorbidities yielded comparable results (OR 1.25; 95% CI 1.17–1.35; p < 0.001). Internal validation, comparing the derivation and validation cohorts, yielded equivalent results. Based on similarity of in-hospital mortality rates for the partial year in 2015 for non-tertiary hospitals (3.9%) and tertiary hospitals (4.4%), we derived the OR for a fully adjusted model. The derivation and validation cohorts were not consistent with the full study results (Table 3).
1.25 1.17–1.35 <0.001
DISCUSSION
Table 2. Unadjusted and Fully Adjusted Odds Ratio of InHospital Mortality of Adults Hospitalized for SIRS 2010–2015 (N = 86,378) Model Type Unadjusted model Controlling for year Controlling for year, age, sex, payer, ICU admission, transfers Controlling for year, age, sex, payer, ICU admission, transfers, admission outside hospital referral region, comorbidities
OR
95% CI
p-Value
Reference group = discharges from tertiary hospitals. SIRS = systemic inflammatory response syndrome; OR = odds ratio; CI = confidence interval; ICU = intensive care unit.
In a study of 86,378 patients hospitalized for sepsis in the State of Washington between 2010 and 2015, rates of mortality differed between tertiary and non-tertiary hospitals. Higher rates of mortality among patients in
Sepsis Mortality in Tertiary vs. Non-Tertiary Hospitals
5
Table 3. Unadjusted and Fully Adjusted Odds Ratio of In-Hospital Mortality of Adults Hospitalized for SIRS 2015 (N = 15,205) Derivation Cohort
Validation Cohort
Model Type
OR
95% CI
p-Value
OR
95% CI
p-Value
Unadjusted model Controlling for year, age, sex, payer, ICU admission, transfers Controlling for year, age, sex, payer, ICU admission, transfers, admission outside hospital referral region, comorbidities
0.66 0.68
0.60–0.99 0.52–0.88
0.04 0.004
1.02 0.82
0.77–1.35 0.62–1.10
0.88 0.20
0.68
0.52–0.89
0.004
0.81
0.60–1.08
0.16
Reference group = discharges from tertiary hospitals. SIRS = systemic inflammatory response syndrome; OR = odds ratio; CI = confidence interval; ICU = intensive care unit.
non-tertiary hospitals persisted in multivariable models controlling for demographic and clinical characteristics and in internally validated analyses. One of the major goals of national treatment guidelines is to identify and foster widespread implementation of evidence-based best practices to reduce mortality and morbidity from conditions that affect large numbers of patients. Despite the overall success of such highvisibility efforts as the Guidelines for the Early Management of Ischemic Stroke, and the Guideline for the Management of Heart Failure, there has been a persistent discrepancy in mortality outcomes between tertiary and non-tertiary hospitals (19,22). This has been true across a broad range of conditions, including those with and without national campaigns to encourage uniform treatment protocols (16). Our results, finding a similar tertiary/non-tertiary discrepancy for sepsis, show that the discrepancy persists even in the absence of an obvious resource-disparity explanation. Successful adherence to national guidelines for cardiovascular events requires prompt access to sophisticated diagnostic equipment and an on-call team capable of revascularization within stated time limits. These resources are not found at non-tertiary hospitals. For sepsis, however, the mortality gap cannot be accounted for with a similar explanation. Unlike stroke or myocardial infarction, which can be reliably diagnosed using a computed tomography scan, biomarkers, and electrocardiogram findings, sepsis diagnosis consistent with the SSC guidelines involves a cluster of clinical findings that can vary individual-to-individual or based on the context of clinical presentation (10,23). Whereas diagnoses requiring specialized technology or other scarce resources would be expected to differ by facility type, the SSC guidelines require nothing but clinical skills and laboratory capabilities common to both tertiary and non-tertiary hospitals. Yet based on our findings, a mortality gap exists for sepsis. The reasons for the disparity in sepsis mortality between tertiary and non-tertiary hospitals requires further investigation. Because data were not available about
sepsis bundle implementation, it is not possible for us to determine if there were significant differences between the use of the guidelines in tertiary vs. non-tertiary hospitals, or if the rate of uptake was different. Also, unknown from the available data is whether there are significant differences in nursing or physician awareness training, in-service education, availability of resources, or inhouse review for adherence to the guidelines, any one or a combination of which could account, in whole or part, for the observed discrepancy. The efficacy of sepsis-bundled interventions was established in teaching hospitals often classified as providing tertiary care. Reducing mortality rates and increasing diagnosis of sepsis has most often occurred in environments where sepsis interventions have been implemented and compliance with bundles is high (24,25). Achieving acceptance, implementation, and compliance with the SSC guidelines is a complex task that can be affected by hospital culture, available financial and other resources, institutional commitment, and many other factors that tend to favor tertiary care hospitals. Our results also support previous reports of both increasing incidence of sepsis and declining mortality rates for patients admitted to U.S. hospitals in the treatment of sepsis (26–28). Although we found lower mortality rates in tertiary hospitals, the tertiary/non-tertiary differences narrowed over the study span. In fact, the mortality rates in tertiary and non-tertiary hospitals converged by 2015. Though this was only partial-year data due to the ICD-9/10 changeover, it may point to a slower but eventually complete uptake and implementation of the SSC guidelines, and is another reason that there is urgent need for a prospective, large-scale study capable of elucidating the reasons for any confirmed discrepancy. Our finding of declining mortality in the treatment of hospitalized patients with sepsis confirms previous reports (29–31). During our study, the absolute numbers of patients dying from sepsis declined sharply in tertiary hospitals, but remained relatively stable in nontertiary hospitals. Substantial reductions in mortality since implementation of the sepsis bundles are a strong
6
G. G. Salvatierra et al.
indication that coordinated, preemptive action, increased level of suspicion, and prompt response combined with therapeutic advances hold the promise of further reductions in the sepsis mortality rate. Attributing reductions in mortality rates to bundled interventions, however, requires a more granular understanding of sepsis causality and mortality and experimental study designs. Limitations Our study findings should be understood within the context of several limitations. First, we did not know the compliance rates for implementing sepsis bundles at any of the hospitals in our study. In addition, the CHARS dataset did not include any information regarding which, if any, sepsis diagnosis criteria were used. We identified sepsis based on ICD-9-CM codes, and in Washington State, ICD-9-CM codes were discontinued starting October 1, 2015. Partial-year data for 2015 may account, in part, for the anomalous mortality results for that year. Hospital abstracts historically fail to capture complete data on patient race and ethnicity. Finally, the use of a retrospective study design does not allow any inferences of causality. We attempted to control for methodological inadequacies and inconsistencies through our study design. We assembled a large set of hospital abstracts over multiple years, from all civilian hospitals in Washington State. Our study design ensured adequate power and generalizability. We controlled for multiple prognostic factors including measures of severity of illness and comorbidities. We also selected hospitals within a single state that reported data in a standardized fashion. Finally, we internally validated our study findings using the splitfile approach. CONCLUSIONS We observed differences in the sepsis mortality rates between tertiary and non-tertiary hospitals for the years 2010–2014. Differences in mortality rates disappeared in 2015, although the exact reason for the convergence of mortality rates is unclear. Sepsis is a severe and urgent health care issue. There is need for a strong prospective study to determine, not only the degree to which there is a sepsis mortality gap, but also the causal factor or factors that create or contribute to such a gap. A prospective study could help inform efforts to achieve optimal treatment and diagnosis of sepsis, and thus reduce avoidable hospital mortality. REFERENCES 1. Singer M, Deutschman C, Seymour C, et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA 2016;315:801–10.
2. De Backer D, Dorman T. Surviving Sepsis Guideline: a continuous move toward better care of patients with sepsis. JAMA 2017;317: 807–8. 3. Angus D, Linde-Zwirble W, Lidicker J, et al. Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med 2001;29:1303–10. 4. Mayr F, Yende S, Angus D. Epidemiology of severe sepsis. Virulence 2014;5(1):4–11. 5. Liu V, Escobar G, Greene J, et al. Hospital deaths in patients with sepsis from 2 independent cohorts. JAMA 2014;312:90–2. 6. Gaieski DF, Edwards JM, Kallan M, Karr BG. Benchmarking the incidence and mortality of severe sepsis in the United States. Crit Care Med 2013;41:1167–74. 7. Lagu T, Rothberg M, Shieh M, et al. Hospitalizations, costs, and outcomes of severe sepsis in the United States 2003 to 2007. Crit Care Med 2012;40:754–61. 8. Dellinger R, Levy M, Rhodes A, et al. Surviving Sepsis Campaign Guidelines Committee including The Pediatric S. Surviving Sepsis Campaign: international guidelines for management of severe sepsis and septic shock, 2012. Intensive Care Med 2013;39:165– 228. 9. Khan P, Divatia J. Severe sepsis bundles. Indian J Crit Care Med 2010;14:8–13. 10. Cohen J, Vincent JL, Adhikari NK, et al. Sepsis: a roadmap for future research. Lancet Infect Dis 2015;15:581–614. 11. Levy M, Rhodes A, Phillips G, et al. Surviving Sepsis Campaign: association between performance metrics and outcomes in a 7.5year study. Crit Care Med 2015;43:3–12. 12. Rhodes A, Phillips G, Beale R, et al. The Surviving Sepsis Campaign bundles and outcome: results from the International Multicentre Prevalence Study on Sepsis (the IMPreSS study). Intensive Care Med 2015;41:1620–8. 13. Miller RR 3rd, Dong L, Nelson NC, et al. Multicenter implementation of a severe sepsis and septic shock treatment bundle. Am J Respir Crit Care Med 2013;188:77–82. 14. Seymour CW, Gesten F, Prescott HC, et al. Time to treatment and mortality during mandated emergency care for sepsis. N Engl J Med 2017;376:2235–44. 15. ProCESS Investigators. A randomized trial of protocol-based care for early septic shock. N Engl J Med 2014;370:1683–93. 16. Villapiano N, Iwashyna TJ, Davis MM. Worsening rural-urban gap in hospital mortality. J Am Board Fam Med 2017;30(6):816–23. 17. Baldwin LM, Chan L, Andrilla CHA, Huff ED, Hart LG. Quality of care for myocardial infarction in rural and urban hospitals. J Rural Health 2010;26:51–7. 18. Leira EC, Hess DC, Torner JC, Adams HP Jr. Rural-urban differences in acute stroke management practices: a modifiable disparity. Arch Neurol 2008;65:887–91. 19. Yancy C, Jessup M, Bozkurt B, et al. 2017 ACC/AHA/HFSA focused update of the 2013 ACCF/AHA Guideline for the Management of Heart Failure: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice guidelines and the Heart Failure Society of America. Circulation 2017;136:e137–61. 20. Johns Hopkins Medicine. Tertiary care definition. Available at: https://www.hopkinsmedicine.org/patient_care/pay_bill/insurance_ footnotes.html. Accessed April 14, 2017. 21. Elixhauser A, Steiner C, Harris D, et al. Comorbidity measures for use with administrative data. Med Care 1998;36:8–27. 22. Jauch EC, Saver JL, Adams HP, et al. Guidelines for the early management of patients with acute ischemic stroke. 2013;44:870–947. 23. Levy MM, et al. 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. Intensive Care Med 2003;29:530–8. 24. Levy M, Dellinger P, Townsend S, et al. The Surviving Sepsis Campaign: results of an international guideline-based performance improvement program targeting severe sepsis. Intensive Care Med 2010;36:222–31. 25. Henriksen D, Laursen C, Jensen T, et al. Incidence rate of community-acquired sepsis among hospitalized acute medical patients—a population-based survey. Crit Care Med 2015;43:13–21.
Sepsis Mortality in Tertiary vs. Non-Tertiary Hospitals 26. Fleischmann C, Scherag A, Adhikari N, et al. Assessment of global incidence and mortality of hospital-treated sepsis. Current estimates and limitations. Am J Respir Crit Care Med 2016;193:259–72. 27. Walkey A, Lagu T, Lindenauer P. Trends in sepsis and infection sources in the United States. A population-based study. Ann Am Thorac Soc 2015;12:216–20. 28. Zimmerman JE, Kramer AA, Knaus WA. Changes in hospital mortality for United States intensive care unit admissions from 1988 to 2012. Crit Care 2013;17:R81.
7 29. Kaukonen K, Bailey M, Suzuki S, et al. Mortality related to severe sepsis and septic shock among critically ill patients in Australia and New Zealand, 2000–2012. JAMA 2014;311:1308–16. 30. Stevenson E, Rubenstein A, Radin G, et al. Two decades of mortality trends among patients with severe sepsis: a comparative metaanalysis. Crit Care Med 2014;42:625–31. 31. Liu V, Escobar G, Greene J, et al. Multicenter implementation of a treatment bundle for patients with sepsis and intermediate lactate values. Am J Respir Crit Care Med 2016;193:1264–70.
8
G. G. Salvatierra et al.
ARTICLE SUMMARY 1. Why is this topic important? Sepsis is a leading cause of hospital deaths and one of several conditions for which there are widely accepted national treatment guidelines. Unlike the other leading causes of in-hospital mortality, diagnosis, and treatment does not rely on resources that should differ between tertiary and non-tertiary hospitals, so any discrepancy in outcomes needs further exploration and explanation. 2. What does this study attempt to show? The study attempts to show that there are significant differences in sepsis mortality between tertiary and nontertiary hospitals for one 5-year period in one state. 3. What are the key findings? Non-tertiary hospitals in Washington State over a 5year span had higher in-hospital mortality due to sepsis than did tertiary hospitals. 4. How is patient care impacted? The majority of patients are seen in non-tertiary hospitals. Unlike some conditions for which guidelines are published, sepsis recognition and treatment does not call for exceptional human or technical resources, so discrepancies in such resources are not an explanation for the sepsis mortality gap. Research should be directed at better understanding the factors that cause non-tertiary hospitals to have a higher rate of sepsis mortality.
https://doi.org/10.1016/j.jemermed.2018.01.027
Original Contributions
IN-HOSPITAL SEPSIS MORTALITY RATES COMPARING TERTIARY AND NON-TERTIARY HOSPITALS IN WASHINGTON STATE Gail G. Salvatierra, RN, PHD,* Bernice G. Gulek, ACNP-BC, MS, RN,† Baran Erdik, MD, MHPA,† Deborah Bennett, RN, PHD,* and Kenn B. Daratha, PHD†‡§ *School of Nursing, California State University San Marcos, San Marcos, California, †College of Nursing, Washington State University, Spokane, Washington, ‡Providence Medical Research Center, Providence Sacred Heart Medical Center, Spokane, Washington, and §Department of Medical Education and Biomedical Informatics, University of Washington, Spokane and Seattle, Washington Reprint Address: Gail G. Salvatierra, RN, PHD, School of Nursing, California State University San Marcos, 333 S. Twin Oaks Valley Rd., San Marcos, CA 92096
, Keywords—in-hospital mortality; sepsis; severe sepsis; SIRS; tertiary hospitals; non-tertiary hospitals
, Abstract—Background: More than a million people a year in the United States experience sepsis or sepsisrelated complications, and sepsis remains the leading cause of in-hospital deaths. Unlike many other leading causes of inhospital mortality, sepsis detection and treatment are not dependent on the presence of any technology or services that differ between tertiary and non-tertiary hospitals. Objective: To compare sepsis mortality rates between tertiary and non-tertiary hospitals in Washington State. Methods: A retrospective longitudinal, observational cohort study of 73 Washington State hospitals for 2010–2015 using data from a standardized state database of hospital abstracts. Abstract records on adult patients (n = 86,378) admitted through the emergency department (ED) from 2010 through 2015 in all tertiary (n = 7) and non-tertiary (n = 66) hospitals in Washington State. Results: The overall mortality rate for all hospitals was 6.5%. In the fully adjusted model, the odds ratio for in-hospital death was higher in non-tertiary hospitals compared with tertiary hospitals (odds ratio 1.25; 95% confidence interval 1.17–1.35; p < 0.001). Conclusions: We observed higher sepsis mortality rates in non-tertiary hospitals, compared with tertiary hospitals. Because most patients who are treated for sepsis are treated outside of tertiary hospitals, and the number of patients treated for sepsis in non-tertiary hospitals seems to be rising, a better understanding of the cause or causes for this differential is crucial. Ó 2018 Elsevier Inc. All rights reserved.
INTRODUCTION Sepsis is ‘‘a life-threatening organ dysfunction due to a dysregulated host response to infection’’ (1). More than a million people a year in the United States experience sepsis (or sepsis-related complications), which is the leading cause of in-hospital deaths, and treatment costs are estimated to exceed $24 billion annually (2–7). Improving the quality of care for sepsis has gained widespread attention for more than a decade. In 2004, the Surviving Sepsis Campaign (SSC) published the first bundled guidelines, which provided a framework for widespread standardization of early treatment and management of sepsis (8,9). Since the first guidelines were released, studies have consistently reported increases in sepsis awareness and incidence (10). Studies have also reported a decrease in mortality and length of stay, despite the increased incidence of sepsis (8,9,11–13). Most sepsis studies, however, have been conducted in larger urban centers and teaching hospitals (14,15). Whether patients with sepsis or sepsis-related diagnoses treated in tertiary and non-tertiary hospitals are
RECEIVED: 21 October 2017; FINAL SUBMISSION RECEIVED: 6 January 2018; ACCEPTED: 19 January 2018 1
2
G. G. Salvatierra et al.
achieving comparable benefits is uncertain. Much of the research to date that includes sepsis comparisons, such as that of Villapiano et al. (2017), has made a rural/urban distinction that is not fully comparable with tertiary/nontertiary (16). Several studies reporting mortality rates between non-tertiary and tertiary hospitals found higher mortality rates for non-tertiary hospitals for conditions with specific treatment guidelines, such as acute myocardial infarction and stroke (17–19). Sepsis is unique among the leading causes of inhospital mortality because its diagnosis and treatment, as defined by the SSC, requires neither technology nor other resources that are differentially distributed between tertiary (mostly urban) and non-tertiary (mostly rural) hospitals (9). The diagnosis of sepsis, per guidelines, is by careful clinical observation and common laboratory tests. Treatment is through standard antibiotics and supportive therapies. The purpose of our study was to compare in-hospital mortality among multiple geographically contiguous tertiary and non-tertiary care hospitals in the state of Washington. MATERIALS AND METHODS Our retrospective longitudinal cohort study included 86,378 adult patients aged 18 years or older who were hospitalized with a diagnosis of sepsis between January 2010 and September 2015 in civilian, acute care hospitals in Washington State. The study period terminated with the changeover from International Classification of Diseases, 9th Revision (ICD-9) to ICD-10 coding in October 2015. Our project was determined to be exempt from review by the institutional review board at Washington State University (institutional review board exempt application #15607). We utilized de-identified patient data from the Washington State Comprehensive Hospital Abstract Reporting System (CHARS). The CHARS database contains abstracted information on all discharges from civilian hospitals in Washington State. The CHARS database includes data from all payers for hospitalization. The CHARS database does not include information on hospitalizations from federal, Veterans Affairs, or military institutions. Each of the selected hospitals was categorized as a tertiary care or a non-tertiary hospital based on services provided. Tertiary care institutions provide 24-h emergency services and access to all medical and surgical specialty services (20).
through September 30, 2015. Study participants included patients with an ICD-9-Clinical Modification (CM) primary diagnosis of septicemia (038*) or systemic inflammatory response syndrome (995.9*). An ICD-9-CM secondary diagnosis further classified systemic inflammatory response syndrome as not otherwise specified (995.90), sepsis due to infectious process without acute organ dysfunction (995.91), severe sepsis with acute organ dysfunction (995.92), systemic inflammatory response syndrome due to noninfectious process without acute organ dysfunction (995.93), and systemic inflammatory response syndrome due to noninfectious process with acute organ dysfunction (995.94). Study participants were identified as admitted through the ED by the presence of a revenue code of 450–459. To control for varying severity of illness at entry into the study, we calculated the number of previous hospitalizations within the past 12 months for each participant included in the study. We identified comorbidity variables based on the hospitalization and any other previous hospitalizations (past 12 months) using a set of 30 comprehensive comorbidity definitions employing the Elixhauser method, which allows for the inclusion of 7000 diagnostic codes (ICD-9-CM) that have been categorized into broader diagnosis groupings and assigned to each case according to conditions at hospitalization that worsen patient outcomes (21). Statistical Analysis We compared rates of in-hospital mortality in tertiary care hospitals to rates of mortality in non-tertiary care hospitals, and calculated a relative risk for mortality. To examine the risk for mortality between the two cohorts, we used binary logistic regression to control for comorbidities and severity of illness. We calculated the odds ratio (OR) and confidence interval (95% CI) for adjusted and unadjusted hospital mortality for each of the hospitals. After performing unadjusted analyses for both tertiary and non-tertiary centers, we adjusted for potential confounders in the following order: index in-hospital mortality, age, sex, intensive care unit (ICU) admissions, transfers, and comorbidity. Tests were two-tailed, and statistical significance set to 0.05. SPSS software package (Version 20.1; IBM, Armonk, NY) was used for all analyses. The internal validation of main study findings was completed using split-file validation in which the analysis was repeated after cases were randomly assigned to derivation and validation cohorts.
Inclusion Criteria and Hospitalization Characteristics RESULTS We included in the analysis adult patients (>18 years of age) admitted through the emergency department (ED) and hospitalized in Washington from January 1, 2010
Among 86,378 adult patients who were hospitalized with sepsis in Washington State hospitals from 2010 through
Sepsis Mortality in Tertiary vs. Non-Tertiary Hospitals
3
September 2015, there were demographic and clinical differences (Table 1). Care for 76% of the sample was provided in 66 non-tertiary hospitals, compared with patients admitted to seven tertiary hospitals (24% of the sample). Patients who were admitted to tertiary care hospitals were younger, on average, by almost 5 years. The proportion of patients covered by Medicare insurance was thus higher in non-tertiary hospitals, compared with tertiary hospitals. In addition, slight statistically significant differences were observed in other demographic characteristics. Transfers from skilled nursing facilities, other hospitals, and admission from outside the hospital referral region were statistically higher in tertiary hospitals compared with non-tertiary hospitals (p < 0.001). Clinical characteristics for patients admitted to both hospital types were similar. No statistical differences were detected in number of comorbidities, proportion of patients with diabetes, length of stay, or classification of sepsis. Although rates of in-hospital deaths have declined for both non-tertiary and tertiary hospitals (Figure 1), abso-
lute counts differed. Comparing the years 2010–2014, the number of discharges from non-tertiary hospitals for sepsis has nearly doubled (from 7317 patients in 2010 to 14,481 patients in 2014). By contrast, the number of discharges from tertiary hospitals for sepsis has increased by 39% (from 2951 patients in 2010 to 4113 patients in 2014). Within non-tertiary hospitals, the number of deaths has changed little over time, ranging from a low of 758 in-hospital deaths in 2010 to 889 in-hospital deaths in 2012. Comparing 2010 with 2014 in-hospital deaths, the absolute number has declined 2%, with the biggest reduction occurring in 2014, where the number of discharges were the highest (14,481) and the number of deaths the lowest (758). In contrast, the number of deaths within tertiary hospitals has declined each full year of the study, from 232 in-hospital deaths in 2010 to 134 inhospital deaths in 2014. Comparing 2010 with 2014 inhospital deaths, the absolute number has declined 42%. A multivariable model that controlled for differences in patient populations served (Table 2) confirmed that there was increased risk for in-hospital death in non-
Table 1. Characteristics of Adults Hospitalized for Sepsis by Type of Hospital (N = 86,378) Non-Tertiary 66 Hospitals (n = 66,021)
Patient Characteristics
Patient age (years)
p-Value
Median
SD
Median
SD
65.2
19.0
60.3
19.5
Median
IQR
Median
IQR
2 4
1–2 3–6
2 4
1–2 3–6
Number of comorbidities Length of stay (days)
Females Insurance payer Medicare Medicaid Commercial Self-pay and charitable Dual enrolled Medicare and Medicaid Transfers from SNF Transfers from other hospital ICU admission Admissions outside HRR Comorbidities Heart failure Hypertension w/o complications Hypertension w/complications Chronic pulmonary disease Diabetes w/o complications Diabetes w/complications Kidney failure Drug abuse Sepsis disease classification 995.91 995.92 995.90 or 995.93 or 995.94
Tertiary 7 Hospitals (n = 20,357)
<0.001
0.86 0.39
n
%
n
%
34,965
53.0
10,381
51.0
<0.001
28,914 7897 16,041 3062 10,107 1737 778 21,833 4410
43.8 12.0 24.3 4.6 15.3 2.6 1.2 33.1 6.7
6077 3656 5977 1402 3245 826 714 6982 1606
29.9 18.0 29.4 6.9 15.9 4.1 3.5 34.3 7.9
<0.001
<0.001 <0.001 0.001 <0.001
8824 12,648 5004 12,204 9118 3028 6396 4062
13.4 19.2 7.6 18.5 13.8 4.6 9.7 6.2
2305 3456 1321 3918 2794 981 1876 1794
11.3 17.0 6.5 19.2 13.7 4.8 9.2 8.8
<0.001 <0.001 < 0.001 0.02 0.76 0.17 0.05 <0.001
45,753 20,053 215
69.3 30.4 0.3
14,009 6264 84
68.8 30.8 0.4
0.09
SNF = skilled nursing facilities; ICU = intensive care unit; HRR = hospital referral region as defined by the Dartmouth Atlas.
4
G. G. Salvatierra et al.
Figure 1. In-hospital mortality rate of adults hospitalized for sepsis by hospital type (N = 86,378). ICD10 = International Classification of Diseases, 10th Revision.
tertiary hospitals compared with tertiary hospitals. The overall mortality rate for all hospitals was 6.5% (5645 in-hospital deaths divided by 86,378 discharges). The unadjusted OR for in-hospital death was higher in nontertiary hospitals compared with tertiary hospitals (OR 1.37; 95% CI 1.28–1.47; p < 0.001). The fully adjusted
1.37 1.28–1.47 <0.001 1.43 1.34–1.53 <0.001 1.28 1.19–1.37 <0.001
model, controlling for year of discharge, age, sex, insurance payer, admission to the ICU, transfers from skilled nursing facilities, transfers from other hospitals, admissions outside the hospital referral region, and number of comorbidities yielded comparable results (OR 1.25; 95% CI 1.17–1.35; p < 0.001). Internal validation, comparing the derivation and validation cohorts, yielded equivalent results. Based on similarity of in-hospital mortality rates for the partial year in 2015 for non-tertiary hospitals (3.9%) and tertiary hospitals (4.4%), we derived the OR for a fully adjusted model. The derivation and validation cohorts were not consistent with the full study results (Table 3).
1.25 1.17–1.35 <0.001
DISCUSSION
Table 2. Unadjusted and Fully Adjusted Odds Ratio of InHospital Mortality of Adults Hospitalized for SIRS 2010–2015 (N = 86,378) Model Type Unadjusted model Controlling for year Controlling for year, age, sex, payer, ICU admission, transfers Controlling for year, age, sex, payer, ICU admission, transfers, admission outside hospital referral region, comorbidities
OR
95% CI
p-Value
Reference group = discharges from tertiary hospitals. SIRS = systemic inflammatory response syndrome; OR = odds ratio; CI = confidence interval; ICU = intensive care unit.
In a study of 86,378 patients hospitalized for sepsis in the State of Washington between 2010 and 2015, rates of mortality differed between tertiary and non-tertiary hospitals. Higher rates of mortality among patients in
Sepsis Mortality in Tertiary vs. Non-Tertiary Hospitals
5
Table 3. Unadjusted and Fully Adjusted Odds Ratio of In-Hospital Mortality of Adults Hospitalized for SIRS 2015 (N = 15,205) Derivation Cohort
Validation Cohort
Model Type
OR
95% CI
p-Value
OR
95% CI
p-Value
Unadjusted model Controlling for year, age, sex, payer, ICU admission, transfers Controlling for year, age, sex, payer, ICU admission, transfers, admission outside hospital referral region, comorbidities
0.66 0.68
0.60–0.99 0.52–0.88
0.04 0.004
1.02 0.82
0.77–1.35 0.62–1.10
0.88 0.20
0.68
0.52–0.89
0.004
0.81
0.60–1.08
0.16
Reference group = discharges from tertiary hospitals. SIRS = systemic inflammatory response syndrome; OR = odds ratio; CI = confidence interval; ICU = intensive care unit.
non-tertiary hospitals persisted in multivariable models controlling for demographic and clinical characteristics and in internally validated analyses. One of the major goals of national treatment guidelines is to identify and foster widespread implementation of evidence-based best practices to reduce mortality and morbidity from conditions that affect large numbers of patients. Despite the overall success of such highvisibility efforts as the Guidelines for the Early Management of Ischemic Stroke, and the Guideline for the Management of Heart Failure, there has been a persistent discrepancy in mortality outcomes between tertiary and non-tertiary hospitals (19,22). This has been true across a broad range of conditions, including those with and without national campaigns to encourage uniform treatment protocols (16). Our results, finding a similar tertiary/non-tertiary discrepancy for sepsis, show that the discrepancy persists even in the absence of an obvious resource-disparity explanation. Successful adherence to national guidelines for cardiovascular events requires prompt access to sophisticated diagnostic equipment and an on-call team capable of revascularization within stated time limits. These resources are not found at non-tertiary hospitals. For sepsis, however, the mortality gap cannot be accounted for with a similar explanation. Unlike stroke or myocardial infarction, which can be reliably diagnosed using a computed tomography scan, biomarkers, and electrocardiogram findings, sepsis diagnosis consistent with the SSC guidelines involves a cluster of clinical findings that can vary individual-to-individual or based on the context of clinical presentation (10,23). Whereas diagnoses requiring specialized technology or other scarce resources would be expected to differ by facility type, the SSC guidelines require nothing but clinical skills and laboratory capabilities common to both tertiary and non-tertiary hospitals. Yet based on our findings, a mortality gap exists for sepsis. The reasons for the disparity in sepsis mortality between tertiary and non-tertiary hospitals requires further investigation. Because data were not available about
sepsis bundle implementation, it is not possible for us to determine if there were significant differences between the use of the guidelines in tertiary vs. non-tertiary hospitals, or if the rate of uptake was different. Also, unknown from the available data is whether there are significant differences in nursing or physician awareness training, in-service education, availability of resources, or inhouse review for adherence to the guidelines, any one or a combination of which could account, in whole or part, for the observed discrepancy. The efficacy of sepsis-bundled interventions was established in teaching hospitals often classified as providing tertiary care. Reducing mortality rates and increasing diagnosis of sepsis has most often occurred in environments where sepsis interventions have been implemented and compliance with bundles is high (24,25). Achieving acceptance, implementation, and compliance with the SSC guidelines is a complex task that can be affected by hospital culture, available financial and other resources, institutional commitment, and many other factors that tend to favor tertiary care hospitals. Our results also support previous reports of both increasing incidence of sepsis and declining mortality rates for patients admitted to U.S. hospitals in the treatment of sepsis (26–28). Although we found lower mortality rates in tertiary hospitals, the tertiary/non-tertiary differences narrowed over the study span. In fact, the mortality rates in tertiary and non-tertiary hospitals converged by 2015. Though this was only partial-year data due to the ICD-9/10 changeover, it may point to a slower but eventually complete uptake and implementation of the SSC guidelines, and is another reason that there is urgent need for a prospective, large-scale study capable of elucidating the reasons for any confirmed discrepancy. Our finding of declining mortality in the treatment of hospitalized patients with sepsis confirms previous reports (29–31). During our study, the absolute numbers of patients dying from sepsis declined sharply in tertiary hospitals, but remained relatively stable in nontertiary hospitals. Substantial reductions in mortality since implementation of the sepsis bundles are a strong
6
G. G. Salvatierra et al.
indication that coordinated, preemptive action, increased level of suspicion, and prompt response combined with therapeutic advances hold the promise of further reductions in the sepsis mortality rate. Attributing reductions in mortality rates to bundled interventions, however, requires a more granular understanding of sepsis causality and mortality and experimental study designs. Limitations Our study findings should be understood within the context of several limitations. First, we did not know the compliance rates for implementing sepsis bundles at any of the hospitals in our study. In addition, the CHARS dataset did not include any information regarding which, if any, sepsis diagnosis criteria were used. We identified sepsis based on ICD-9-CM codes, and in Washington State, ICD-9-CM codes were discontinued starting October 1, 2015. Partial-year data for 2015 may account, in part, for the anomalous mortality results for that year. Hospital abstracts historically fail to capture complete data on patient race and ethnicity. Finally, the use of a retrospective study design does not allow any inferences of causality. We attempted to control for methodological inadequacies and inconsistencies through our study design. We assembled a large set of hospital abstracts over multiple years, from all civilian hospitals in Washington State. Our study design ensured adequate power and generalizability. We controlled for multiple prognostic factors including measures of severity of illness and comorbidities. We also selected hospitals within a single state that reported data in a standardized fashion. Finally, we internally validated our study findings using the splitfile approach. CONCLUSIONS We observed differences in the sepsis mortality rates between tertiary and non-tertiary hospitals for the years 2010–2014. Differences in mortality rates disappeared in 2015, although the exact reason for the convergence of mortality rates is unclear. Sepsis is a severe and urgent health care issue. There is need for a strong prospective study to determine, not only the degree to which there is a sepsis mortality gap, but also the causal factor or factors that create or contribute to such a gap. A prospective study could help inform efforts to achieve optimal treatment and diagnosis of sepsis, and thus reduce avoidable hospital mortality. REFERENCES 1. Singer M, Deutschman C, Seymour C, et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA 2016;315:801–10.
2. De Backer D, Dorman T. Surviving Sepsis Guideline: a continuous move toward better care of patients with sepsis. JAMA 2017;317: 807–8. 3. Angus D, Linde-Zwirble W, Lidicker J, et al. Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med 2001;29:1303–10. 4. Mayr F, Yende S, Angus D. Epidemiology of severe sepsis. Virulence 2014;5(1):4–11. 5. Liu V, Escobar G, Greene J, et al. Hospital deaths in patients with sepsis from 2 independent cohorts. JAMA 2014;312:90–2. 6. Gaieski DF, Edwards JM, Kallan M, Karr BG. Benchmarking the incidence and mortality of severe sepsis in the United States. Crit Care Med 2013;41:1167–74. 7. Lagu T, Rothberg M, Shieh M, et al. Hospitalizations, costs, and outcomes of severe sepsis in the United States 2003 to 2007. Crit Care Med 2012;40:754–61. 8. Dellinger R, Levy M, Rhodes A, et al. Surviving Sepsis Campaign Guidelines Committee including The Pediatric S. Surviving Sepsis Campaign: international guidelines for management of severe sepsis and septic shock, 2012. Intensive Care Med 2013;39:165– 228. 9. Khan P, Divatia J. Severe sepsis bundles. Indian J Crit Care Med 2010;14:8–13. 10. Cohen J, Vincent JL, Adhikari NK, et al. Sepsis: a roadmap for future research. Lancet Infect Dis 2015;15:581–614. 11. Levy M, Rhodes A, Phillips G, et al. Surviving Sepsis Campaign: association between performance metrics and outcomes in a 7.5year study. Crit Care Med 2015;43:3–12. 12. Rhodes A, Phillips G, Beale R, et al. The Surviving Sepsis Campaign bundles and outcome: results from the International Multicentre Prevalence Study on Sepsis (the IMPreSS study). Intensive Care Med 2015;41:1620–8. 13. Miller RR 3rd, Dong L, Nelson NC, et al. Multicenter implementation of a severe sepsis and septic shock treatment bundle. Am J Respir Crit Care Med 2013;188:77–82. 14. Seymour CW, Gesten F, Prescott HC, et al. Time to treatment and mortality during mandated emergency care for sepsis. N Engl J Med 2017;376:2235–44. 15. ProCESS Investigators. A randomized trial of protocol-based care for early septic shock. N Engl J Med 2014;370:1683–93. 16. Villapiano N, Iwashyna TJ, Davis MM. Worsening rural-urban gap in hospital mortality. J Am Board Fam Med 2017;30(6):816–23. 17. Baldwin LM, Chan L, Andrilla CHA, Huff ED, Hart LG. Quality of care for myocardial infarction in rural and urban hospitals. J Rural Health 2010;26:51–7. 18. Leira EC, Hess DC, Torner JC, Adams HP Jr. Rural-urban differences in acute stroke management practices: a modifiable disparity. Arch Neurol 2008;65:887–91. 19. Yancy C, Jessup M, Bozkurt B, et al. 2017 ACC/AHA/HFSA focused update of the 2013 ACCF/AHA Guideline for the Management of Heart Failure: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice guidelines and the Heart Failure Society of America. Circulation 2017;136:e137–61. 20. Johns Hopkins Medicine. Tertiary care definition. Available at: https://www.hopkinsmedicine.org/patient_care/pay_bill/insurance_ footnotes.html. Accessed April 14, 2017. 21. Elixhauser A, Steiner C, Harris D, et al. Comorbidity measures for use with administrative data. Med Care 1998;36:8–27. 22. Jauch EC, Saver JL, Adams HP, et al. Guidelines for the early management of patients with acute ischemic stroke. 2013;44:870–947. 23. Levy MM, et al. 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. Intensive Care Med 2003;29:530–8. 24. Levy M, Dellinger P, Townsend S, et al. The Surviving Sepsis Campaign: results of an international guideline-based performance improvement program targeting severe sepsis. Intensive Care Med 2010;36:222–31. 25. Henriksen D, Laursen C, Jensen T, et al. Incidence rate of community-acquired sepsis among hospitalized acute medical patients—a population-based survey. Crit Care Med 2015;43:13–21.
Sepsis Mortality in Tertiary vs. Non-Tertiary Hospitals 26. Fleischmann C, Scherag A, Adhikari N, et al. Assessment of global incidence and mortality of hospital-treated sepsis. Current estimates and limitations. Am J Respir Crit Care Med 2016;193:259–72. 27. Walkey A, Lagu T, Lindenauer P. Trends in sepsis and infection sources in the United States. A population-based study. Ann Am Thorac Soc 2015;12:216–20. 28. Zimmerman JE, Kramer AA, Knaus WA. Changes in hospital mortality for United States intensive care unit admissions from 1988 to 2012. Crit Care 2013;17:R81.
7 29. Kaukonen K, Bailey M, Suzuki S, et al. Mortality related to severe sepsis and septic shock among critically ill patients in Australia and New Zealand, 2000–2012. JAMA 2014;311:1308–16. 30. Stevenson E, Rubenstein A, Radin G, et al. Two decades of mortality trends among patients with severe sepsis: a comparative metaanalysis. Crit Care Med 2014;42:625–31. 31. Liu V, Escobar G, Greene J, et al. Multicenter implementation of a treatment bundle for patients with sepsis and intermediate lactate values. Am J Respir Crit Care Med 2016;193:1264–70.
8
G. G. Salvatierra et al.
ARTICLE SUMMARY 1. Why is this topic important? Sepsis is a leading cause of hospital deaths and one of several conditions for which there are widely accepted national treatment guidelines. Unlike the other leading causes of in-hospital mortality, diagnosis, and treatment does not rely on resources that should differ between tertiary and non-tertiary hospitals, so any discrepancy in outcomes needs further exploration and explanation. 2. What does this study attempt to show? The study attempts to show that there are significant differences in sepsis mortality between tertiary and nontertiary hospitals for one 5-year period in one state. 3. What are the key findings? Non-tertiary hospitals in Washington State over a 5year span had higher in-hospital mortality due to sepsis than did tertiary hospitals. 4. How is patient care impacted? The majority of patients are seen in non-tertiary hospitals. Unlike some conditions for which guidelines are published, sepsis recognition and treatment does not call for exceptional human or technical resources, so discrepancies in such resources are not an explanation for the sepsis mortality gap. Research should be directed at better understanding the factors that cause non-tertiary hospitals to have a higher rate of sepsis mortality.