- Email: [email protected]
The Association Between Sepsis and Potential Medical Injury Among Hospitalized Patients
CHEST
Original Research CRITICAL CARE
The Association Between Sepsis and Potential Medical Injury Among Hospitalized Patients Vincent Liu, MD; Benjamin J. Turk, BA; Norman W. Rizk, MD, FCCP; Patricia Kipnis, PhD; and Gabriel J. Escobar, MD
Background: Patient safety remains a national priority, but the role of disease-specific characteristics in safety is not well characterized. Methods: We identified potentially preventable medical injuries using patient safety indicators (PSIs) and annual data from the Nationwide Inpatient Sample between 2003 and 2007. We compared the rate of selected PSIs among patients hospitalized with and without sepsis. Among patients with sepsis, we also compared PSI rates across severity strata. Using multivariable casecontrol matching and regression analyses, we estimated the excess adverse outcomes associated with PSI events in patients with sepsis. Results: Patients hospitalized with sepsis accounted for 2% to 4% of hospital discharges; however, they accounted for 9% to 26% of all potential medical injuries. PSI rates varied considerably; among patients hospitalized for sepsis, they were lowest for accidental puncture or laceration and highest for postoperative respiratory failure. Nearly all PSI rates were higher among patients with sepsis compared with patients without sepsis. Among those with sepsis, most PSI rates increased as sepsis severity increased. Compared with matched sepsis control subjects, increased length of stay and hospital charges were associated with PSI events in sepsis cases. However, only decubitus ulcer, iatrogenic pneumothorax, and postoperative metabolic and physiologic derangement or respiratory failure were associated with excess mortality. Conclusion: Patients hospitalized for sepsis, compared with the general hospital population, were at a substantially increased risk of potential medical injury; their risk rose as disease severity increased. Future patient safety efforts may benefit from focusing on medically vulnerable populations. CHEST 2012; 142(3):606–613 Abbreviations: CVC 5 central venous catheter; DRG 5 diagnosis-related group; ICD-9-CM 5 International Classification of Diseases, 9th Edition, Clinical Modification; LOS 5 length of stay; NIS 5 Nationwide Inpatient Sample; PE 5 pulmonary embolism; PSI 5 patient safety indicator
Err is Human estimated that as many as To98,000 hospitalized patients die annually from pre1
ventable medical errors. Subsequent attention has aimed to identify and implement interventions that improve patient safety.2-5 Although efforts to reduce specific harms have been promising, large-scale
Manuscript received October 7, 2011; revision accepted February 1, 2012. Affiliations: From the Division of Research and Systems Research Initiative (Drs Liu, Kipnis, and Escobar and Mr Turk), and the Department of Management, Information, and Analysis (Dr Kipnis), Kaiser Permanente, Oakland; the Department of Critical Care (Dr Liu), Kaiser Permanente, Santa Clara; the Department of Inpatient Pediatrics (Dr Escobar), Kaiser Permanente, Walnut Creek; and the Division of Pulmonary and Critical Care Medicine (Dr Rizk), Stanford University, Stanford, CA. 606
improvements in safety have been limited.6-11 Few studies have evaluated how disease-specific characteristics or severity of illness might contribute to patient safety.1,12-14 To further investigate the role of disease severity in patient safety, we chose to describe the scope of potential medical injury among patients hospitalized Funding/Support: Dr Liu was supported by the Agency for Healthcare Research and Quality [Grant F32HS019181-01]. Correspondence to: Vincent Liu, MD, 2000 Broadway, Oakland, CA, 94612; e-mail: [email protected] © 2012 American College of Chest Physicians. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians. See online for more details. DOI: 10.1378/chest.11-2556 Original Research
with sepsis. Sepsis is common among inpatients and associated with significant morbidity, mortality, and resource use.15-17 Guidelines aimed at improving care among patients with sepsis also recommend intensive care and invasive procedures for selected patients, potentially increasing their risk of adverse events during hospitalization.15-20 Using data from the Nationwide Inpatient Sample (NIS) between 2003 and 2007, we describe the impact of sepsis on patient safety by comparing the rates of selected potentially preventable medical injuries among 33 million inpatients with and without sepsis. We also evaluate whether increased sepsis severity was associated with worsened patient safety and estimate the burden of potential medical injuries on outcomes among patients with sepsis.
Materials and Methods Data The NIS datasets contain discharge abstracts from up to 8 million hospitalizations annually and approximate a 20% sample of US acute-care community hospitals.21 Data elements available for each hospitalization include International Classification of Diseases, 9th Edition, Clinical Modification (ICD-9-CM) diagnosis and procedure codes, diagnosis-related groups (DRGs), discharge vital status, length of stay (LOS), hospital charges, and patient demographics. Sepsis Definitions We defined sepsis hospitalizations in two categories (narrow and broad) based on ICD-9-CM diagnosis coding.15-17 The narrow definition included hospitalizations for which sepsis codes were listed as the primary hospital diagnosis (ie, reason for hospitalization): 020.0 (septicemic), 038 and subtypes (septicemia), 785.52 (septic shock), 790.7 (bacteremia), 995.91 (sepsis), and 995.92 (severe sepsis). In the broad definition, we included hospitalizations for which sepsis was listed in either the primary or secondary diagnosis positions, to ensure consistency with definitions based on previous literature.15-17 We determined sepsis severity based on the presence of organ system dysfunction or failure by ICD-9-CM coding.15,16 For example, respiratory dysfunction was defined by the following secondary ICD-9-CM codes: 518.5 (acute respiratory distress syndrome after shock or trauma), 518.81 (acute respiratory failure), 518.82 (ARDS), 786.09 (respiratory insufficiency), 799.1 (respiratory arrest), or 96.7 (ventilator management). Severity was categorized by dysfunction in none, one or two, or three or more organ systems.15,16 Severity of Comorbid Illness To account for preexisting illness, we created a comorbidity index based on the risk of hospital mortality.22 The regression coefficients for predicted mortality based on Elixhauser categories were estimated using discharges in the 2007 NIS while controlling for age, sex, race/ethnicity, insurance status, DRG fixed effects, and all patient refined-DRG mortality risk subclasses.23 The coefficients were then used as relative weights to summarize a comorbidity index.24 This index was applied to all discharges journal.publications.chestnet.org
between 2003 and 2007; discharges without comorbidity data (18.0%; n 5 7,234,780) were removed. Identifying Medical Injuries The patient safety indicators (PSIs), version 4.2, were used to identify potentially preventable medical injuries. These indicators were designed to screen for potential adverse events resulting from exposure to health care.25 Each PSI is determined by select inclusion and exclusion criteria to describe a specific group of at-risk patients (denominator) and individuals who experience a potential adverse event (numerator). We evaluated the following PSIs: accidental puncture or laceration, decubitus ulcer, iatrogenic pneumothorax, and central venous catheter (CVC)-related bloodstream infection, as well as postoperative events including hemorrhage or hematoma, hip fracture, physiologic and metabolic derangement, pulmonary embolism (PE) or DVT, respiratory failure, and wound dehiscence. We excluded PSIs that overlapped with our inclusion criteria or outcomes (postoperative death, death in low-mortality DRGs, and postoperative sepsis), were extremely rare (transfusion reaction), or were specific to childbirth (obstetric vaginal trauma). Because our primary interest was evaluating patients hospitalized for sepsis, we revised PSIs that captured events following elective surgery (physiologic and metabolic derangement and respiratory failure) to assess only those who had surgery during nonelective hospital admission. Comparing Sepsis and Nonsepsis Events To evaluate the frequency of potential medical injury, we measured the annual rate of PSI events among patients with and without sepsis. We defined the PSI sepsis rate ratio as the rate of each PSI occurring among sepsis hospitalizations divided by the rate of the same PSI occurring among nonsepsis hospitalizations for both narrowly and broadly defined sepsis groups. We determined the mean PSI sepsis rate ratio and 95% CI by calculating a weighted average of annual values. PSI rates were also compared among patients with sepsis stratified by sepsis severity. For PSIs that identified specific organ system failure (eg, postoperative respiratory failure), ICD-9-CM coding in the same organ system (eg, respiratory dysfunction) was not counted toward sepsis severity or organ failure to avoid double-counting organ dysfunction as a measure of severity and outcome. We also evaluated the likelihood of PSI events among patients with and without sepsis after adjustment for hospital LOS based on two methods. First, we calculated LOS-weighted PSI sepsis rate ratios in which individual hospitalizations were divided by their hospital LOS. For example, the LOS-adjusted PSI contribution for a patient with a 10-day hospitalization was 0.1; comparatively, the PSI contribution from a 2-day hospitalization was 0.5. We also performed a bivariate logistic regression for the odds of each PSI adjusted for a diagnosis of sepsis and hospital LOS. Excess Adverse Outcomes Among Patients With Sepsis We estimated excess LOS, total charges, and mortality associated with potential adverse events among patients with a narrow sepsis diagnosis by two methods. First, we performed multivariable matching to minimize confounding resulting from unmeasured differences in clinical and hospital characteristics.22,26-30 Each PSI sepsis case was matched with up to four randomly selected sepsis control subjects from the same hospital and year; cases and control subjects were also matched by sex, race, age by decade, severity of organ dysfunction, and comorbid illness. Cases and control subjects without any comorbid illness were matched; CHEST / 142 / 3 / SEPTEMBER 2012
607
Table 1—Baseline Characteristics and Hospital Outcomes of Hospitalizations With and Without a Primary Hospital Diagnosis of Sepsis Between 2003 and 2007 Primary Hospital Diagnosis Category Characteristic or Outcome Characteristics Age, %, y ⱕ 17 18-39 40-64 ⱖ 65 Male Race/ethnicity, white Primary insurance Medicare Medicaid Private Other Comorbidity Without comorbidity Mean comorbidity indexa Hospital outcome Mortality Median length of stay, d Total charges, $
Nonsepsis (n 5 32,688,780)
Sepsis (n 5 534,416)
178,949 (0.6) 8,556,403 (26.2) 10,715,634 (32.8) 13,237,794 (40.5) 12,621,118 (38.6) 16,657,662 (51.0)
0 (0.0) 30,743 (5.8) 145,842 (27.3) 357,831 (67.0) 247,916 (46.4) 281,260 (52.6)
14,323,980 (43.9) 4,756,645 (14.6) 10,662,802 (32.7) 2,885,509 (8.8)
383,411 (71.8) 43,299 (8.1) 83,139 (15.6) 23,995 (4.5)
25,012,546 (77.4) 9.1 (9.6)
213,708 (40.3) 11.6 (10.9)
709,002 (2.2) 3 (2-5) 24,438 (38,762)
97,895 (18.3) 6 (4-11) 47,593 (72,648)
Values represent frequency (%) or median (interquartile range). Mean value when comorbidity present.
Second, we estimated excess adverse outcomes with linear regression so that estimates could be directly compared with those from matching analysis.22 Regression analyses controlled for patient age, sex, race, severity of organ dysfunction, comorbid illness index, and year; hospital fixed effects were adjusted for but not calculated. The regression analyses included all case and control subjects from each sepsis PSI risk pool. For cross-validation of estimates for mortality, we also performed conditional logistic regression using the multivariable matched sets and unmatched sets. A P value of , .05 was considered statistically significant for all analyses. Data were described as mean (SD), frequency (%), and rate (95% CI).
Results During the study period, 534,416 hospitalizations (1.6%) had a primary hospital diagnosis of sepsis (Table 1). Patients with sepsis were older and more frequently had preexisting comorbid illness (59.7%) than patients without sepsis (22.5%). Mortality was 18.3% among patients with sepsis and 2.2% among patients without sepsis (P , .01). Nearly one-half of hospitalized patients with sepsis had organ dysfunction (Table 2). Outcomes worsened as the severity of organ dysfunction increased; for example, mortality was 8.3% for patients without organ dysfunction and 53.7% for those with dysfunction in three or more systems.
a
PSI Rates those with comorbidities were matched to within a 2.5% difference in risk of death due to comorbidities. For each PSI, excess LOS was calculated as the difference between the LOS for a case and the mean LOS for matched controls. Pooled mean excess LOS was then determined using all matched sets; the paired t test was used to evaluate whether the pooled excess LOS was significantly different from zero. Similar methods were used to measure excess hospital charges and mortality. To increase matching frequency, we also performed multivariable matching without consideration of unique hospitals.
A total of 628,998 total PSI events were identified; 58,295 (9.3%) were among hospitalizations with a primary hospital diagnosis of sepsis. PSI rates varied considerably in patients with and without sepsis (Table 3); for example, among patients with sepsis, PSI rates varied from 0.9 (accidental puncture or laceration) to 302.9 (respiratory failure) per thousand at-risk discharges. The PSI sepsis rate ratio was greater than
Table 2—Characteristics and Outcomes of Patients With Sepsis by Severity Sepsis Severity (No. of Dysfunctional Organ Systems) Characteristic or Outcome No. (%) Outcomes Mortality Length of stay, d Total charges, $ Organ failure frequency Cardiovascular Hematologic Hepatic Metabolic Neurologic Renal Respiratory
0 273,427 (51.1) 22,700 (8.3) 7.5 (7.4) 31,572 (42,463) … … … … … … …
1-2 224,800 (42.1) 55,790 (24.8) 10.1 (11.5) 58,870 (83,575) 38,431 (17.1) 33,671 (15.0) 3,748 (1.7) 25,520 (11.4) 11,305 (5.0) 121,568 (54.1) 61,318 (27.3)
31 36,189 (6.8) 19,405 (53.7) 12.0 (14.4) 99,339 (125,198) 12,053 (33.3) 16,927 (46.8) 6,185 (17.1) 20,774 (57.4) 4,784 (13.2) 31,750 (87.7) 27,618 (76.3)
Overall 534,416 97,895 (18.3) 8.9 (10.0) 47,593 (72,648) 50,484 (9.5) 50,598 (9.5) 9,933 (1.9) 46,294 (8.7) 16,089 (3.0) 153,318 (28.7) 88,936 (16.6)
Severity was categorized by the number of dysfunctional organ systems as indicated by organ-specific secondary ICD-9-CM coding. Patients with sepsis include only those with a primary hospital diagnosis of sepsis. ICD-9-CM 5 International Classification of Diseases, 9th Edition, Clinical Modification. 608
Original Research
Table 3—Annual Rates of PSI Events (per 1,000 At-Risk Hospital Discharges) Stratified by Patients With and Without Sepsis and the Overall PSI Sepsis Rate Ratio Annual PSI Rate per 1,000 At-Risk Discharges Patient Safety Indicator Accidental puncture or laceration Sepsis Nonsepsis CVC-related bloodstream infection Sepsis Nonsepsis Decubitus ulcer Sepsis Nonsepsis Iatrogenic pneumothorax Sepsis Nonsepsis Postoperative hemorrhage or hematoma Sepsis Nonsepsis Postoperative hip fracture Sepsis Nonsepsis Postoperative physiologic or metabolic derangement Sepsis Nonsepsis Postoperative PE/DVT Sepsis Nonsepsis Postoperative respiratory failure Sepsis Nonsepsis Postoperative wound dehiscence Sepsis Nonsepsis
2003
2004
2005
2006
2007
P Trend
PSI Sepsis Rate Ratio (95% CI)
0.9 3.2
1.3 3.2
1.6 3.3
1.3 3.3
1.7 3.4
a
0.4 (0.4-0.5)
10.3 2.0
12.2 2.0
12.5 2.0
11.4 1.9
9.8 1.5
a
6.0 (5.7-6.4)
109.1 21.3
119.0 22.5
121.5 22.3
121.9 23.7
130.2 24.4
a
5.3 (5.2-5.4)
1.2 0.5
1.5 0.5
1.5 0.5
1.6 0.5
1.7 0.5
a
2.8 (2.5-3.2)
3.6 2.8
2.7 2.7
4.4 2.7
4.6 2.6
4.2 2.7
… …
1.5 (1.2-1.7)
1.6 0.2
1.7 0.2
1.1 0.2
1.5 0.2
1.2 0.2
… a
6.7 (6.0-7.3)
38.9 3.7
36.6 4.1
48.4 4.1
41.8 4.0
49.2 4.4
a
10.6 (9.8-11.4)
41.5 9.7
45.8 10.2
49.4 10.3
44.3 11.1
49.7 11.6
a
4.4 (4.2-4.6)
221.3 44.2
247.4 49.0
279.4 42.9
282.1 45.6
302.9 46.9
a …
5.9 (5.3-6.4)
4.1 1.8
7.1 2.1
6.6 2.3
5.2 2.3
5.6 2.2
… a
2.7 (2.3-3.0)
a
a
a
a
a
a
The PSI sepsis rate ratio is the mean ratio of annual ratios calculated as each PSI among patients with a primary hospital diagnosis of sepsis divided by the same PSI among patients without sepsis. CVC 5 central venous catheter; PE 5 pulmonary embolism; PSI 5 patient safety indicator. aP , .05 for trend.
unity for all PSIs except accidental puncture or laceration (Table 3). The remaining sepsis rate ratios varied from 1.5 for hemorrhage or hematoma to 10.6 for physiologic and metabolic derangement. Sepsis hospitalizations under the broad definition accounted for 162,872 (25.9%) of all identified PSI events. The PSI sepsis rate ratios were also greater than unity for sepsis hospitalizations under the broad definition (e-Table 1). After adjustment for hospital LOS, the results remained similar; only accidental puncture or laceration was less likely among patients with sepsis compared with patients without sepsis by logistic regression (OR, 0.4; P , .001) or LOS-weighted comparisons (e-Tables 2, 3). Most PSI rates increased as sepsis-related organ dysfunction worsened (Fig 1). The greatest rate increase across severity strata was seen for iatrogenic pneumothorax, increasing from 0.5 to 5.1 per thousand at-risk discharges across strata. Changes in PSI rates across severity strata were journal.publications.chestnet.org
not significantly different for decubitus ulcer, PE/DVT, wound dehiscence, and hip fracture. Excess Adverse Outcomes Multivariable matching of patients with sepsis within hospital facilities produced at least one matched control subject for 52% of all PSI cases; matching frequency was lowest for wound dehiscence (25%) and highest for accidental puncture or laceration (83%). Matching without consideration for hospital facilities produced high matching frequencies, ranging from 96% to 100%. Estimates of excess adverse outcomes associated with PSI events are shown in Figure 2. Significant increases in hospital LOS were present for all PSIs except hemorrhage or hematoma and hip fracture. Similarly, hospital charges were increased for all PSIs except postoperative hip fracture. Decubitus ulcer, iatrogenic pneumothorax, metabolic and physiologic CHEST / 142 / 3 / SEPTEMBER 2012
609
Figure 1. Percent change from baseline of patient safety indicator (PSI) event rates by level of sepsis severity. Sepsis hospitalizations include only those with a primary hospital diagnosis of sepsis. Filled circles and lines represent statistically significant differences by linear regression analysis; open circles and dotted lines represent nonstatistically significant differences. Abbreviated PSI descriptions include those for postoperative complications including physiologic derangement, respiratory failure, hemorrhage, PE/DVT, wound dehiscence, and hip fracture. Others include accidental puncture and CVC-infection. Accidental puncture 5 accidental puncture or laceration; CVC-infection 5 central venous catheter-related bloodstream infection; hemorrhage 5 hemorrhage or hematoma; PE 5 pulmonary embolism; physiologic derangement 5 physiologic or metabolic derangement.
The heightened vulnerability to potential injury among patients with sepsis is multifactorial and likely reflects the inherent risk associated with the disease, the underlying characteristics of patients with sepsis, as well as the adverse events resulting from treatment. Patients with sepsis have high mortality compared with other inpatients. In this study, overall sepsis mortality was 18% and was nearly 29% among patients with any organ dysfunction; these rates match those reported in prior studies.15,16 The increased age and frequency of comorbid illness among patients with sepsis also contribute to an increased risk of adverse events.13,17,31,32 Nearly one-half of patients with sepsis in this study had evidence of organ dysfunction. Current care guidelines for patients with severe sepsis recommend treatments that often require intensive care monitoring and invasive interventions.18 Adverse events, often life-threatening, occur more frequently among
derangement, and respiratory failure were associated with excess mortality (Fig 2). Patients with sepsis with respiratory failure had the highest excess mortality (25.3%); iatrogenic pneumothorax was also associated with high excess mortality (16.2%). Excess hospital mortality was not present for accidental puncture or laceration, CVC-associated bloodstream infection, hemorrhage or hematoma, hip fracture, and PE/DVT. Mortality estimates were similar by conditional logistic regression. Discussion This study demonstrated that patients with sepsis are at high risk for potential medical injuries when compared with the general hospital population. Using the PSIs, patients discharged with a primary hospital diagnosis of sepsis were nearly three times as likely to have iatrogenic pneumothorax, more than six times as likely to have CVC-related bloodstream infection or postoperative respiratory failure, and more than 10 times as likely to have postoperative physiologic or metabolic derangement than patients without sepsis. Although , 2% of hospital discharges had a primary hospital diagnosis of sepsis, they accounted for nearly 10% of all potential adverse events. Similarly, the 4% of hospitalizations with a broad diagnosis of sepsis accounted for 26% of all PSI events, highlighting the impact that specific disease conditions have on patient safety. 610
Figure 2. Estimates of excess adverse outcomes among patients with sepsis with PSI events. A, Length of stay. B, Hospital charges. C, Mortality. Patients with sepsis include only those with a primary hospital diagnosis of sepsis. For each PSI, point estimates (markers) and 95% CIs (lines) derived from three methods are displayed. Open circles and squares represent multivariable matching analysis within hospital facility and without hospital facility, respectively. Filled triangles are estimates from linear regression analysis. See Figure 1 legend for expansion of abbreviations. Original Research
patients admitted to the ICU.9,13,33-39 Although the NIS data did not allow us to specifically identify which patients underwent intensive care, we found that for most PSIs, the highest rates of potential medical injury were present among those patients with the most severe sepsis who would also likely require critical care. Invasive procedures also increase the risk of injury among patients with sepsis. For patients with severe sepsis, standard treatment algorithms rely on CVCguided therapy.18,40 Although these interventions may be associated with improved mortality,20,40 CVC insertion can result in iatrogenic injuries, including pneumothorax and bloodstream infection.41,42 We found that the rates of PSI-identified iatrogenic pneumothorax and CVC-related bloodstream infections increased as organ dysfunction worsened; for example, the rate of pneumothorax was tenfold higher among patients with the most severe organ dysfunction than those without dysfunction. Importantly, when they occurred in patients with sepsis, most PSI events were associated with increased hospital LOS and charges. However, only selected PSI events were associated with excess hospital mortality. This study has several strengths. We used a nationally representative sample of hospitalizations to compare rates of potential medical injury among sepsis and general hospital populations over 5 years, providing a broad perspective on patient safety in hospitals. The large sample size also provided sufficient power to evaluate adverse hospital outcomes associated with PSIs among patients with sepsis and resulted in robust estimates. However, this study also has notable limitations. Since this study was based on administrative data, we did not have detailed clinical variables necessary to ensure adequate risk adjustment for severity of illness between patients with sepsis as well as patients without sepsis. These data also lacked information about the timing of events within a hospitalization, thus hampering the ability to define causal relationships between PSIs and outcomes. Using the PSIs alone to identify medical injuries also has several limitations. First, they depend on accurate ICD-9-CM coding; inaccuracies can result from multiple sources and impact the screening capability of PSIs.43,44 Second, studies have suggested that some PSIs identify conditions that were “present on admission” rather than those that resulted from hospitalization.45,46 A handful of PSIs, such as decubitus ulcer, may be particularly affected. Unfortunately, in this study, coding for ICD-9-CM diagnoses present at hospital admission was not available; nonetheless, even after removing all decubitus ulcer events, patients with broadly-defined sepsis still accounted for 24% of all other identified PSI events. journal.publications.chestnet.org
Third, the PSIs screen for a subset of all potential adverse hospital events. For example, they do not evaluate injuries that result from medication errors, an important source of medical injury.47 As a result, the overall rates of injuries derived from the PSIs may substantially underestimate the true rate of adverse events.48 Finally, the PSIs were not designed to definitively identify medical injuries, since indicators may flag events that occur during the natural course of disease or treatment rather than events that result from failures in patient care. This limitation might be particularly acute in the case of sepsis, since the overwhelming inflammatory response to infection can affect nearly all organ systems. Thus, the broadly defined PSIs may lack some discrimination necessary to differentiate the sequelae of disease (eg, especially for postoperative metabolic and physiologic derangement or respiratory failure, which are common in sepsis) or the precursors of disease (eg, CVC-related infections) from actual preventable medical injuries. Thus, although PSIs may not be adequate instruments to benchmark and compare hospital quality or safety,45,49,50 they can highlight comparative differences in safety and potential deficiencies in care. In conclusion, this study found that sepsis was associated with an increased risk for potential medical injury; this risk rose as organ dysfunction increased. Sepsis hospitalizations accounted for between 9% to 26% of all identified potential adverse events. These findings suggest that severity of illness strongly influences patient safety and interventions aimed at reducing patient harm should focus especially on medically vulnerable populations. Acknowledgments Author contributions: Dr Liu had full access to the data and takes responsibility for the integrity of the data and the accuracy of the data analysis. Dr Liu: contributed to study design; data collection, analysis, and interpretation; preparation of the manuscript; and review and approval of the final manuscript. Mr Turk: contributed to study design, data analysis and interpretation, preparation of the manuscript, and review and approval of the final manuscript. Dr Rizk: contributed to study design, data interpretation, preparation of the manuscript, and review and approval of the final manuscript. Dr Kipnis: contributed to study design, data analysis and interpretation, preparation of the manuscript, and review and approval of the final manuscript. Dr Escobar: contributed to study design, data analysis and interpretation, preparation of the manuscript, and review and approval of the final manuscript. Financial/nonfinancial disclosures: The authors have reported to CHEST the following conflicts of interest: Dr Escobar received grant support from MedImmune, LLC (for studies on respiratory syncytial virus infection in infants), and Novartis AG (for a study on ICU-acquired infection). Drs Liu, Rizk, and Kipnis, and Mr Turk have reported that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article. Role of sponsors: The sponsor was not involved in the design and conduct of the study; collection, management, analysis, and CHEST / 142 / 3 / SEPTEMBER 2012
611
interpretation of the data; and preparation, review, or approval of the manuscript. Additional information: The e-Tables can be found in the “Supplemental Materials” area of the online article.
References 1. Kohn LT, Corrigan JM, Donaldson MS. To Err Is Human: Building a Safer Health System. Washington, DC: National Academies Press; 1999. 2. Agency for Healthcare Research and Quality. Medical errors and patient safety. Agency for Healthcare Research and Quality website. http://www.ahrq.gov/qual/patientsafetysix.htm. Published 2010. October 1, 2011. 3. Clancy CM. Ten years after To Err is Human. Am J Med Qual. 2009;24(6):525-528. 4. Pronovost PJ, Nolan T, Zeger S, Miller M, Rubin H. How can clinicians measure safety and quality in acute care? Lancet. 2004;363(9414):1061-1067. 5. Berwick DM, Calkins DR, McCannon CJ, Hackbarth AD. The 100,000 lives campaign: setting a goal and a deadline for improving health care quality. JAMA. 2006;295(3):324-327. 6. Pronovost P, Needham D, Berenholtz S, et al. An intervention to decrease catheter-related bloodstream infections in the ICU. N Engl J Med. 2006;355(26):2725-2732. 7. Haynes AB, Weiser TG, Berry WR, et al; Safe Surgery Saves Lives Study Group. A surgical safety checklist to reduce morbidity and mortality in a global population. N Engl J Med. 2009;360(5):491-499. 8. Bates DW, Leape LL, Cullen DJ, et al. Effect of computerized physician order entry and a team intervention on prevention of serious medication errors. JAMA. 1998;280(15): 1311-1316. 9. Landrigan CP, Rothschild JM, Cronin JW, et al. Effect of reducing interns’ work hours on serious medical errors in intensive care units. N Engl J Med. 2004;351(18):1838-1848. 10. Landrigan CP, Parry GJ, Bones CB, Hackbarth AD, Goldmann DA, Sharek PJ. Temporal trends in rates of patient harm resulting from medical care. N Engl J Med. 2010;363(22):2124-2134. 11. Rosen AK, Zhao S, Rivard P, et al. Tracking rates of patient safety indicators over time: lessons from the Veterans Administration. Med Care. 2006;44(9):850-861. 12. Pronovost PJ, Goeschel CA, Marsteller JA, Sexton JB, Pham JC, Berenholtz SM. Framework for patient safety research and improvement. Circulation. 2009;119(2):330-337. 13. Brennan TA, Leape LL, Laird NM, et al. Incidence of adverse events and negligence in hospitalized patients. Results of the Harvard Medical Practice Study I. N Engl J Med. 1991;324(6):370-376. 14. Naessens J, Campbell CR, Shah N, et al. Effect of illness severity and comorbidity on patient safety and adverse events. Am J Med Qual. 2012;27(1):48-56. 15. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR. Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med. 2001;29(7):1303-1310. 16. Martin GS, Mannino DM, Eaton S, Moss M. The epidemiology of sepsis in the United States from 1979 through 2000. N Engl J Med. 2003;348(16):1546-1554. 17. Dombrovskiy VY, Martin AA, Sunderram J, Paz HL. Rapid increase in hospitalization and mortality rates for severe sepsis in the United States: a trend analysis from 1993 to 2003. Crit Care Med. 2007;35(5):1244-1250. 18. Dellinger RP, Levy MM, Carlet JM, et al; International Surviving Sepsis Campaign Guidelines Committee; American 612
19.
20.
21.
22. 23. 24. 25.
26.
27. 28. 29.
30. 31.
32. 33. 34.
Association of Critical-Care Nurses; American College of Chest Physicians; American College of Emergency Physicians; Canadian Critical Care Society; European Society of Clinical Microbiology and Infectious Diseases; European Society of Intensive Care Medicine; European Respiratory Society; International Sepsis Forum; Japanese Association for Acute Medicine; Japanese Society of Intensive Care Medicine; Society of Critical Care Medicine; Society of Hospital Medicine; Surgical Infection Society; World Federation of Societies of Intensive and Critical Care Medicine. Surviving Sepsis Campaign: international guidelines for management of severe sepsis and septic shock: 2008. Crit Care Med. 2008;36(1):296-327. Ferrer R, Artigas A, Levy MM, et al; Edusepsis Study Group. Improvement in process of care and outcome after a multicenter severe sepsis educational program in Spain. JAMA. 2008;299(19):2294-2303. Levy MM, Dellinger RP, Townsend SR, et al. The Surviving Sepsis Campaign: results of an international guideline-based performance improvement program targeting severe sepsis. Crit Care Med. 2010;38(2):367-374. Introduction to the HCUP Nationwide Inpatient Sample (NIS). Healthcare Cost and Utilization Project website. http://www.hcup-us.ahrq.gov/db/nation/nis/NIS_Introduction_ 2007.jsp. Published 2007. Accessed November 16, 2010. Zhan C, Miller MR. Excess length of stay, charges, and mortality attributable to medical injuries during hospitalization. JAMA. 2003;290(14):1868-1874. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998; 36(1):8-27. Romano PS, Chan BK. Risk-adjusting acute myocardial infarction mortality: are APR-DRGs the right tool? Health Serv Res. 2000;34(7):1469-1489. Patient Safety Indicators Overview. AHRQ quality indicators. Agency for Healthcare Research and Quality website http:// www.qualityindicators.ahrq.gov/modules/psi_overview.aspx. Published February 2006. Accessed November, 16, 2010. Classen DC, Pestotnik SL, Evans RS, Lloyd JF, Burke JP. Adverse drug events in hospitalized patients. Excess length of stay, extra costs, and attributable mortality. JAMA. 1997; 277(4):301-306. Bates DW, Spell N, Cullen DJ, et al; Adverse Drug Events Prevention Study Group. The costs of adverse drug events in hospitalized patients. JAMA. 1997;277(4):307-311. Bates DW, Miller EB, Cullen DJ, et al; ADE Prevention Study Group. Patient risk factors for adverse drug events in hospitalized patients. Arch Intern Med. 1999;159(21):2553-2560. Eber MR, Laxminarayan R, Perencevich EN, Malani A. Clinical and economic outcomes attributable to health careassociated sepsis and pneumonia. Arch Intern Med. 2010; 170(4):347-353. Rivard PE, Luther SL, Christiansen CL, et al. Using patient safety indicators to estimate the impact of potential adverse events on outcomes. Med Care Res Rev. 2008;65(1):67-87. Pittet D, Thiévent B, Wenzel RP, Li N, Gurman G, Suter PM. Importance of pre-existing co-morbidities for prognosis of septicemia in critically ill patients. Intensive Care Med. 1993; 19(5):265-272. Knaus WA, Wagner DP, Draper EA, et al. The APACHE III prognostic system. Risk prediction of hospital mortality for critically ill hospitalized adults. Chest. 1991;100(6):1619-1636. Andrews LB, Stocking C, Krizek T, et al. An alternative strategy for studying adverse events in medical care. Lancet. 1997;349(9048):309-313. Leape LL, Brennan TA, Laird N, et al. The nature of adverse events in hospitalized patients. Results of the Harvard Medical Practice Study II. N Engl J Med. 1991;324(6):377-384. Original Research
35. Giraud T, Dhainaut JF, Vaxelaire JF, et al. Iatrogenic complications in adult intensive care units: a prospective two-center study. Crit Care Med. 1993;21(1):40-51. 36. Cullen DJ, Sweitzer BJ, Bates DW, Burdick E, Edmondson A, Leape LL. Preventable adverse drug events in hospitalized patients: a comparative study of intensive care and general care units. Crit Care Med. 1997;25(8):1289-1297. 37. Rothschild JM, Landrigan CP, Cronin JW, et al. The Critical Care Safety Study: The incidence and nature of adverse events and serious medical errors in intensive care. Crit Care Med. 2005;33(8):1694-1700. 38. Gaba DM, Howard SK. Patient safety: fatigue among clinicians and the safety of patients. N Engl J Med. 2002;347(16): 1249-1255. 39. Bion JF, Heffner JE. Challenges in the care of the acutely ill. Lancet. 2004;363(9413):970-977. 40. Rivers E, Nguyen B, Havstad S, et al; Early Goal-Directed Therapy Collaborative Group. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med. 2001;345(19):1368-1377. 41. McGee DC, Gould MK. Preventing complications of central venous catheterization. N Engl J Med. 2003;348(12):1123-1133. 42. Sadeghi B, Baron R, Zrelak P, et al. Cases of iatrogenic pneumothorax can be identified from ICD-9-CM coded data. Am J Med Qual. 2010;25(3):218-224.
journal.publications.chestnet.org
43. Iezzoni LI. Risk Adjustment for Measuring Health Care Outcomes. Chicago, IL: Health Administration Press; 2003. 44. Simborg DW. DRG creep: a new hospital-acquired disease. N Engl J Med. 1981;304(26):1602-1604. 45. Bahl V, Thompson MA, Kau TY, Hu HM, Campbell DA Jr. Do the AHRQ patient safety indicators flag conditions that are present at the time of hospital admission? Med Care. 2008;46(5):516-522. 46. Houchens RL, Elixhauser A, Romano PS. How often are potential patient safety events present on admission? Jt Comm J Qual Patient Saf. 2008;34(3):154-163. 47. Bates DW, Cullen DJ, Laird N, et al; ADE Prevention Study Group. Incidence of adverse drug events and potential adverse drug events. Implications for prevention. JAMA. 1995;274(1):29-34. 48. Classen DC, Resar R, Griffin F, et al. ‘Global trigger tool’ shows that adverse events in hospitals may be ten times greater than previously measured. Health Aff (Millwood). 2011;30(4):581-589. 49. Isaac T, Jha AK. Are patient safety indicators related to widely used measures of hospital quality? J Gen Intern Med. 2008; 23(9):1373-1378. 50. Romano PS, Geppert JJ, Davies S, Miller MR, Elixhauser A, McDonald KM. A national profile of patient safety in U.S. hospitals. Health Aff (Millwood). 2003;22(2):154-166.
CHEST / 142 / 3 / SEPTEMBER 2012
613
Original Research CRITICAL CARE
The Association Between Sepsis and Potential Medical Injury Among Hospitalized Patients Vincent Liu, MD; Benjamin J. Turk, BA; Norman W. Rizk, MD, FCCP; Patricia Kipnis, PhD; and Gabriel J. Escobar, MD
Background: Patient safety remains a national priority, but the role of disease-specific characteristics in safety is not well characterized. Methods: We identified potentially preventable medical injuries using patient safety indicators (PSIs) and annual data from the Nationwide Inpatient Sample between 2003 and 2007. We compared the rate of selected PSIs among patients hospitalized with and without sepsis. Among patients with sepsis, we also compared PSI rates across severity strata. Using multivariable casecontrol matching and regression analyses, we estimated the excess adverse outcomes associated with PSI events in patients with sepsis. Results: Patients hospitalized with sepsis accounted for 2% to 4% of hospital discharges; however, they accounted for 9% to 26% of all potential medical injuries. PSI rates varied considerably; among patients hospitalized for sepsis, they were lowest for accidental puncture or laceration and highest for postoperative respiratory failure. Nearly all PSI rates were higher among patients with sepsis compared with patients without sepsis. Among those with sepsis, most PSI rates increased as sepsis severity increased. Compared with matched sepsis control subjects, increased length of stay and hospital charges were associated with PSI events in sepsis cases. However, only decubitus ulcer, iatrogenic pneumothorax, and postoperative metabolic and physiologic derangement or respiratory failure were associated with excess mortality. Conclusion: Patients hospitalized for sepsis, compared with the general hospital population, were at a substantially increased risk of potential medical injury; their risk rose as disease severity increased. Future patient safety efforts may benefit from focusing on medically vulnerable populations. CHEST 2012; 142(3):606–613 Abbreviations: CVC 5 central venous catheter; DRG 5 diagnosis-related group; ICD-9-CM 5 International Classification of Diseases, 9th Edition, Clinical Modification; LOS 5 length of stay; NIS 5 Nationwide Inpatient Sample; PE 5 pulmonary embolism; PSI 5 patient safety indicator
Err is Human estimated that as many as To98,000 hospitalized patients die annually from pre1
ventable medical errors. Subsequent attention has aimed to identify and implement interventions that improve patient safety.2-5 Although efforts to reduce specific harms have been promising, large-scale
Manuscript received October 7, 2011; revision accepted February 1, 2012. Affiliations: From the Division of Research and Systems Research Initiative (Drs Liu, Kipnis, and Escobar and Mr Turk), and the Department of Management, Information, and Analysis (Dr Kipnis), Kaiser Permanente, Oakland; the Department of Critical Care (Dr Liu), Kaiser Permanente, Santa Clara; the Department of Inpatient Pediatrics (Dr Escobar), Kaiser Permanente, Walnut Creek; and the Division of Pulmonary and Critical Care Medicine (Dr Rizk), Stanford University, Stanford, CA. 606
improvements in safety have been limited.6-11 Few studies have evaluated how disease-specific characteristics or severity of illness might contribute to patient safety.1,12-14 To further investigate the role of disease severity in patient safety, we chose to describe the scope of potential medical injury among patients hospitalized Funding/Support: Dr Liu was supported by the Agency for Healthcare Research and Quality [Grant F32HS019181-01]. Correspondence to: Vincent Liu, MD, 2000 Broadway, Oakland, CA, 94612; e-mail: [email protected] © 2012 American College of Chest Physicians. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians. See online for more details. DOI: 10.1378/chest.11-2556 Original Research
with sepsis. Sepsis is common among inpatients and associated with significant morbidity, mortality, and resource use.15-17 Guidelines aimed at improving care among patients with sepsis also recommend intensive care and invasive procedures for selected patients, potentially increasing their risk of adverse events during hospitalization.15-20 Using data from the Nationwide Inpatient Sample (NIS) between 2003 and 2007, we describe the impact of sepsis on patient safety by comparing the rates of selected potentially preventable medical injuries among 33 million inpatients with and without sepsis. We also evaluate whether increased sepsis severity was associated with worsened patient safety and estimate the burden of potential medical injuries on outcomes among patients with sepsis.
Materials and Methods Data The NIS datasets contain discharge abstracts from up to 8 million hospitalizations annually and approximate a 20% sample of US acute-care community hospitals.21 Data elements available for each hospitalization include International Classification of Diseases, 9th Edition, Clinical Modification (ICD-9-CM) diagnosis and procedure codes, diagnosis-related groups (DRGs), discharge vital status, length of stay (LOS), hospital charges, and patient demographics. Sepsis Definitions We defined sepsis hospitalizations in two categories (narrow and broad) based on ICD-9-CM diagnosis coding.15-17 The narrow definition included hospitalizations for which sepsis codes were listed as the primary hospital diagnosis (ie, reason for hospitalization): 020.0 (septicemic), 038 and subtypes (septicemia), 785.52 (septic shock), 790.7 (bacteremia), 995.91 (sepsis), and 995.92 (severe sepsis). In the broad definition, we included hospitalizations for which sepsis was listed in either the primary or secondary diagnosis positions, to ensure consistency with definitions based on previous literature.15-17 We determined sepsis severity based on the presence of organ system dysfunction or failure by ICD-9-CM coding.15,16 For example, respiratory dysfunction was defined by the following secondary ICD-9-CM codes: 518.5 (acute respiratory distress syndrome after shock or trauma), 518.81 (acute respiratory failure), 518.82 (ARDS), 786.09 (respiratory insufficiency), 799.1 (respiratory arrest), or 96.7 (ventilator management). Severity was categorized by dysfunction in none, one or two, or three or more organ systems.15,16 Severity of Comorbid Illness To account for preexisting illness, we created a comorbidity index based on the risk of hospital mortality.22 The regression coefficients for predicted mortality based on Elixhauser categories were estimated using discharges in the 2007 NIS while controlling for age, sex, race/ethnicity, insurance status, DRG fixed effects, and all patient refined-DRG mortality risk subclasses.23 The coefficients were then used as relative weights to summarize a comorbidity index.24 This index was applied to all discharges journal.publications.chestnet.org
between 2003 and 2007; discharges without comorbidity data (18.0%; n 5 7,234,780) were removed. Identifying Medical Injuries The patient safety indicators (PSIs), version 4.2, were used to identify potentially preventable medical injuries. These indicators were designed to screen for potential adverse events resulting from exposure to health care.25 Each PSI is determined by select inclusion and exclusion criteria to describe a specific group of at-risk patients (denominator) and individuals who experience a potential adverse event (numerator). We evaluated the following PSIs: accidental puncture or laceration, decubitus ulcer, iatrogenic pneumothorax, and central venous catheter (CVC)-related bloodstream infection, as well as postoperative events including hemorrhage or hematoma, hip fracture, physiologic and metabolic derangement, pulmonary embolism (PE) or DVT, respiratory failure, and wound dehiscence. We excluded PSIs that overlapped with our inclusion criteria or outcomes (postoperative death, death in low-mortality DRGs, and postoperative sepsis), were extremely rare (transfusion reaction), or were specific to childbirth (obstetric vaginal trauma). Because our primary interest was evaluating patients hospitalized for sepsis, we revised PSIs that captured events following elective surgery (physiologic and metabolic derangement and respiratory failure) to assess only those who had surgery during nonelective hospital admission. Comparing Sepsis and Nonsepsis Events To evaluate the frequency of potential medical injury, we measured the annual rate of PSI events among patients with and without sepsis. We defined the PSI sepsis rate ratio as the rate of each PSI occurring among sepsis hospitalizations divided by the rate of the same PSI occurring among nonsepsis hospitalizations for both narrowly and broadly defined sepsis groups. We determined the mean PSI sepsis rate ratio and 95% CI by calculating a weighted average of annual values. PSI rates were also compared among patients with sepsis stratified by sepsis severity. For PSIs that identified specific organ system failure (eg, postoperative respiratory failure), ICD-9-CM coding in the same organ system (eg, respiratory dysfunction) was not counted toward sepsis severity or organ failure to avoid double-counting organ dysfunction as a measure of severity and outcome. We also evaluated the likelihood of PSI events among patients with and without sepsis after adjustment for hospital LOS based on two methods. First, we calculated LOS-weighted PSI sepsis rate ratios in which individual hospitalizations were divided by their hospital LOS. For example, the LOS-adjusted PSI contribution for a patient with a 10-day hospitalization was 0.1; comparatively, the PSI contribution from a 2-day hospitalization was 0.5. We also performed a bivariate logistic regression for the odds of each PSI adjusted for a diagnosis of sepsis and hospital LOS. Excess Adverse Outcomes Among Patients With Sepsis We estimated excess LOS, total charges, and mortality associated with potential adverse events among patients with a narrow sepsis diagnosis by two methods. First, we performed multivariable matching to minimize confounding resulting from unmeasured differences in clinical and hospital characteristics.22,26-30 Each PSI sepsis case was matched with up to four randomly selected sepsis control subjects from the same hospital and year; cases and control subjects were also matched by sex, race, age by decade, severity of organ dysfunction, and comorbid illness. Cases and control subjects without any comorbid illness were matched; CHEST / 142 / 3 / SEPTEMBER 2012
607
Table 1—Baseline Characteristics and Hospital Outcomes of Hospitalizations With and Without a Primary Hospital Diagnosis of Sepsis Between 2003 and 2007 Primary Hospital Diagnosis Category Characteristic or Outcome Characteristics Age, %, y ⱕ 17 18-39 40-64 ⱖ 65 Male Race/ethnicity, white Primary insurance Medicare Medicaid Private Other Comorbidity Without comorbidity Mean comorbidity indexa Hospital outcome Mortality Median length of stay, d Total charges, $
Nonsepsis (n 5 32,688,780)
Sepsis (n 5 534,416)
178,949 (0.6) 8,556,403 (26.2) 10,715,634 (32.8) 13,237,794 (40.5) 12,621,118 (38.6) 16,657,662 (51.0)
0 (0.0) 30,743 (5.8) 145,842 (27.3) 357,831 (67.0) 247,916 (46.4) 281,260 (52.6)
14,323,980 (43.9) 4,756,645 (14.6) 10,662,802 (32.7) 2,885,509 (8.8)
383,411 (71.8) 43,299 (8.1) 83,139 (15.6) 23,995 (4.5)
25,012,546 (77.4) 9.1 (9.6)
213,708 (40.3) 11.6 (10.9)
709,002 (2.2) 3 (2-5) 24,438 (38,762)
97,895 (18.3) 6 (4-11) 47,593 (72,648)
Values represent frequency (%) or median (interquartile range). Mean value when comorbidity present.
Second, we estimated excess adverse outcomes with linear regression so that estimates could be directly compared with those from matching analysis.22 Regression analyses controlled for patient age, sex, race, severity of organ dysfunction, comorbid illness index, and year; hospital fixed effects were adjusted for but not calculated. The regression analyses included all case and control subjects from each sepsis PSI risk pool. For cross-validation of estimates for mortality, we also performed conditional logistic regression using the multivariable matched sets and unmatched sets. A P value of , .05 was considered statistically significant for all analyses. Data were described as mean (SD), frequency (%), and rate (95% CI).
Results During the study period, 534,416 hospitalizations (1.6%) had a primary hospital diagnosis of sepsis (Table 1). Patients with sepsis were older and more frequently had preexisting comorbid illness (59.7%) than patients without sepsis (22.5%). Mortality was 18.3% among patients with sepsis and 2.2% among patients without sepsis (P , .01). Nearly one-half of hospitalized patients with sepsis had organ dysfunction (Table 2). Outcomes worsened as the severity of organ dysfunction increased; for example, mortality was 8.3% for patients without organ dysfunction and 53.7% for those with dysfunction in three or more systems.
a
PSI Rates those with comorbidities were matched to within a 2.5% difference in risk of death due to comorbidities. For each PSI, excess LOS was calculated as the difference between the LOS for a case and the mean LOS for matched controls. Pooled mean excess LOS was then determined using all matched sets; the paired t test was used to evaluate whether the pooled excess LOS was significantly different from zero. Similar methods were used to measure excess hospital charges and mortality. To increase matching frequency, we also performed multivariable matching without consideration of unique hospitals.
A total of 628,998 total PSI events were identified; 58,295 (9.3%) were among hospitalizations with a primary hospital diagnosis of sepsis. PSI rates varied considerably in patients with and without sepsis (Table 3); for example, among patients with sepsis, PSI rates varied from 0.9 (accidental puncture or laceration) to 302.9 (respiratory failure) per thousand at-risk discharges. The PSI sepsis rate ratio was greater than
Table 2—Characteristics and Outcomes of Patients With Sepsis by Severity Sepsis Severity (No. of Dysfunctional Organ Systems) Characteristic or Outcome No. (%) Outcomes Mortality Length of stay, d Total charges, $ Organ failure frequency Cardiovascular Hematologic Hepatic Metabolic Neurologic Renal Respiratory
0 273,427 (51.1) 22,700 (8.3) 7.5 (7.4) 31,572 (42,463) … … … … … … …
1-2 224,800 (42.1) 55,790 (24.8) 10.1 (11.5) 58,870 (83,575) 38,431 (17.1) 33,671 (15.0) 3,748 (1.7) 25,520 (11.4) 11,305 (5.0) 121,568 (54.1) 61,318 (27.3)
31 36,189 (6.8) 19,405 (53.7) 12.0 (14.4) 99,339 (125,198) 12,053 (33.3) 16,927 (46.8) 6,185 (17.1) 20,774 (57.4) 4,784 (13.2) 31,750 (87.7) 27,618 (76.3)
Overall 534,416 97,895 (18.3) 8.9 (10.0) 47,593 (72,648) 50,484 (9.5) 50,598 (9.5) 9,933 (1.9) 46,294 (8.7) 16,089 (3.0) 153,318 (28.7) 88,936 (16.6)
Severity was categorized by the number of dysfunctional organ systems as indicated by organ-specific secondary ICD-9-CM coding. Patients with sepsis include only those with a primary hospital diagnosis of sepsis. ICD-9-CM 5 International Classification of Diseases, 9th Edition, Clinical Modification. 608
Original Research
Table 3—Annual Rates of PSI Events (per 1,000 At-Risk Hospital Discharges) Stratified by Patients With and Without Sepsis and the Overall PSI Sepsis Rate Ratio Annual PSI Rate per 1,000 At-Risk Discharges Patient Safety Indicator Accidental puncture or laceration Sepsis Nonsepsis CVC-related bloodstream infection Sepsis Nonsepsis Decubitus ulcer Sepsis Nonsepsis Iatrogenic pneumothorax Sepsis Nonsepsis Postoperative hemorrhage or hematoma Sepsis Nonsepsis Postoperative hip fracture Sepsis Nonsepsis Postoperative physiologic or metabolic derangement Sepsis Nonsepsis Postoperative PE/DVT Sepsis Nonsepsis Postoperative respiratory failure Sepsis Nonsepsis Postoperative wound dehiscence Sepsis Nonsepsis
2003
2004
2005
2006
2007
P Trend
PSI Sepsis Rate Ratio (95% CI)
0.9 3.2
1.3 3.2
1.6 3.3
1.3 3.3
1.7 3.4
a
0.4 (0.4-0.5)
10.3 2.0
12.2 2.0
12.5 2.0
11.4 1.9
9.8 1.5
a
6.0 (5.7-6.4)
109.1 21.3
119.0 22.5
121.5 22.3
121.9 23.7
130.2 24.4
a
5.3 (5.2-5.4)
1.2 0.5
1.5 0.5
1.5 0.5
1.6 0.5
1.7 0.5
a
2.8 (2.5-3.2)
3.6 2.8
2.7 2.7
4.4 2.7
4.6 2.6
4.2 2.7
… …
1.5 (1.2-1.7)
1.6 0.2
1.7 0.2
1.1 0.2
1.5 0.2
1.2 0.2
… a
6.7 (6.0-7.3)
38.9 3.7
36.6 4.1
48.4 4.1
41.8 4.0
49.2 4.4
a
10.6 (9.8-11.4)
41.5 9.7
45.8 10.2
49.4 10.3
44.3 11.1
49.7 11.6
a
4.4 (4.2-4.6)
221.3 44.2
247.4 49.0
279.4 42.9
282.1 45.6
302.9 46.9
a …
5.9 (5.3-6.4)
4.1 1.8
7.1 2.1
6.6 2.3
5.2 2.3
5.6 2.2
… a
2.7 (2.3-3.0)
a
a
a
a
a
a
The PSI sepsis rate ratio is the mean ratio of annual ratios calculated as each PSI among patients with a primary hospital diagnosis of sepsis divided by the same PSI among patients without sepsis. CVC 5 central venous catheter; PE 5 pulmonary embolism; PSI 5 patient safety indicator. aP , .05 for trend.
unity for all PSIs except accidental puncture or laceration (Table 3). The remaining sepsis rate ratios varied from 1.5 for hemorrhage or hematoma to 10.6 for physiologic and metabolic derangement. Sepsis hospitalizations under the broad definition accounted for 162,872 (25.9%) of all identified PSI events. The PSI sepsis rate ratios were also greater than unity for sepsis hospitalizations under the broad definition (e-Table 1). After adjustment for hospital LOS, the results remained similar; only accidental puncture or laceration was less likely among patients with sepsis compared with patients without sepsis by logistic regression (OR, 0.4; P , .001) or LOS-weighted comparisons (e-Tables 2, 3). Most PSI rates increased as sepsis-related organ dysfunction worsened (Fig 1). The greatest rate increase across severity strata was seen for iatrogenic pneumothorax, increasing from 0.5 to 5.1 per thousand at-risk discharges across strata. Changes in PSI rates across severity strata were journal.publications.chestnet.org
not significantly different for decubitus ulcer, PE/DVT, wound dehiscence, and hip fracture. Excess Adverse Outcomes Multivariable matching of patients with sepsis within hospital facilities produced at least one matched control subject for 52% of all PSI cases; matching frequency was lowest for wound dehiscence (25%) and highest for accidental puncture or laceration (83%). Matching without consideration for hospital facilities produced high matching frequencies, ranging from 96% to 100%. Estimates of excess adverse outcomes associated with PSI events are shown in Figure 2. Significant increases in hospital LOS were present for all PSIs except hemorrhage or hematoma and hip fracture. Similarly, hospital charges were increased for all PSIs except postoperative hip fracture. Decubitus ulcer, iatrogenic pneumothorax, metabolic and physiologic CHEST / 142 / 3 / SEPTEMBER 2012
609
Figure 1. Percent change from baseline of patient safety indicator (PSI) event rates by level of sepsis severity. Sepsis hospitalizations include only those with a primary hospital diagnosis of sepsis. Filled circles and lines represent statistically significant differences by linear regression analysis; open circles and dotted lines represent nonstatistically significant differences. Abbreviated PSI descriptions include those for postoperative complications including physiologic derangement, respiratory failure, hemorrhage, PE/DVT, wound dehiscence, and hip fracture. Others include accidental puncture and CVC-infection. Accidental puncture 5 accidental puncture or laceration; CVC-infection 5 central venous catheter-related bloodstream infection; hemorrhage 5 hemorrhage or hematoma; PE 5 pulmonary embolism; physiologic derangement 5 physiologic or metabolic derangement.
The heightened vulnerability to potential injury among patients with sepsis is multifactorial and likely reflects the inherent risk associated with the disease, the underlying characteristics of patients with sepsis, as well as the adverse events resulting from treatment. Patients with sepsis have high mortality compared with other inpatients. In this study, overall sepsis mortality was 18% and was nearly 29% among patients with any organ dysfunction; these rates match those reported in prior studies.15,16 The increased age and frequency of comorbid illness among patients with sepsis also contribute to an increased risk of adverse events.13,17,31,32 Nearly one-half of patients with sepsis in this study had evidence of organ dysfunction. Current care guidelines for patients with severe sepsis recommend treatments that often require intensive care monitoring and invasive interventions.18 Adverse events, often life-threatening, occur more frequently among
derangement, and respiratory failure were associated with excess mortality (Fig 2). Patients with sepsis with respiratory failure had the highest excess mortality (25.3%); iatrogenic pneumothorax was also associated with high excess mortality (16.2%). Excess hospital mortality was not present for accidental puncture or laceration, CVC-associated bloodstream infection, hemorrhage or hematoma, hip fracture, and PE/DVT. Mortality estimates were similar by conditional logistic regression. Discussion This study demonstrated that patients with sepsis are at high risk for potential medical injuries when compared with the general hospital population. Using the PSIs, patients discharged with a primary hospital diagnosis of sepsis were nearly three times as likely to have iatrogenic pneumothorax, more than six times as likely to have CVC-related bloodstream infection or postoperative respiratory failure, and more than 10 times as likely to have postoperative physiologic or metabolic derangement than patients without sepsis. Although , 2% of hospital discharges had a primary hospital diagnosis of sepsis, they accounted for nearly 10% of all potential adverse events. Similarly, the 4% of hospitalizations with a broad diagnosis of sepsis accounted for 26% of all PSI events, highlighting the impact that specific disease conditions have on patient safety. 610
Figure 2. Estimates of excess adverse outcomes among patients with sepsis with PSI events. A, Length of stay. B, Hospital charges. C, Mortality. Patients with sepsis include only those with a primary hospital diagnosis of sepsis. For each PSI, point estimates (markers) and 95% CIs (lines) derived from three methods are displayed. Open circles and squares represent multivariable matching analysis within hospital facility and without hospital facility, respectively. Filled triangles are estimates from linear regression analysis. See Figure 1 legend for expansion of abbreviations. Original Research
patients admitted to the ICU.9,13,33-39 Although the NIS data did not allow us to specifically identify which patients underwent intensive care, we found that for most PSIs, the highest rates of potential medical injury were present among those patients with the most severe sepsis who would also likely require critical care. Invasive procedures also increase the risk of injury among patients with sepsis. For patients with severe sepsis, standard treatment algorithms rely on CVCguided therapy.18,40 Although these interventions may be associated with improved mortality,20,40 CVC insertion can result in iatrogenic injuries, including pneumothorax and bloodstream infection.41,42 We found that the rates of PSI-identified iatrogenic pneumothorax and CVC-related bloodstream infections increased as organ dysfunction worsened; for example, the rate of pneumothorax was tenfold higher among patients with the most severe organ dysfunction than those without dysfunction. Importantly, when they occurred in patients with sepsis, most PSI events were associated with increased hospital LOS and charges. However, only selected PSI events were associated with excess hospital mortality. This study has several strengths. We used a nationally representative sample of hospitalizations to compare rates of potential medical injury among sepsis and general hospital populations over 5 years, providing a broad perspective on patient safety in hospitals. The large sample size also provided sufficient power to evaluate adverse hospital outcomes associated with PSIs among patients with sepsis and resulted in robust estimates. However, this study also has notable limitations. Since this study was based on administrative data, we did not have detailed clinical variables necessary to ensure adequate risk adjustment for severity of illness between patients with sepsis as well as patients without sepsis. These data also lacked information about the timing of events within a hospitalization, thus hampering the ability to define causal relationships between PSIs and outcomes. Using the PSIs alone to identify medical injuries also has several limitations. First, they depend on accurate ICD-9-CM coding; inaccuracies can result from multiple sources and impact the screening capability of PSIs.43,44 Second, studies have suggested that some PSIs identify conditions that were “present on admission” rather than those that resulted from hospitalization.45,46 A handful of PSIs, such as decubitus ulcer, may be particularly affected. Unfortunately, in this study, coding for ICD-9-CM diagnoses present at hospital admission was not available; nonetheless, even after removing all decubitus ulcer events, patients with broadly-defined sepsis still accounted for 24% of all other identified PSI events. journal.publications.chestnet.org
Third, the PSIs screen for a subset of all potential adverse hospital events. For example, they do not evaluate injuries that result from medication errors, an important source of medical injury.47 As a result, the overall rates of injuries derived from the PSIs may substantially underestimate the true rate of adverse events.48 Finally, the PSIs were not designed to definitively identify medical injuries, since indicators may flag events that occur during the natural course of disease or treatment rather than events that result from failures in patient care. This limitation might be particularly acute in the case of sepsis, since the overwhelming inflammatory response to infection can affect nearly all organ systems. Thus, the broadly defined PSIs may lack some discrimination necessary to differentiate the sequelae of disease (eg, especially for postoperative metabolic and physiologic derangement or respiratory failure, which are common in sepsis) or the precursors of disease (eg, CVC-related infections) from actual preventable medical injuries. Thus, although PSIs may not be adequate instruments to benchmark and compare hospital quality or safety,45,49,50 they can highlight comparative differences in safety and potential deficiencies in care. In conclusion, this study found that sepsis was associated with an increased risk for potential medical injury; this risk rose as organ dysfunction increased. Sepsis hospitalizations accounted for between 9% to 26% of all identified potential adverse events. These findings suggest that severity of illness strongly influences patient safety and interventions aimed at reducing patient harm should focus especially on medically vulnerable populations. Acknowledgments Author contributions: Dr Liu had full access to the data and takes responsibility for the integrity of the data and the accuracy of the data analysis. Dr Liu: contributed to study design; data collection, analysis, and interpretation; preparation of the manuscript; and review and approval of the final manuscript. Mr Turk: contributed to study design, data analysis and interpretation, preparation of the manuscript, and review and approval of the final manuscript. Dr Rizk: contributed to study design, data interpretation, preparation of the manuscript, and review and approval of the final manuscript. Dr Kipnis: contributed to study design, data analysis and interpretation, preparation of the manuscript, and review and approval of the final manuscript. Dr Escobar: contributed to study design, data analysis and interpretation, preparation of the manuscript, and review and approval of the final manuscript. Financial/nonfinancial disclosures: The authors have reported to CHEST the following conflicts of interest: Dr Escobar received grant support from MedImmune, LLC (for studies on respiratory syncytial virus infection in infants), and Novartis AG (for a study on ICU-acquired infection). Drs Liu, Rizk, and Kipnis, and Mr Turk have reported that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article. Role of sponsors: The sponsor was not involved in the design and conduct of the study; collection, management, analysis, and CHEST / 142 / 3 / SEPTEMBER 2012
611
interpretation of the data; and preparation, review, or approval of the manuscript. Additional information: The e-Tables can be found in the “Supplemental Materials” area of the online article.
References 1. Kohn LT, Corrigan JM, Donaldson MS. To Err Is Human: Building a Safer Health System. Washington, DC: National Academies Press; 1999. 2. Agency for Healthcare Research and Quality. Medical errors and patient safety. Agency for Healthcare Research and Quality website. http://www.ahrq.gov/qual/patientsafetysix.htm. Published 2010. October 1, 2011. 3. Clancy CM. Ten years after To Err is Human. Am J Med Qual. 2009;24(6):525-528. 4. Pronovost PJ, Nolan T, Zeger S, Miller M, Rubin H. How can clinicians measure safety and quality in acute care? Lancet. 2004;363(9414):1061-1067. 5. Berwick DM, Calkins DR, McCannon CJ, Hackbarth AD. The 100,000 lives campaign: setting a goal and a deadline for improving health care quality. JAMA. 2006;295(3):324-327. 6. Pronovost P, Needham D, Berenholtz S, et al. An intervention to decrease catheter-related bloodstream infections in the ICU. N Engl J Med. 2006;355(26):2725-2732. 7. Haynes AB, Weiser TG, Berry WR, et al; Safe Surgery Saves Lives Study Group. A surgical safety checklist to reduce morbidity and mortality in a global population. N Engl J Med. 2009;360(5):491-499. 8. Bates DW, Leape LL, Cullen DJ, et al. Effect of computerized physician order entry and a team intervention on prevention of serious medication errors. JAMA. 1998;280(15): 1311-1316. 9. Landrigan CP, Rothschild JM, Cronin JW, et al. Effect of reducing interns’ work hours on serious medical errors in intensive care units. N Engl J Med. 2004;351(18):1838-1848. 10. Landrigan CP, Parry GJ, Bones CB, Hackbarth AD, Goldmann DA, Sharek PJ. Temporal trends in rates of patient harm resulting from medical care. N Engl J Med. 2010;363(22):2124-2134. 11. Rosen AK, Zhao S, Rivard P, et al. Tracking rates of patient safety indicators over time: lessons from the Veterans Administration. Med Care. 2006;44(9):850-861. 12. Pronovost PJ, Goeschel CA, Marsteller JA, Sexton JB, Pham JC, Berenholtz SM. Framework for patient safety research and improvement. Circulation. 2009;119(2):330-337. 13. Brennan TA, Leape LL, Laird NM, et al. Incidence of adverse events and negligence in hospitalized patients. Results of the Harvard Medical Practice Study I. N Engl J Med. 1991;324(6):370-376. 14. Naessens J, Campbell CR, Shah N, et al. Effect of illness severity and comorbidity on patient safety and adverse events. Am J Med Qual. 2012;27(1):48-56. 15. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR. Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med. 2001;29(7):1303-1310. 16. Martin GS, Mannino DM, Eaton S, Moss M. The epidemiology of sepsis in the United States from 1979 through 2000. N Engl J Med. 2003;348(16):1546-1554. 17. Dombrovskiy VY, Martin AA, Sunderram J, Paz HL. Rapid increase in hospitalization and mortality rates for severe sepsis in the United States: a trend analysis from 1993 to 2003. Crit Care Med. 2007;35(5):1244-1250. 18. Dellinger RP, Levy MM, Carlet JM, et al; International Surviving Sepsis Campaign Guidelines Committee; American 612
19.
20.
21.
22. 23. 24. 25.
26.
27. 28. 29.
30. 31.
32. 33. 34.
Association of Critical-Care Nurses; American College of Chest Physicians; American College of Emergency Physicians; Canadian Critical Care Society; European Society of Clinical Microbiology and Infectious Diseases; European Society of Intensive Care Medicine; European Respiratory Society; International Sepsis Forum; Japanese Association for Acute Medicine; Japanese Society of Intensive Care Medicine; Society of Critical Care Medicine; Society of Hospital Medicine; Surgical Infection Society; World Federation of Societies of Intensive and Critical Care Medicine. Surviving Sepsis Campaign: international guidelines for management of severe sepsis and septic shock: 2008. Crit Care Med. 2008;36(1):296-327. Ferrer R, Artigas A, Levy MM, et al; Edusepsis Study Group. Improvement in process of care and outcome after a multicenter severe sepsis educational program in Spain. JAMA. 2008;299(19):2294-2303. Levy MM, Dellinger RP, Townsend SR, et al. The Surviving Sepsis Campaign: results of an international guideline-based performance improvement program targeting severe sepsis. Crit Care Med. 2010;38(2):367-374. Introduction to the HCUP Nationwide Inpatient Sample (NIS). Healthcare Cost and Utilization Project website. http://www.hcup-us.ahrq.gov/db/nation/nis/NIS_Introduction_ 2007.jsp. Published 2007. Accessed November 16, 2010. Zhan C, Miller MR. Excess length of stay, charges, and mortality attributable to medical injuries during hospitalization. JAMA. 2003;290(14):1868-1874. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998; 36(1):8-27. Romano PS, Chan BK. Risk-adjusting acute myocardial infarction mortality: are APR-DRGs the right tool? Health Serv Res. 2000;34(7):1469-1489. Patient Safety Indicators Overview. AHRQ quality indicators. Agency for Healthcare Research and Quality website http:// www.qualityindicators.ahrq.gov/modules/psi_overview.aspx. Published February 2006. Accessed November, 16, 2010. Classen DC, Pestotnik SL, Evans RS, Lloyd JF, Burke JP. Adverse drug events in hospitalized patients. Excess length of stay, extra costs, and attributable mortality. JAMA. 1997; 277(4):301-306. Bates DW, Spell N, Cullen DJ, et al; Adverse Drug Events Prevention Study Group. The costs of adverse drug events in hospitalized patients. JAMA. 1997;277(4):307-311. Bates DW, Miller EB, Cullen DJ, et al; ADE Prevention Study Group. Patient risk factors for adverse drug events in hospitalized patients. Arch Intern Med. 1999;159(21):2553-2560. Eber MR, Laxminarayan R, Perencevich EN, Malani A. Clinical and economic outcomes attributable to health careassociated sepsis and pneumonia. Arch Intern Med. 2010; 170(4):347-353. Rivard PE, Luther SL, Christiansen CL, et al. Using patient safety indicators to estimate the impact of potential adverse events on outcomes. Med Care Res Rev. 2008;65(1):67-87. Pittet D, Thiévent B, Wenzel RP, Li N, Gurman G, Suter PM. Importance of pre-existing co-morbidities for prognosis of septicemia in critically ill patients. Intensive Care Med. 1993; 19(5):265-272. Knaus WA, Wagner DP, Draper EA, et al. The APACHE III prognostic system. Risk prediction of hospital mortality for critically ill hospitalized adults. Chest. 1991;100(6):1619-1636. Andrews LB, Stocking C, Krizek T, et al. An alternative strategy for studying adverse events in medical care. Lancet. 1997;349(9048):309-313. Leape LL, Brennan TA, Laird N, et al. The nature of adverse events in hospitalized patients. Results of the Harvard Medical Practice Study II. N Engl J Med. 1991;324(6):377-384. Original Research
35. Giraud T, Dhainaut JF, Vaxelaire JF, et al. Iatrogenic complications in adult intensive care units: a prospective two-center study. Crit Care Med. 1993;21(1):40-51. 36. Cullen DJ, Sweitzer BJ, Bates DW, Burdick E, Edmondson A, Leape LL. Preventable adverse drug events in hospitalized patients: a comparative study of intensive care and general care units. Crit Care Med. 1997;25(8):1289-1297. 37. Rothschild JM, Landrigan CP, Cronin JW, et al. The Critical Care Safety Study: The incidence and nature of adverse events and serious medical errors in intensive care. Crit Care Med. 2005;33(8):1694-1700. 38. Gaba DM, Howard SK. Patient safety: fatigue among clinicians and the safety of patients. N Engl J Med. 2002;347(16): 1249-1255. 39. Bion JF, Heffner JE. Challenges in the care of the acutely ill. Lancet. 2004;363(9413):970-977. 40. Rivers E, Nguyen B, Havstad S, et al; Early Goal-Directed Therapy Collaborative Group. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med. 2001;345(19):1368-1377. 41. McGee DC, Gould MK. Preventing complications of central venous catheterization. N Engl J Med. 2003;348(12):1123-1133. 42. Sadeghi B, Baron R, Zrelak P, et al. Cases of iatrogenic pneumothorax can be identified from ICD-9-CM coded data. Am J Med Qual. 2010;25(3):218-224.
journal.publications.chestnet.org
43. Iezzoni LI. Risk Adjustment for Measuring Health Care Outcomes. Chicago, IL: Health Administration Press; 2003. 44. Simborg DW. DRG creep: a new hospital-acquired disease. N Engl J Med. 1981;304(26):1602-1604. 45. Bahl V, Thompson MA, Kau TY, Hu HM, Campbell DA Jr. Do the AHRQ patient safety indicators flag conditions that are present at the time of hospital admission? Med Care. 2008;46(5):516-522. 46. Houchens RL, Elixhauser A, Romano PS. How often are potential patient safety events present on admission? Jt Comm J Qual Patient Saf. 2008;34(3):154-163. 47. Bates DW, Cullen DJ, Laird N, et al; ADE Prevention Study Group. Incidence of adverse drug events and potential adverse drug events. Implications for prevention. JAMA. 1995;274(1):29-34. 48. Classen DC, Resar R, Griffin F, et al. ‘Global trigger tool’ shows that adverse events in hospitals may be ten times greater than previously measured. Health Aff (Millwood). 2011;30(4):581-589. 49. Isaac T, Jha AK. Are patient safety indicators related to widely used measures of hospital quality? J Gen Intern Med. 2008; 23(9):1373-1378. 50. Romano PS, Geppert JJ, Davies S, Miller MR, Elixhauser A, McDonald KM. A national profile of patient safety in U.S. hospitals. Health Aff (Millwood). 2003;22(2):154-166.
CHEST / 142 / 3 / SEPTEMBER 2012
613