The epidemiologic characteristics, temporal trends, predictors of death, and discharge disposition in patients with a diagnosis of sepsis: A cross-sectional retrospective cohort study

The epidemiologic characteristics, temporal trends, predictors of death, and discharge disposition in patients with a diagnosis of sepsis: A cross-sectional retrospective cohort study

Journal of Critical Care 39 (2017) 48–55 Contents lists available at ScienceDirect Journal of Critical Care journal homepage: www.jccjournal.org Th...

259KB Sizes 0 Downloads 4 Views

Journal of Critical Care 39 (2017) 48–55

Contents lists available at ScienceDirect

Journal of Critical Care journal homepage: www.jccjournal.org

The epidemiologic characteristics, temporal trends, predictors of death, and discharge disposition in patients with a diagnosis of sepsis: A cross-sectional retrospective cohort study Sarah Elfeky, MD a, Pegah Golabi, MD b, Munkhzul Otgonsuren, MPH b, Svetolik Djurkovic, MD a, Mary E. Schmidt, MD, MPH a, Zobair M. Younossi, MD, MPH a,b,⁎ a b

Department of Medicine, Inova Fairfax Medical Campus, Falls Church, VA Betty and Guy Beatty Center for Integrated Research, Inova Health System, Falls Church, VA

a r t i c l e

i n f o

Available online xxxx Keywords: Sepsis Disposition Mortality Insurance

a b s t r a c t Purpose: To assess recent epidemiologic characteristics, temporal trends, and predictors of death and discharge disposition in patients with sepsis. Material and methods: This is a cross-sectional retrospective cohort study using the US National Inpatient Sample (NIS) data from 2009 to 2012. The study population included adults (18 years and older) with sepsis-related International Classification of Diseases, Ninth Revision, Clinical Modification codes at the time of discharge. Factors associated with in-hospital mortality and patient discharge disposition were derived from multivariate analyses using multinomial logistic models by SAS PROC LOGISTIC with GLOGIT link. Results: Of 1 303 640 patients admitted, 15% died, 30% were discharged to home without home care, 34% were transferred to a skilled outpatient facility, and 4% were transferred to another short-term hospital. In-hospital mortality decreased from 16.5% to 13.8% (P b .001) across time. Length of stay also decreased from 6.7 to 5.9 days (P b .001). Reductions in mortality and length of stay were seen despite an increase in the number of comorbidities (P b .001). Multivariate analysis revealed that the strongest predictors of in-hospital mortality were respiratory, cardiovascular, and hepatic failures, and neurologic events. The predictors of transfer to an outpatient facility were a major operative procedure, neurologic event, respiratory failure, and weight loss. Weight loss was also an independent predictor of in-hospital mortality. Conclusion: Certain comorbidities and organ failures were associated with death and discharge to a skilled outpatient facility. © 2017 Elsevier Inc. All rights reserved.

1. Introduction Sepsis is an increasingly important contributor to hospitalizations, in-hospital deaths, and transfer to a short-term acute care hospital or a skilled outpatient health care facility. Although the implementation of specific guidelines to assist with earlier recognition of and treatment for patients with severe sepsis and septic shock has been shown to decrease morbidity and mortality, both still remain high [1-4]. Other factors associated with morbidity and mortality need to be identified so that targeted interventions can be implemented and studied. Hence, this reinforces the need for continuous monitoring of trends as done in this study.

⁎ Corresponding author at: Betty and Guy Beatty Center for Integrated Research, Claude Moore Health Education and Research Building, 3300 Gallows Rd, Falls Church, VA 22042. Tel.: +1 703 776 2540; fax: +1 703 776 4386. E-mail address: [email protected] (Z.M. Younossi).

http://dx.doi.org/10.1016/j.jcrc.2017.01.006 0883-9441/© 2017 Elsevier Inc. All rights reserved.

Sepsis, defined as a syndrome of physiologic, pathologic, and biochemical abnormalities induced by infection, accounted for more than $20 billion (5.2%) of total US hospital costs in 2011 [5]. In fact, sepsis consumes almost half of intensive care unit (ICU) resources. Multiple reports have suggested that the incidence of septicemia, sepsis, and severe sepsis has been increasing steadily for the past several decades. The Centers for Disease Control and Prevention's National Center for Health Statistics estimates that the number of hospitalization for sepsis increased from 621 000 in 2000 to 1 141 000 in 2008 [6]. Septicemia was the most expensive condition in 2009 and its costs grew fastest between 1997 and 2009 [7]. Furthermore, there is increasing awareness that patients who survive sepsis often have long-term physical, psychological, and cognitive disabilities with significant health care and social implications [8-11]. The reported incidence of sepsis is rising, likely reflecting an aging population with more comorbidities, better recognition of sepsis, and the advent of reimbursement-favorable coding [12]. Commonly cited explanations for this emerging trend include increasing use of

S. Elfeky et al. / Journal of Critical Care 39 (2017) 48–55

immunosuppression, invasive procedures, and the spread of multidrugresistant pathogens. Most existing studies on temporal trends, however, are based on analyses of administrative data. It is therefore possible that some of the observed increase in incidence is due to changes in diagnosis and coding practices rather than true increases in disease frequency. The treatment of sepsis involves caring for sicker patients who have longer inpatient stays than those with other diagnoses. Total nationwide inpatient annual costs of treating those hospitalized for septicemia have been rising and were estimated to be $14.6 billion in 2008 [13]. Patients who do survive severe cases are more likely to have negative long-term effects on health and on cognitive and physical functioning [8-11,14]. Only 2% of hospitalizations in 2008 were for septicemia or sepsis, yet accounted for 17% of in-hospital deaths [15]. In-hospital deaths were more than 8 times as likely among patients hospitalized for septicemia or sepsis (17%) compared with other diagnoses (2%). Therefore, understanding the factors that impact disposition of patients with sepsis is important from a societal and economic standpoint. Many studies suggest that acute infections worsen preexisting chronic diseases or result in new chronic diseases, leading to poor long-term outcomes in acute illness survivors [15]. Almost half of sepsis survivors are discharged to skilled nursing facilities. In addition, those hospitalized for septicemia or sepsis were one-half as likely to be discharged home, twice as likely to be transferred to a short-term acute care hospital, and 3 times as likely to be discharged to a skilled outpatient facility, as those with other diagnoses [16]. The goal of this study was to assess epidemiologic characteristics, factors associated with disposition and trends in patients with sepsis who died or were discharged to a short-term acute care hospitals or a skilled outpatient health care facility vs routine discharge to home without home care services. Understanding risk factors associated with mortality, poor functional outcomes, and increased post-discharge care is important for providers to know. This knowledge can improve patients' outcomes and minimize odds of morbidity or mortality if extrapolated conclusions from this data are implemented. Please refer to the discussion section for further elaboration. Also, disclosure of this information to the patient or family can help with post-discharge planning and with decision making during goals of care meetings. In addition, this is important information in regard to health care planning and allocation of governmental, social, and hospital resources. In a time where we are seeking health care reform and where post sepsis syndrome is increasingly recognized, it is imperative to continuously assess such trends to see if there are changes with respect to cost, morbidity, and mortality (eg, insured and noninsured patients)

2. Methods 2.1. Study design and population A series cross-sectional study, the National Inpatient Sample (NIS) contains data on community “non-Federal short-stay” hospitalizations, not patient-level records from states, participating approximately 40 states in the Healthcare Cost and Utilization Project. The NIS is a stratified probability sampling frame of 20% of discharges from community hospitals that represent approximately 95% of the US population (see full information on study design, procedures, and quality control on reference) [17,18]. Patients with sepsis were identified by presence of any listed (up to 25 diagnosis codes per episode) the International Classification of Diseases, Ninth Revision (ICD-9) diagnosis codes as 038 (septicemia), 020.0 (septicemic), 790.7 (bacteremia), 117.9 (disseminated fungal infection), 112.5 (disseminated candida infection), and 112.81 (disseminated fungal endocarditis) in 2009 and 2012. We excluded patients 17 years or younger as well patients with missing information on key variables (discharge status, the total charge, age, sex, length of stay [LOS], demographic variables, approximately 3% of the sepsis cohort).

49

2.2. Outcomes Patients' dispositions were categorized into the following: (1) routine discharge to home without home health services, (2) transfer out to another short-term acute care hospital, (3) transfer to a skilled outpatient health care facility (eg, skilled nursing facility, intermediate care facility, long-term care rehabilitation facility, (4) home health care (includes patients discharged home with intravenous antibiotics and those discharged home with hospice services), (5) against medical advice and (6) died in-hospital. Invalid and unknown destinations were excluded from the analytical cohort. 2.3. Characteristics of study Available variables included patients' demographic information: age; sex; race was recoded as white, black, other (included Hispanic, Asian, and other race), and “not available” for accounting missing (15% in 2009, 11% in 2010, 10% in 2011 and 5% in 2012); elective admission or not; primary payer (Medicare, Medicaid, private including health maintenance organization, uninsured, no charge, other); median income in patient”'s zip code (categorized into quartiles relative to the nationwide distribution for each year); hospital size (estimated using the number of beds, categorized by size as small, medium, large); location/teaching (rural, urban nonteaching, urban teaching); and region (Northeast, Midwest, South, or West). The total number of hospital charges, procedures, diagnoses, and LOS (days) were also included with each discharge. Charges were adjusted for 2012 US dollars by Consumer Price Index [19]. The comorbidities were derived from the coexisting medical conditions that are not related to the principal diagnosis in NIS and were defined by the Agency for Healthcare Research and Quality comorbidity measures [20]. These coexisting medical conditions were based on International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes and include those diagnoses that preceded the hospitalization. Potential comorbidities were grouped as the following: (1)acquired immune deficiency; (2) substance abuse (alcohol abuse, drug abuse); (3)rheumatoid arthritis/collagen vascular diseases; (4)congestive heart failure; (5)hypertension, uncomplicated and complicated; (6)peripheral vascular disorders, disease; (7)pulmonary diseases (chronic pulmonary disease, pulmonary circulation disorders); (8)depression/psychoses; (9)diabetes uncomplicated/with chronic complications; (10)liver disease; (11)renal disorders (fluid and electrolyte disorders, renal failure); (12)cancer (metastatic cancer, lymphoma, solid tumor without metastasis); (13)paralysis/other neurologic disorders were also evaluated; (14)obesity; and (15)weight loss and reported for each discharge together with the severity of illness for the patient measured according to the All Patient Refined Diagnosis Related Groups (classified into minor loss of function, including cases with no comorbidity or complications; moderate loss of function; major loss of function; extreme loss of function). For the purpose of the study, the 7 following organ failures were identified by the ICD-9-CM codes (Table 1) An indication of any organ failure was created in the presence of at least 1 of the 7 organ failures. 2.4. Data analysis Simple descriptive statistics were examined by calendar year and by disposition status. To analyze patient disposition, we used a multivariate multinomial logistic model with routine discharge to home (without home health services) as the reference, and demographic, socioeconomic, hospital, and clinical characteristics as the independent variables. We used backward model selection (P = .10) to select a robust set of independent variables for the final model (Table 5). The natural logarithm transformed the following variables: the total charge, LOS, and number of procedures in the models to highly right skewed variation. A 2-sided P value less than .05 was considered statistically

50

S. Elfeky et al. / Journal of Critical Care 39 (2017) 48–55

Table 1 ICD-9-CM organ failure codes used in the study. Respiratory codes Cardiovascular codes

518.81 acute respiratory failure, 518.82 other pulmonary insufficiency, 518.85, 786.09, 799.1, 96.7 458.0 hypotension, postural, 785.5 shock, 785.51 shock, cardiogenic, 785.59 shock, circulatory or septic, 458.8 hypotension, specified type, not elsewhere classified, 458.9 hypotension, arterial, constitutional, 796.3 hypotension, transient; 584 acute renal failure, 580 acute glomerulonephritis, 585 renal shutdown unspecified, 39.95 hemodialysis; 570 acute hepatic failure or necrosis, 572.2 hepatic encephalopathy, 573.3 hepatitis, septic or unspecified Hematologic codes: 286.2 disseminated intravascular coagulation, 286.6 purpura fulminans, 286.9 coagulopathy, 287.3-5 thrombocytopenia, primary, secondary or unspecified; (6) metabolic codes: 276.2 acidosis, metabolic or lactic 276.2 acidosis, metabolic or lactic 293 transient organic psychosis, 348.1 anoxic brain injury, 348.3 encephalopathy, acute, 780.01 coma; 780.09 altered consciousness, unspecified, 89.14 electroencephalography

Renal codes Hepatic codes Hematologic codes Metabolic codes Neurologic codes

significant. Statistical analyses were performed using SAS, version 9.3 (SAS Institute, Cary, NC).

2.6. Institutional review board approval The study was considered exempt by the Inova Institutional Review Board.

2.5. Role of the funding source The study was funded by internal funds only. There were no external funds or sponsor for this project. Therefore, there were no influences of the funding source regarding the design, conduct, or reporting of the study.

3. Results Patient characteristics by year are reported on Table 2. For 4 years (2009-2012), there were 1 303 640 hospitalizations with sepsis

Table 2 Patient characteristics by calendar year in patients with sepsis, NIS, 2009-2012 Variables

Total

2009

2010

2011

2012

Total no. of sepsis admissions Crude rate per 100 000 all hospital discharges Disposition status, Routine Short-term another hospital Skilled outpatient health care facility Home health care Against medical advice Died in-hospital Total charge ($)

1 303 640 4302

299 992 3840

315 617 4046

354 066 4413

333 965 4577

396 969 (30.40%) 52 073 (4.00%) 438 568 (33.66%)

86 937 (29.65%) 11 821 (4.05%) 97 550 (33.60%)

91 442 (30.02%) 11 839 (3.84%) 102 465 (33.62%)

105 047 (30.63%) 12 800 (3.78%) 117 707 (34.18%)

109 229 (32.71%) 12 654 (3.79%) 109 163 (32.69%)

209 026 (16.03%) 10 165 (0.78%) 196 839 (15.12%) $46 157.35 (22 421.3-101 998.9) 6.40 (3.4-12.4) 67.53 (54.3-79.5) 657 384 (50.41%)

45 107 (15.48%) 2206 (0.76%) 47 932 (16.46%) $45 605.76 (21 860.2-102 640.1) 6.73 (3.5-13.0) 67.72 (54.2-79.6) 147 443 (50.56%)

48 916 (16.17%) 2345 (0.77%) 47 281 (15.57%) $46 110.12 (22 434.1-102 728.8) 6.51 (3.4-12.4) 67.52 (54.2-79.6) 152 890 (50.24%)

55 723 (16.11%) 2594 (0.76%) 49 955 (14.53%) $45 620.76 (22 493.1-98 905.7) 6.13 (3.3-11.6) 67.54 (54.3-79.6) 174 268 (50.69%)

53 915 (16.14%) 2859 (0.86%) 46 145 (13.82%) $44 164.68 (21 950.0-96 054.5) 5.88 (3.1-11.1) 67.17 (54.1-79.3) 167 603 (50.19%)

820 302 (62.90%) 187 212 (14.47%) 178 986 (13.75%) 117 140 (8.88%)

173 514 (59.65%) 37 492 (12.86%) 39 992 (13.74%) 40 555 (13.75%)

189 576 (62.25%) 45 016 (15.09%) 39 910 (13.01%) 29 786 (9.64%)

217 556 (63.20%) 51 052 (15.07%) 45 441 (13.30%) 29 777 (8.43%)

222 695 (66.68%) 47 306 (14.16%) 48 676 (14.58%) 15 288 (4.58%)

$391 286 (30.93%) $323 319 (25.40%) $303 621 (23.78%) $253 228 (19.89%)

$85 203 (30.02%) $74 740 (26.16%) $66 806 (23.51%) $56 861 (20.31%)

$89 147 (30.43%) $76 603 (25.76%) $69 505 (23.44%) $60 415 (20.37%)

$100 631 (30.68%) $83 190 (24.75%) $84 945 (24.84%) $67 958 (19.73%)

$103 804 (31.83%) $81 411 (24.96%) $76 493 (23.46%) $64 410 (19.75%)

149 762 (11.68%) 550 531 (42.69%) 588 178 (45.63%)

34 498 (11.78%) 122 544 (44.21%) 128 777 (44.01%)

36 872 (11.92%) 128 567 (42.77%) 135 221 (45.30%)

35 180 (11.77%) 144 878 (42.11%) 157 961 (46.13%)

38 255 (11.45%) 130 093 (38.95%) 165 617 (49.59%)

147 786 (11.53%) 944 521 (73.06%) 196 164 (15.41%)

32 986 (11.43%) 209 116 (73.35%) 43 717 (15.23%)

39 424 (13.48%) 22,3031 (73.52%) 38 205 (13.00%)

31 907 (9.77%) 256 877 (75.30%) 49 235 (14.92%)

40 572 (12.15%) 248 304 (74.35%) 45 089 (13.50%)

167 734 (12.74%) 313 444 (24.51%) 807 293 (62.76%)

32 587 (11.04%) 69 775 (24.74%) 183 457 (64.22%)

37 276 (11.34%) 70 020 (22.81%) 193 364 (65.85%)

37 034 (10.61%) 83 494 (25.21%) 217 491 (64.18%)

43 011 (12.88%) 87 711 (26.26%) 203 243 (60.86%)

813 954 (64.23%) 141 801 (11.34%) 230 702 (18.20%) 53 689 (4.23%)

185 190 (65.23%) 31 745 (11.13%) 55 969 (19.60%) 11 589 (4.04%)

192 705 (64.96%) 35 124 (11.84%) 56 049 (18.95%) 12 421 (4.24%)

221 187 (66.05%) 37 572 (11.29%) 61 636 (18.33%) 14 451 (4.34%)

214 872 (66.21%) 37 360 (11.51%) 57 048 (17.58%) 15 228 (4.69%)

LOS (d) Age (y) Female Race White Black Other Missing Median household income $0-24 999 $25 000-34 999 $35 000-44 999 $45 000+ Hospital location/teaching status Rural Urban nonteaching Urban teaching Hospital control/ownership Government, nonfederal Private, nonprofit Private, invest-own Hospital bed-size Small Medium Large Primary payment Medicare Medicaid Private including HMO Uninsured/no charge

P

b.0001

.4713 b.0001 .1338 .1454 b.0001

.8803

.4988

.4723

.3598

.0264

Data were represented as median (interquartile range) for numeric variables and frequency (proportion) for categorical variables. P values were reported by linear trend for numerical variables and χ2 test for categorical variables.

S. Elfeky et al. / Journal of Critical Care 39 (2017) 48–55

diagnosis. The annual number of hospitalizations with sepsis increased from 299 992 in 2009 to 333 965 in 2012, amounting to an 11% increase across the time span of the study period, or 3800 to 4600 per 100 000 admissions. This is interesting because the observed 11% increase in total sepsis admissions does not seem to result from an increase in the number of total admissions throughout this period. Appendix Table 1 shows that there was an overall decline in all hospitalizations from 2009 to 2012 by about 7.2%. One may argue that there was an increasing trend from 2009 to 2011, but this was only a modest 2.9%. The increase in sepsis hospitalizations was quite consistent throughout this period and occurred among both men and women. Age, sex, and race remained stable over the study period. Forty-six percent of patients were discharged from an urban teaching hospital and more than 30% of patients were from a lower-income bracket in all years studied. Sixtythree percent of patients were white which was less than that in comparison to 74.4% in the 2010 US census [21] Whites are also underrepresented in all patients who have been hospitalized during this period, consisting of 60.3%. There was no significant change in average inflation-adjusted charge during the study period (from $45 605 to $44 164; P = .47). The pattern of discharge status only slightly changed. For example, Medicare was the predominant payer covering 65% of patients in 2009 and 66% of patients in 2014. In the NIS, a total of 26 069 030 adult (18 years or older) patients were admitted to approximately 1000 hospitals between 2009 and 2012 in the United States. We excluded 625 738 admissions with missing information on key variables (discharge status, the total charge, LOS, age/sex) resulting in 25 443 292 admissions. Temporal trends in discharge status, hospital utilizations, and demographic variables are shown in Appendix Table 1. The trend of sepsis hospitalization in this period seems to contrast that of all hospitalizations in some aspects, although some parallels exist. For example, there was a 2.9% annual decrease in in-hospital mortality among patients hospitalized for sepsis, whereas there was an

51

annual 0.76% decrease in in-hospital mortality in all patients. Interestingly, there was a 3.4% annual increase in sepsis hospitalizations resulting in routine discharge to home, whereas there was a steady annual decline of 0.82% among all hospitalizations. There was an annual 2.1% decrease in disposition of sepsis hospitalizations to another acute care facility, which was similar to the observed annual 2.2% decrease in disposition of all hospitalizations to another acute care facility. All hospitalizations observed a steady annual increase in discharges to skilled outpatient facilities (1.8%) and to home with home health services (3.1%). Hospitalizations for sepsis also observed an increase in discharges to home with home health services (1.4%) but had a steady annual decrease in discharges to a skilled outpatient health care facility (0.91%). Of note, the absolute number of patients with sepsis requiring transfer to a skilled outpatient has been rising from 97 550 (33.60%) in 2009 to 109 163 (32.71%) in 2012. Comorbidities and organ failures associated with the diagnosis of sepsis are reported in Table 3. Comorbidities that were most commonly associated with sepsis were hypertension (54.2%), renal disorders (65.3%), diabetes (32.4%), pulmonary diseases (27.4%), and neurologic disorders (19.2%). The distribution of comorbidities was similar in 2009 and 2012, but the proportion of patients admitted for sepsis with hypertension (from 51.49% in 2009 to 57.15% in 2012) and obesity (from 9.22% in 2009 to 13.07% in 2012) increased. Mortality rates for the entire cohort declined over the 4-year period, averaging from 16.5% in 2009 to 13.8% in 2012 (P b .001) as well as the hospital LOS (6.75.9 days [P b .001]). The reductions in mortality and LOS were seen despite an increase in the number of chronic conditions (P b .0001) and proportion of patients with any organ failure (P = .0005). Renal, cardiovascular, and respiratory failures were the most likely organ system failures throughout the study period. Almost 70% of patients had one organ failure. The organs that failed most frequently were renal (49%), cardiovascular (28%), and respiratory (24%). Less frequent was hepatic failure (4%). The incidence of respiratory, cardiovascular, renal, and hepatic

Table 3 Comorbidities, organ failures, and procedures associated with sepsis, NIS, 2009-2012 Variables

Total

2009

2010

2011

2012

P

No. of chronic conditions Comorbidity AIDS Substance abuse Rheumatoid arthritis/collagen vascular disease CHF Peripheral vascular disorders Hypertension Pulmonary diseases Depression Diabetes Liver disease Renal disorders Cancer Paralysis/other neurologic disorders Obesity Weight loss Severity of illness Minor/moderate loss of function Major/extreme loss of function Any organ failure, Organ failure status Respiratory Cardiovascular Renal Hepatic Hematologic Metabolic Neurologic No. of procedures Major operating room procedure

5.45 (3.3-7.8)

4.96 (3.0-7.2)

5.13 (3.1-7.5)

5.71 (3.6-8.0)

5.90 (3.7-8.2)

b.0001

3620 (0.28%) 86 422 (6.63%) 46 162 (3.53%) 265 434 (20.53%) 108 908 (8.50%) 693 434 (54.20%) 358 258 (27.43%) 198 998 (15.23%) 423 467 (32.43%) 69 060 (5.29%) 852 730 (65.34%) 138 985 (10.66%) 251 135 (19.25%) 139 696 (10.86%) 218 954 (17.13%)

966 (0.33%) 17 165 (5.86%) 9678 (3.31%) 59 213 (20.34%) 23 753 (8.15%) 150 193 (51.49%) 76 457 (26.23%) 39 433 (13.53%) 90 766 (31.09%) 14 161 (4.84%) 185 664 (63.65%) 31 453 (10.81%) 53 805 (18.47%) 27 019 (9.22%) 47 805 (16.35%)

997 (0.33%) 19 204 (6.33%) 10 087 (3.32%) 61 960 (20.34%) 24 280 (8.00%) 159 539 (52.40%) 80 213 (26.33%) 43 518 (14.28%) 94 281 (30.93%) 15 568 (5.13%) 195 607 (64.30%) 32 692 (10.75%) 57 229 (18.76%) 28 538 (9.38%) 49 608 (16.32%)

877 (0.25%) 23 529 (6.87%) 12 725 (3.69%) 74 540 (21.65%) 31 082 (9.02%) 192 844 (55.99%) 96 893 (28.14%) 55 629 (16.11%) 114 343 (33.29%) 18 973 (5.50%) 230 739 (67.04%) 36 815 (10.67%) 67 796 (19.74%) 40 489 (11.75%) 63 852 (18.60%)

731 (0.22%) 25 257 (7.56%) 12 771 (3.82%) 69 721 (20.88%) 29 793 (8.92%) 190 858 (57.15%) 96 017 (28.75%) 55 939 (16.75%) 113 423 (33.96%) 19 327 (5.79%) 223 589 (66.95%) 35 646 (10.67%) 65 245 (19.54%) 43 650 (13.07%) 57 689 (17.27%)

.003 b.0001 b.0001 .0011 b.0001 b.0001 b.0001 b.0001 b.0001 b.0001 b.0001 .9469 0.0001 b.0001 .0003 b.0001

166 348 (12.78%) 1 108 679 (%) 909 279 (69.70%)

33 585 (11.53%) 257 960 (88.47%) 200 408 (68.68%)

35 645 (11.67%) 268 624 (88.33%) 211 746 (69.64%)

39 904 (11.67%) 303 766 (88.33%) 242 618 (70.47%)

55 424 (16.61%) 278 329 (83.39%) 233 573 (69.94%)

.0005

307 807 (23.60%) 361 282 (27.67%) 640 302 (49.07%) 52 088 (3.99%) 168 523 (12.88%) 180 599 (13.81%) 180 295 (13.82%) 1.30 (0.0-3.9) 249 448 (19.14%)

68 932 (23.60%) 79 985 (27.38%) 140 709 (48.21%) 11 259 (3.85%) 34 394 (11.76%) 37 135 (12.70%) 34 711 (11.88%) 1.43 (0.0-4.0) 58 470 (20.02%)

72 404 (23.83%) 85 849 (28.24%) 150 021 (49.34%) 12 152 (4.00%) 38 419 (12.64%) 40 352 (13.27%) 39 468 (13.01%) 1.40 (0.0-4.1) 59 698 (19.73%)

81 108 (23.55%) 98 017 (28.39%) 172 005 (49.93%) 14 130 (4.09%) 47 086 (13.58%) 49 665 (14.36%) 49 404 (14.40%) 1.24 (0.0-3.8) 65 555 (18.95%)

77 502 (23.21%) 92 771 (27.78%) 163 542 (48.97%) 13 814 (4.14%) 45 893 (13.74%) 51 303 (15.36%) 50 361 (15.08%) 1.13 (0.0-3.7) 61 385 (18.38%)

.3455 .1976 .0069 .1012 b.0001 b.0001 b.0001 .0023 .0002

Data were represented as median (interquartile range) for numeric variables and frequency (proportion) for categorical variables. CHF indicates chronic heart failure. P values were reported by linear trend for numerical variables and χ2 test for categorical variables.

52

S. Elfeky et al. / Journal of Critical Care 39 (2017) 48–55

Table 4 Characteristics of study by disposition status, NIS, 2009-2012 Variables

Routine

Total no. of 396 969 (30.45%) sepsis discharges (%) Total charge ($) $30 806.03 (16 634.1-60 646.1) LOS (d) 4.53 (2.7-7.9) Age (y) 58.62 (45.8-71.2) Female 198 383 (49.95%) Race White 239 953 (60.37%) Black 55 484 (14.07%) Other 66 681 (16.86%) Missing 34 851 (8.70%) Median household income $0-24 999 $120 417 (31.33%) $25 000-34 $99 613 (25.75%) 999 $35 000-44 $91 352 (23.54%) 999 $45 000+ $75 070 (19.37%)

Short-term another hospital

Another health care facility including a long-term care

Home health care

Against medical advice

Died in hospital

52 073 (3.99%)

438 568 (33.64%)

209 026 (16.03%)

10 165 (0.78%)

196 839 (15.10%)

$41 001.54 (18 125.0-98 581.1) 4.94 (1.7-11.8) 63.47 (52.0-74.7) 23 335 (44.85%)

$60 875.72 (29 583.2-132 232.2) 8.73 (5.0-16.1) 74.26 (62.2-83.1) 227 888 (51.94%)

$49 828.81 (26 484.8-97 890.4) 7.31 (4.4-12.8) 66.53 (54.1-78.2) 106 187 (50.79%)

$29 301.71 (14 744.5-61 236.1) 3.37 (1.4-7.3) 47.15 (34.6-57.0) 4131 (40.59%)

$61 438.56 (24 921.4-149 750.9) 5.58 (1.6-13.3) 72.91 (60.9-82.7) 97 460 (49.50%)

b.0001

33 125 (63.64%) 6538 (12.67%) 6979 (13.34%) 5431 (10.35%)

282 883 (64.50%) 65 758 (15.10%) 49 466 (11.29%) 40 461 (9.11%)

134 215 (64.25%) 29 212 (14.07%) 26 605 (12.72%) 18 994 (8.96%)

5236 (51.43%) 2544 (25.18%) 1714 (16.91%) 671 (6.48%)

124 890 (63.41%) 27 676 (14.19%) 27 541 (14.01%) 16 732 (8.39%)

b.0001 b.0001 b.0001 b.0001

$16 565 (32.94%) $13 763 (27.19%)

$129 148 (30.25%) $108 727 (25.31%)

$62 282 (30.57%) $51 575 (25.18%)

$4136 (43.80%) $2317 (24.42%)

$58 738 (30.87%) $47 324 (24.68%)

b.0001 b.0001

$11 348 (22.25%)

$104 538 (24.27%)

$49 478 (24.09%)

$1891 (19.95%)

$45 014 (23.43%)

b.0001

$8930 (17.62%)

$86 507 (20.17%)

$41 344 (20.17%)

$1124 (11.83%)

$40 253 (21.02%)

b.0001

P

.4996 b.0001 b.0001

Data were represented as median (interquartile range) for numeric variables and frequency (proportion) for categorical variables. P values were reported by linear trend for numerical variables and χ2 test for categorical variables.

failure remained stable, whereas the incidence of metabolic and neurologic failure increased significantly (P b .0001). The characteristics, by proportion, of patients who died in the hospital, discharged routinely without homecare, or discharged to a short-term acute care hospital or and a skilled outpatient health care facility are reviewed in Table 4. Those transferred out to a skilled outpatient health care facility tended to have a longer LOS during the acute hospitalization. Thirty percent of surviving patients were discharged home; the rate increased to 33% in 2012. Patients who left the hospital against medical advice were more likely to be men, have a history of substance abuse and AIDS diagnosis, and have Medicaid or no insurance.

In the multivariate-adjusted analysis (Table 5), the strongest predictors of mortality were female sex (odds ratio [OR], 1.13) and weight loss (OR, 2.14). In addition, certain organ system failures were more likely to be associated with mortality including respiratory (OR, 5.70 [95% confidence interval {CI}, 5.56-5.85]), cardiovascular (OR, 3.22 [95% CI, 3.14-3.30]), hepatic (OR, 2.47 [95% CI, 2.38-2.57]), and a neurologic event (OR, 2.76 [95% CI, 2.67-2.86]). Table 5 shows the estimated OR with 95% CI of each independent variable on each category of the dependent variable compared with the reference category (routine care). The strongest predictors of a transfer to a short-term acute care hospital compared with routine discharge to home were hospital characteristics including rural location (OR, 4.53 [95% CI, 4.06-5.06]), nonteaching status for urban hospitals (OR, 2.17

Table 5 Multivariate-adjusted ORs with 95% CIs for different types of disposition vs routine other Routine discharge home without services (as Reference = 1)

Variable name Female No. of chronic conditions Comorbidity CHF Liver disease Weight loss Organ failure Respiratory Cardiovascular Renal Hepatic Hematologic Metabolic Neurologic Major operating room procedure Hospital location/teaching status Urban teaching Rural Urban non-teaching Hospital bed-size Large Small Medium

Short-term another hospital OR (95% CI)

Other health care including a long-term care

Home with home services

Against medical advice

Died in hospital

0.88 (0.86-0.90) 1.05 (1.04-1.06)

1.17 (1.15-1.18) 1.07 (1.07-1.08)

1.09 (1.07-1.10) 1.07 (1.07-1.08)

0.69 (0.66-0.73) 1.01 (1.01-1.02)

1.13 (1.11-1.15) 1.03 (1.02-1.04)

1.37 (1.33-1.41) 1.04 (1.00-1.10) 2.02 (1.92-2.13)

1.84 (1.81-1.88) 0.64 (0.62-0.66) 2.52 (2.44-2.61)

1.42 (1.39-1.46) 0.79 (0.76-0.81) 1.90 (1.83-1.96)

0.97 (0.91-1.04) 1.88 (1.72-2.04) 1.13 (1.05-1.22)

1.82 (1.78-1.86) 0.94 (0.91-0.97) 2.14 (2.06-2.22)

2.42 (2.34-2.51) 1.65 (1.59-1.70) 1.41 (1.37-1.45) 1.88 (1.78-1.99) 1.08 (1.04-1.12) 1.30 (1.25-1.36) 1.98 (1.89-2.08) 1.30 (1.25-1.36)

1.83 (1.79-1.88) 1.38 (1.35-1.41) 1.37 (1.34-1.39) 1.29 (1.24-1.34) 0.82 (0.80-0.84) 1.12 (1.10-1.15) 2.69 (2.61-2.77) 1.85 (1.81-1.90)

1.20 (1.17-1.23) 1.13 (1.10-1.15) 1.11 (1.08-1.13) 1.14 (1.10-1.19) 0.90 (0.88-0.92) 1.09 (1.06-1.12) 1.50 (1.46-1.55) 1.65 (1.62-1.68)

1.27 (1.19-1.36) 0.86 (0.81-0.92) 1.07 (1.01-1.12) 1.17 (1.02-1.33) 0.88 (0.82-0.95) 1.17 (1.08-1.26) 1.26 (1.16-1.38) 0.59 (0.55-0.63)

5.70 (5.56-5.85) 3.22 (3.14-3.30) 1.99 (1.95-2.03) 2.47 (2.38-2.57) 1.18 (1.15-1.21) 1.91 (1.86-1.97) 2.76 (2.67-2.86) 1.25 (1.22-1.29)

Reference 4.53 (4.06-5.06) 2.17 (1.95-2.42)

Reference 1.07 (1.00-1.15) 1.08 (1.01-1.14)

Reference 0.77 (0.71-0.84) 0.94 (0.87-1.00)

Reference 0.73 (0.65-0.82) 1.01 (0.93-1.11)

Reference 1.21 (1.12-1.30) 1.04 (0.98-1.11)

Reference 3.17 (2.88-3.50) 1.56 (1.41-1.72)

Reference 1.27 (1.18-1.36) 1.12 (1.05-1.19)

Reference 0.97 (0.89-1.04) 0.90 (0.83-0.98)

Reference 1.00 (0.89-1.13) 1.13 (1.03-1.25)

Reference 1.33 (1.23-1.44) 1.07 (1.00-1.14)

Data were represented as median (interquartile range) for numeric variables and frequency (proportion) for categorical variables. CHF indicates chronic heart failure.

S. Elfeky et al. / Journal of Critical Care 39 (2017) 48–55

[95% CI,1.95-2.42]), and small size of the hospital (OR, 3.17 [95% CI, 2.883.50]). Respiratory failure and weight loss were highly associated with transfer to a short-term hospital (ORs, 2.42 [2.34-2.51] and 2.02 [1.922.13] respectively) compared with routine discharge. That is to say, patients with sepsis who had a respiratory failure more likely died in hospital (5.70 [5.56-5.85]) and discharged to another short-term hospital (2.42 [2.34-2.51]) than routine discharge when compared with patients with sepsis who did not have a respiratory failure. The predictors of transfer to a skilled outpatient health care facility were having a major procedure in the operating room (OR, 1.85 [95% CI, 1.81-1.90]), a neurologic event (OR, 2.69 [95% CI, 2.61-2.77, cardiovascular disease (OR, 1.38 [95% CI, 1.35-1.41), and respiratory failure (OR, 1.8 [95% CI, 1.79-1.88]). Approximately 20% of the data have missing information on race; hence, the multivariate-adjusted models were not adjusted for race. That is to say, the number of chronic conditions only highly correlated with the number of diagnoses (Pearson ρ = 0.79, P b .0001). Also, the total charge, LOS, age, organ system failure, number of procedures, and severity of illness in the final multivariate-adjusted model were not adjusted for because of high correlation among the above covariates, for example, the total charge and severity (Pearson ρ = 0.47, P b .001), age and primary payer (Pearson ρ =0.49, P b .001). Therefore, from the above risk factors, we adjusted for the number of chronic conditions and other risk factors in the final adjusted model (Table 5). 4. Discussions We investigated mortality rates, comorbidity trends, LOS, and the disposition information of patients diagnosed as having sepsis in the US population-based NIS database for years 2009 through 2012. This study revealed that total number of sepsis cases increased by 11% from 2009 to 2012 which is in agreement with previous publications [22]. We expect an increasing trend in sepsis cases because of the aging population, the increasing number of chronic health conditions in the US population, and the increased use of immunosuppressive treatments for conditions other than autoimmune diseases and cancer. The in-hospital mortality of sepsis declined steadily, from 16.5% to 13.8%, despite the increase in the diagnosis of sepsis and a concomitant increase in the number of comorbidities and associated organ failures. Not surprising, a higher total number of organ failures was associated with death. Survivors experienced a significant increase in the number of discharges to an outpatient skilled health care facility and home health care between 2009 and 2012. Cardiovascular failure, respiratory failure, and renal failure were more likely to be associated with requiring home health care services and death. The average inpatient LOS in our study decreased from 6.7 days in 2009 to 5.9 days in 2012. Parallel to our results, Lagu et al found a 2% yearly decrease in sepsis mortality between 2003 and 2007 as well as a slight decrease in LOS, from 9.9 to 9.2 days [23]. Similarly, AyalaRamirez et al demonstrated a downward trend in sepsis mortality in Spain. In their study, sepsis mortality decreased from 40.0% to 31.8% in men and from 41.6% to 35.2% in women from 2003 to 2011 [24]. Some of the important reasons for the decline in mortality rates include heightened awareness, earlier recognition, aggressive therapy, and the advent and broad use of therapies directed for sepsis. Our national efforts to decrease hospitalization durations may have contributed to the decreased LOS in our study. Although our study showed a decrease in LOS, this finding was not uniform in all sepsis cases. Patients who were ultimately discharged to a skilled outpatient health care facility had twice the LOS compared with patients discharged routinely to home. During the study years, the proportion of patients discharged to a skilled outpatient health care facility home was similar to the number discharged to home. More patients were discharged with a need for home health care or a skilled facility than discharged to home without needing these services. Kumar et al. reported that the proportion of patients discharged home routinely was stable in 2000 to 2007, whereas the proportion of patients

53

discharged to home with home care services and to skilled nursing facilities was rising [15]. We predict that there will continue to be an increasing need for home care services and outpatient skilled health care facilities as more patients with sepsis survive their hospitalization. Our study finds that multiple comorbidities, organ system failures, and need for major surgery predict increased postdischarge health care utilization. Compared with routine discharge, patients with respiratory or hepatic failure were more likely to be discharged to a shortterm acute care hospital. On the other hand, patients with neurologic events were more likely to be discharged to an outpatient skilled health care facility. Preexisting chronic heart failure was associated with mortality and those who did survive were more likely to be discharged to a short-term acute care hospital or a skilled outpatient health care facility. In our NIS data, home health care included simple support and hospice care; thus, we felt that we could accurately describe and discuss the comorbidities and organ failures of this group. The impact of sepsis-related hospitalizations on society extends well beyond lives lost. Although our study describes inpatient data and demonstrated a decrease in in-hospital mortality, the literature is robust in demonstrating long-term morbidity in survivors of sepsis. Unfortunately, we were not able to identify hospitalizations that were actually rehospitalizations secondary to sepsis. Survivors often experience long-term physical and psychological symptoms often resulting in impaired functional status and reduced health care quality of life, now identified as the post–sepsis syndrome [25]. Iwashyna et al [26] found that severe sepsis survivors had a clinically and statistically significant increase in moderate to severe cognitive impairment. Deficits occurred in both older and younger patients and persisted. At 12 months, 34% of older patients and 24% of younger patients had assessments that were similar to scores for patients with moderate traumatic brain injury and scores for patients with mild Alzheimer disease, respectively. Among those who do survive, impaired quality of life, worsening cognitive and/or functional status, and re hospitalization increase health care utilization and cost [27,28]. Sepsis, outpatient or hospital acquired, can worsen preexisting chronic diseases and new chronic diseases may emerge. Those with multiple chronic health conditions have an increased risk of infections and sepsis which adds to the health care burden. Initial inpatient costs represent only 30% of the total medical and societal cost of sepsis and are related to severity and LOS, whereas lost productivity and other direct and indirect medical costs (eg, premature death, early retirement, lost productivity) after hospitalization account for most of the economic burden of sepsis [29,30]. More studies are needed to evaluate how our advances in technology could help care for these patients and decrease the cost of care. Sjoding et al [31] found that the primary diagnoses of patients admitted to critical care units have substantially changed from 1996 to 2010 with a shift from cardiovascular care to infectious diseases. Previously, sepsis was the 11th ranked diagnosis among Medicare beneficiaries older than 64 years who required ICU care. In 2010, sepsis was the top ranked primary discharge diagnosis with a concomitant decline in admissions with a primary diagnosis of cardiovascular disease. These changes reflect the need for hospitals, community physicians, and public health planners to focus efforts on prevention and management of sepsis in the geriatric population given these epidemiologic changes. For the developed world, enhanced patient selection for advanced medical care, especially in the ICU, has great potential to alleviate suffering and reduce cost by aggressively treating infections and appropriately addressing goals of care as soon as possible. In the outpatient setting, the Program of Allinclusive Care for the Elderly helps older individuals to continue to live at home while receiving medical and social services with a goal of benefit to the individual and a reduction in the total cost of health care [32,33]. Sex and race disparities in health care warrant further investigation. Our study indicates that more women with sepsis die in the hospital or are transferred to a skilled outpatient facility. This observed difference in sex mortality is in agreement with the findings of Pietropaoli et al [34]. Understanding this issue not only improves morbidity and

54

S. Elfeky et al. / Journal of Critical Care 39 (2017) 48–55

mortality in women in general but also may help decrease the economic burden of sepsis during all stages of health care during sepsis. Despite advances in therapy and improvements in mortality, treatment disparities between men and women persist [35]. Biological factors may contribute to the observed increase in mortality in female patients hospitalized with sepsis. Women have a more robust immune response that may impact the inflammatory cascade [36]. Several clinical and epidemiologic studies have investigated the impact of sex on outcome in various clinical settings, yielding conflicting results [37-42]. Our study revealed a lower proportion of whites, compared with the 2010 US census report [21], with sepsis compared with the other racial groups in the study. Much is yet to be understood about the genetic influences on wellness and disease, and this information will help us improve the outcome of many health conditions, including sepsis. Our results indicated that weight loss was associated with in hospital mortality and utilization of health care services upon discharge. Hence, aggressive prevention and management of nutrition during the hospital course needs to be addressed. Malnutrition has been associated with poor outcomes among patients in ICUs, as indicated by increased morbidity, mortality, and LOS [43]. Evidence suggests that in patients with a functional gut, nutrition should be administered through the enteral route, largely because of the morbidity associated with other routes of feeding [44,45]. Parenteral nutrition is especially associated with increased infectious complications [46]. Given the long-term morbidity and mortality after discharge, there is a compelling need to improve postdischarge management of sepsis survivors and to further investigate the increased morbidity and mortality experienced by this population. Because morbidity and mortality of

sepsis survivors extend years after discharge, then the extension of home health services (nursing visits, medications checks) for longer periods may prevent future hospitalizations. Simple interventions like hand washing, aggressive nutrition, and avoiding contact with sick people should be emphasized. This study also has several limitations. First, the database that we used, NIS, is an administrative and discharge-level database. Because it lacks clinical information, we were not able to detect multiple hospitalizations for one patient and were unable to do a chart review to confirm the diagnosis. Using ICD-9 codes to identify cases of sepsis is dependent on the accuracy of hospital coding practices. The number of sepsis cases could be overestimated due to changes in coding practices. Rhee et al [47] found that although the incidence of hospitalizations with sepsis codes increased dramatically, the incidence of hospitalizations with positive blood cultures decreased by 17% [47]. Quality of life and longterm (postdischarge) outcomes are also not reported. We, however, analyzed a large, well-validated national database. The data are robust and representative of the entire US population. 4. Conclusion In conclusion, despite an increase in number of sepsis cases, the LOS and in hospital mortality associated with sepsis decreased during our study period and several organ failures and weight loss predict discharge disposition. With the increased need for skilled care after hospital discharge, expansion of skilled nursing and long-term care facilities will be needed and new and innovative outpatient health care programs will be necessary as our population continues to age.

Appendix A Table 1 Patient characteristics by calendar year in all patients, NIS, 2009-2012 Variables

Total

2009

2010

2011

2012

Total no. of all admissions Total no. of sepsis admissions Crude rate per 100 000 all hospital discharges Short-term acute care hospital Skilled outpatient health care facility Home health care Against medical advice Died in-hospital Total charge ($)

25 443 292 1 303 640 4302

6 447 155 299 992 3840

6 382 035 315 617 4046

6 631 809 354 066 4413

5 982 293 333 965 4577

541 670 (2.13%) 3 923 204 (15.42%) 3 002 645 (11.8%) 291 254 (1.14%) 539 011 (2.11%)

141 624 (2.23%) 950 855 (14.97%) 714 377 (11.25%) 72 731 (1.15%) 139 256 (2.18%) $20 047.97 (10 824.7-39 303.4) 2.61 (1.4-4.9) 58.03 (39.1-74.0) 3 829 853 (59.93%)

137 757 (2.18%) 967 449 (15.40%) 753 532 (12.09%) 73 924 (1.18%) 133 780 (2.13%) $21 274.54 (11 356.4-41 792.9) 2.62 (1.4-4.9) 57.93 (39.3-73.8) 3 727 350 (59.37%)

137 618 (2.15%) 1 060 965 (16.29%) 800 070 (12.23%) 73 557 (1.13%) 138 792 (2.14%) $22 338.08 (11 994.9-43 529.4) 2.62 (1.4-4.8) 59.11 (40.3-74.4) 3 881 980 (59.56%)

124 671 (2.08%) 943 935 (15.78%) 734 666 (12.28%) 71 042 (1.19%) 127 183 (2.13%) $22 781.86 (12 229.5-44 384.1) 2.61 (1.4-4.8) 58.70 (39.5-73.8) 3 549 541 (59.33%)

3 723 115 (58.37%) 738 817 (11.60%) 950 977 (14.92%) 972 803 (15.10%)

3 792 297 (60.41%) 865 143 (14.06%) 899 976 (14.34%) 714 294 (11.18%)

4 075 511 (62.60%) 884 061 (13.66%) 933 796 (14.40%) 622 021 (9.34%)

3 909 234 (65.35%) 826 821 (13.82%) 932 972 (15.60%) 313 266 (5.24%)

LOS (d) Age (y) Female Race White Black Other Missing

15 435 157 (60.7%) 3 314 842 (13%) 3 717 721 (14.6%) 2 622 384 (10.31%)

P

.9139 1 .9681 .2476 b.0001 b.0001 b.0001 b.0001

Data were represented as median (interquartile range) for numeric variables and frequency (proportion) for categorical variables; P values were reported by linear trend for numerical variables and χ2 test for categorical variables.

References [1] Castellanos-Ortega A, Suberviola B, Garcia-Astudillo LA, Holanda MS, Ortiz F, Llorca J, et al. Impact of the Surviving Sepsis Campaign protocols on hospital length of stay and mortality in septic shock patients: results of a three-year follow-up quasi-experimental study. Crit Care Med 2010;38:1036–43. [2] Levy MM, Dellinger RP, Townsend SR, Linde-Zwirble WT, Marshall JC, Bion J, 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. [3] Levy MM, Rhodes A, Phillips GS, Townsend SR, Schorr CA, Beale R, et al. Surviving Sepsis Campaign: association between performance metrics and outcomes in a 7.5-year study. Crit Care Med 2015;43:3–12.

[4] Ferrer R, Artigas A, Levy MM, Blanco J, Gonzalez-Diaz G, Garnacho-Montero J, et al. Improvement in process of care and outcome after a multicenter severe sepsis educational program in Spain. JAMA 2008;299:2294–303. [5] Torio CM, Andrews RM. National Inpatient Hospital Costs: The Most Expensive Conditions by Payer, 2011. Statistical Brief #160; 2013. [6] Elixhauser A, Friedman B, Stranges E. Septicemia in US Hospitals, 2009. Healthcare cost and Utilization Project, Statistical Brief #122; 2011. [7] CDC. Centers for Disease Control and Prevention, Prevalence and Most Common Causes of Disability Among Adults, United States, 2005. MMWR Morb Mortal Wkly Rep 2009;58(16):421–6. [8] Iwashyna TJ, Ely EW, Smith DM, Langa KM. Long-term cognitive impairment and functional disability among survivors of severe sepsis. JAMA 2010;304:1787–94. [9] Prescott HC, Langa KM, Liu V, Escobar GJ, Iwashyna TJ. Increased 1-year healthcare use in survivors of severe sepsis. Am J Respir Crit Care Med 2014;190:62–9.

S. Elfeky et al. / Journal of Critical Care 39 (2017) 48–55 [10] Winters BD, Eberlein M, Leung J, Needham DM, Pronovost PJ, Sevransky JE. Longterm mortality and quality of life in sepsis: a systematic review. Crit Care Med 2010;38:1276–83. [11] Yende S, D'Angelo G, Kellum JA, Weissfeld L, Fine J, Welch RD, et al. Inflammatory markers at hospital discharge predict subsequent mortality after pneumonia and sepsis. Am J Respir Crit Care Med 2008;177:1242–7. [12] Wier LM, Pfuntner A, Maeda J, Strangers E, Ryan K, Jagadish P. Healthcare Cost and Utilization Project (HCUP) facts and figures: statistics on hospital- based care in the United States, 2009. Agency for Healthcare Research and Quality; 2011. [13] Hall MJ, Williams SN, DeFrances CJ, Golosinskiy A. Inpatient care for septicemia or sepsis: a challenge for patients and hospitals. NCHS Data Brief 2011:1–8. [14] Yende S, Iwashyna TJ, Angus DC. Interplay between sepsis and chronic health. Trends Mol Med 2014;20:234–8. [15] Kumar G, Kumar N, Taneja A, Kaleekal T, Tarima S, McGinley E, et al. Nationwide trends of severe sepsis in the 21st century (2000-2007). Chest 2007;140:1223–31. [16] Mayr FB, Yende S, Angus DC. Epidemiology of severe sepsis. Virulence 2014;5:4–11. [17] HCUP. Agency for Healthcare Research and Quality. Rockville, MD: HCUP Quality Control Procedures; 1998. [18] HCUP. Comparative analysis of HCUP and NHDS inpatient discharge data. The HCUP3 nationwide inpatient sample, release 5, 1996. 5th ed. Rockville, MD: Agency for Healthcare Policy and Research; 1998. [19] Shoemaker O. Variance estimates for price changes in the consumer price index January-December 2012. Consumer Price Index, Bureau of Labor Statistics; 2013. [20] Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care 1998;36:8–27. [21] 2010 Census Data; 2011. [22] 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;348:1546–54. [23] Lagu T, Rothberg MB, Shieh MS, Pekow PS, Steingrub JS, Lindenauer PK. Hospitalizations, costs, and outcomes of severe sepsis in the United States 2003 to 2007. Crit Care Med 2012;40:754–61. [24] Ayala-Ramirez OH, Dominguez-Berjon MF, Esteban-Vasallo MD. Trends in hospitalizations of patients with sepsis and factors associated with inpatient mortality in the Region of Madrid, 2003-2011. Eur J Clin Microbiol Infect Dis 2014;33:411–21. [25] Angus DC, Laterre PF, Helterbrand J, Ely EW, Ball DE, Garg R, et al. The effect of drotrecogin alfa (activated) on long-term survival after severe sepsis. Crit Care Med 2004;32:2199–206. [26] Iwashyna TJ, Cooke CR, Wunsch H, Kahn JM. Population burden of long-term survivorship after severe sepsis in older Americans. J Am Geriatr Soc 2012;60:1070–7. [27] Tiru B, DiNino EK, Orenstein A, Mailloux PT, Pesaturo A, Gupta A, et al. The Economic and Humanistic Burden of Severe Sepsis. PharmacoEconomics 2015;33:925–37. [28] Weycker D, Akhras KS, Edelsberg J, Angus DC, Oster G. Long-term mortality and medical care charges in patients with severe sepsis. Crit Care Med 2003;31: 2316–23.

55

[29] Schmid A, Pugin J, Chevrolet JC, Marsch S, Ludwig S, Stocker R, et al. Burden of illness imposed by severe sepsis in Switzerland. Swiss Med Wkly 2004;134:97–102. [30] Schmid A, Burchardi H, Clouth J, Schneider H. Burden of illness imposed by severe sepsis in Germany. Eur J Health Econ 2002;3:77–82. [31] Sjoding MW, Prescott HC, Wunsch H, Iwashyna TJ, Cooke CR. Longitudinal Changes in ICU Admissions Among Elderly Patients in the United States. Crit Care Med 2016; 44:1353–60. [32] Kane RL, Homyak P, Bershadsky B, Flood S. Variations on a theme called PACE. J Gerontol A Biol Sci Med Sci 2006;61:689–93. [33] Rylander C. Recommendations of the Texas Cmptroller: Chapter 8: Health and Human Services 'Expand the use of an effective long-term care program'; 2000. [34] Pietropaoli AP, Glance LG, Oakes D, Fisher SG. Gender differences in mortality in patients with severe sepsis or septic shock. Gend Med 2010;7:422–37. [35] Graham G, Xiao YY, Rappoport D, Siddiqi S. Population-level differences in revascularization treatment and outcomes among various United States subpopulations. World J Cardiol 2016;8:24–40. [36] Fairweather D, Frisancho-Kiss S, Rose NR. Sex differences in autoimmune disease from a pathological perspective. Am J Pathol 2008;173:600–9. [37] Croce MA, Fabian TC, Malhotra AK, Bee TK, Miller PR. Does gender difference influence outcome? J Trauma 2002;53:889–94. [38] Valentin A, Jordan B, Lang T, Hiesmayr M, Metnitz PG. Gender-related differences in intensive care: a multiple-center cohort study of therapeutic interventions and outcome in critically ill patients. Crit Care Med 2003;31:1901–7. [39] Reinikainen M, Niskanen M, Uusaro A, Ruokonen E. Impact of gender on treatment and outcome of ICU patients. Acta Anaesthesiol Scand 2005;49:984–90. [40] Adrie C, Azoulay E, Francais A, Clec'h C, Darques L, Schwebel C, et al. Influence of gender on the outcome of severe sepsis: a reappraisal. Chest 2007;132:1786–93. [41] Fowler RA, Sabur N, Li P, Juurlink DN, Pinto R, Hladunewich MA, et al. Sex-and agebased differences in the delivery and outcomes of critical care. CMAJ 2007;177: 1513–9. [42] Combes A, Luyt CE, Trouillet JL, Nieszkowska A, Chastre J. Gender impact on the outcomes of critically ill patients with nosocomial infections. Crit Care Med 2009;37: 2506–11. [43] Robinson G, Goldstein M, Levine GM. Impact of nutritional status on DRG length of stay. JPEN J Parenter Enteral Nutr 1987;11:49–51. [44] Sullivan DH, Sun S, Walls RC. Protein-energy undernutrition among elderly hospitalized patients: a prospective study. JAMA 1999;281:2013–9. [45] Haydock DA, Hill GL. Improved wound healing response in surgical patients receiving intravenous nutrition. Br J Surg 1987;74:320–3. [46] Chandra RK. Nutrition, immunity, and infection: present knowledge and future directions. Lancet 1983;1:688–91. [47] Rhee C, Murphy MV, Li L, Platt R, Klompas M. Comparison of trends in sepsis incidence and coding using administrative claims versus objective clinical data. Clin Infect Dis 2015;60:88–95.