In-Hospital Mortality and Economic Burden Associated With Hepatic Encephalopathy in the United States From 2005 to 2009

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2012;10: In-Hospital Mortality and Economic Burden Associated With Hepatic Encephalopathy in the United States From 2005 to 2009 MARIA STEPANOVA, ALITA MISHRA, CHAPY VENKATESAN, and ZOBAIR M. YOUNOSSI Center for Liver Diseases, Department of Medicine, Inova Fairfax Hospital; and Betty and Guy Beatty Center for Integrated Research, Inova Health System, Falls Church, Virginia BACKGROUND & AIMS: Hepatic encephalopathy (HE) is a major complication of cirrhosis that causes substantial mortality and utilization of resources. METHODS: We analyzed 5 cycles of the Nationwide Inpatient Sample, conducted between 2005 and 2009, to determine national estimates of incidence, prevalence, inpatient mortality, severity of illness, and resource utilization for inpatients with HE. RESULTS: The yearly inpatient incidence of HE ranged from 20,918 (2005) to 22,931 (2009) (P.2226), comprising approximately 0.33% of all hospitalizations in the United States. Over the 5-year period of analysis, mortality of inpatients with HE remained relatively stable, at 14.13% 15.61% (P.062); however, the proportion of patients with major and extreme severity of illness increased (P.0001). The average length of inpatient stay increased from 8.1 to 8.5 days (P.019). The average total inpatient charges increased from $46,663 to $63,108 per case (P.0001). Furthermore, total national charges related to HE increased from $ million (2005) to $ million (2009). In multivariate analysis, independent predictors of inpatient mortality included the number of diagnoses per admission (odds ratio [OR] 1.022; 95% confidence interval [CI], per diagnosis), number of procedures per admission (OR per procedure; 95% CI, ), and major or extreme severity of illness (OR 3.16; 95% CI, ). The most important predictors of cost, charge, and length of stay were admission to a large, urban hospital; use of Medicaid or Medicaid as the payer; major or extreme severity of illness; number of diagnoses at discharge; and procedures per admission (P.05). CONCLUSIONS: Resource utilization increased from 2005 to 2009 for patients discharged from US hospitals with the diagnosis of HE. The inpatient mortality rate, however, remained stable, despite a trend toward more severe disease. Keywords: Cost Analysis; Chronic Liver Disease; Treatment; Management. Hepatic encephalopathy (HE) is a major complication of end-stage chronic liver disease. 1 In the absence of chronic liver disease, HE can also manifest as a complication of portalsystemic shunting. 1 The spectrum of HE ranges from minimal brain function deficits, known as minimal HE, to hepatic coma. 2 4 Overt HE occurs in approximately 30% 45% of patients with cirrhosis, while minimal HE may affect up to 60% of patients with chronic liver disease and up to 80% with cirrhosis. 5 7 However, the accurate data on the true incidence and prevalence of HE is lacking, mainly because of large differences in the etiology and severity of HE and the difficulty in diagnosing minimal HE Development of HE is associated with a poor prognosis. Specifically, in the presence of chronic liver disease, HE typically heralds hepatic decompensation, and its development is usually associated with high mortality, indicating the need for liver transplantation The management of patients hospitalized with decompensated liver disease is known to have a substantial economic impact. In previous studies that analyzed national administrative data from the early 2000s, the total economic burden of decompensated cirrhosis was reported to be between 1 and 2 billion dollars annually with an increasing statistical trend. 9,15 16 In this study, we used the Nationwide Inpatient Sample (NIS), a large, nationally representative, inpatient database, to evaluate recent trends in hospitalizations of patients with HE in the United States and to estimate its nationwide inpatient resource utilization related to HE. Methods Patient Population The data used for the study included NIS collected between 2005 and The NIS is an all-payer database of hospital discharges maintained as part of the Healthcare Cost and Utilization Project (HCUP) by the Agency for Healthcare Research and Quality. The NIS approximates a 20% stratified sample of community hospitals in the United States. The sampling frame in each of the surveys used for the study is a sample of hospitals that comprise 90% to 95% of all hospital discharges in the respective year in the United States. Specifically, between 2005 and 2009, each NIS cycle included information on approximately 8 million discharges from approximately 1000 hospitals located in 37 (in 2005) to 44 (in 2009) states. Each record in the NIS represents a single hospital discharge and includes a unique identifier, basic demographic information for the patient, admission and disposition type, up to 15 diagnoses, a list of comorbid conditions, up to 15 procedures, expected primary insurance payer, total hospital charges (not including physician or other professional services), and length of stay. Furthermore, Abbreviations used in this paper: CCR, cost-to-charge ratio; CI, confidence interval; HCUP, Healthcare Cost and Utilization Project; HE, hepatic encephalopathy; NIS, Nationwide Inpatient Sample; OR, odds ratio by the AGA Institute /$

2 September 2012 IN-HOSPITAL AND ECONOMIC BURDEN OF HE 1035 the data on each participating hospital includes its basic description such as type (teaching/nonteaching, member of a health system), size, location, and mean cost-to-charge ratio (CCR). Study Definitions In this study, we included all adult (age 18 years) hospital discharges between the years of 2005 and 2009 with HE listed as a primary or secondary diagnosis (International Classification of Diseases 9th Version code 572.2) at the time of discharge from the hospital. Additional data used for the study included patients demographic information, admission source (classified into: admission from emergency room; another hospital; another facility including long-term care; court or law enforcement facility; routine admission), admission type (elective or not), disposition type (classified into: routine; transfer to shortterm hospital; other transfers, including skilled nursing facility, intermediate care, and another type of facility; home health care; against medical advice; died in hospital), primary payer (Medicare; Medicaid; private including health maintenance organization; self-pay; no charge), 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 into tertiles relative to statewide), location (rural or urban), region (Northeast, Midwest, South, or West), membership in a health system (for the cycles only) and teaching status of the hospital. The total number of procedures and diagnoses were also included with each record. Furthermore, 29 potential comorbidities were also evaluated and reported for each discharge together with the severity of illness for the patient measured according to the All Patient Refined Diagnosis Related Groups 17 (classified into: minor loss of function, includes cases with no comorbidity or complications; moderate loss of function; major loss of function; extreme loss of function). Finally, we also studied a number of procedures that we expected to be typically performed in patients with HE admitted to the US hospitals. For the purpose of the study, comorbid diagnoses and procedures were identified by the Clinical Classification Software categorization scheme developed by the Agency for Healthcare Research and Quality as a part of HCUP. Major outcomes evaluated in this study were in-hospital mortality and resource utilization parameters. The latter included total charges (adjusted to the year of 2009 with the coefficient of inflation equal to 3% annually), total costs (estimated for each patient using the average CCR for the respective hospital), and the length of stay measured in days. Statistical Analysis National estimates were obtained for the total number of hospitalizations with HE and total resource utilization parameters by calendar year using individual discharge sampling weights. Because not all participating hospitals had reported their average CCRs, the national estimates of total costs were reweighted to account for all discharges where cost estimates were missing, as recommended by the HCUP. 18 All available clinical and socioeconomic parameters were further compared between years using stratum-specific 2 test for independence (for binary or categorical parameters) or a t test for a contrasted mean (for continuous parameters) to identify the parameters that changed significantly over time. P values of.05 or less were considered statistically significant. In the analysis, the stratum units (used in the NIS to sample hospitals based on geographic region, control, location/teaching status, and bed size) accounted for the survey design effects using Taylor series linearization. Multiple logistic regression models were used to identify factors associated with in-hospital deaths while accounting for potential confounders. Similarly, factors that impacted hospital charges and length of stay were assessed using multiple linear regressions after logarithmic transformation of the respective outcomes. Coefficients from these models were exponentiated to yield a percentage change in the outcomes associated with each predictor. All analyses were run with SAS 9.1 and SUDAAN 10.0 (SAS Institute, Inc, Cary, NC). The study was approved by the Inova Institutional Review Board. Results Overall, 33.3 million hospitalizations were included in the 5 yearly cycles of NIS, representing million hospitalizations nationwide. Of those, the overall number of hospitalizations with HE was 111,090 ( %) in 5 years, representing total 548,000 hospitalizations in the United States. Between 2005 and 2009, there was no change in the incidence of HE (P.223). The demographic characteristics of patients hospitalized with the diagnosis of HE are summarized in Table 1. Specifically, between 2005 and 2009, the proportion of patients of 65 years of age or older increased from 30.2% to 33.0% (P.0001), as did the proportion of female patients with HE (from 39.2% to 41.6%, P.003). Concurrently with the target population becoming older, the proportion of those covered by Medicare also increased to near statistical significance from 42.9% to 45.1% (P.059) while the proportion of Medicaid decreased from 22.9% to 20.2% (P.0007). The remaining demographic and socioeconomic parameters, hospital characteristics, and sources of admission did not change over the study period. On the other hand, the proportion of HE patients with routine discharges decreased from 43.1% to 40.2% (P.0009), while the proportion of dispositions to skilled nursing facility, intermediate care, and another type of facility increased from 22.0% to 24.7% (P.001). Clinical Parameters and In-Hospital Mortality in Patients With HE The most prevalent comorbidities among patients with HE included alcohol abuse, coagulopathy, renal failure, and fluid/electrolyte disorders (Table 2). Between 2005 and 2009, of 29 reported comorbidities, only 3 (alcohol abuse, congestive heart failure, and peptic ulcer disease) did not demonstrate significant changes. A close to significant trend to decrease was detected for alcohol abuse (P.066). The rates of 4 comorbid conditions (AIDS, chronic blood loss anemia, drug abuse, lymphoma) decreased, while the rates for the remaining 22 comorbidities increased significantly. The most notable increase was observed for obesity which increased from 3.0% to 7.9% (P.0001); consistent with the increase in the rate of obesity and its complications observed nationwide (Table 2).

3 1036 STEPANOVA ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 10, No. 9 Table 1. Demographic Characteristics of Patients Hospitalized With HE in Parameter P Total number of hospital discharges in the United States, millions a Nationwide incidence of HE a 102, , , , , Prevalence (%) Age, y (%) Female (%) Ethnicity (%) White Black Hispanic Asian Native American Neighborhood income relative to national (%) 1st quartile nd quartile rd quartile th quartile Hospital characteristics (%) Small Medium Large Urban Teaching Member of a health system Not available Not available Hospital location (%) Northeast Midwest South West Primary payer (%) Medicare Medicaid Private including HMO Self pay No charge Admission source (%) Emergency room Another hospital Another facility Law enforcement Routine Elective admission Disposition (%) Routine Another hospital Another facility Home health care Against medical advice HMO, health maintenance organization. a The nation-wide estimates for both all hospital discharges and discharges with HE are calculated using the NIS sampling weights. The overall number of diagnoses for patients with HE steadily increased over the 5 year time period from 10.0 per record in 2005 to 13.3 per record in 2009 (P.0001). The pattern for the severity of illness also changed significantly toward a more severe disability (minor or moderate severity indexes decreased from 22.1% to 14.8%; extremely severe index increased from 35.9% to 43.1%, P.0001). Despite increasing complexity and severity of illness, the inpatient mortality for patients with HE decreased from 15.6% in 2005 to 14.4% in 2009 (P.062) with the lowest being 14.1%

4 September 2012 IN-HOSPITAL AND ECONOMIC BURDEN OF HE 1037 Table 2. Clinical Characteristics of Patients Hospitalized With HE in Parameter P Severity of illness (%) Minor Moderate Major Extreme Number of diagnoses Presence of comorbidity (%) AIDS Alcohol abuse Anemia Rheumatoid arthritis/collagen vascular diseases Chronic blood loss Congestive heart failure Chronic pulmonary disease Coagulopathy Depression Diabetes, uncomplicated Diabetes with chronic complications Drug abuse Hypertension Hypothyroidism Lymphoma Fluid, electrolyte disorders Metastatic cancer Other neurological disorders Obesity Paralysis Peripheral vascular disorders Psychoses Pulmonary circulation disorders Renal failure Solid tumor without metastasis Peptic ulcer disease excluding bleeding Valvular disease Weight loss In-hospital mortality, (%) in Independent predictors of in-hospital mortality in patients with HE are summarized in Table 3. Specifically, both the number of concomitant diagnoses on the record (odds ratio [OR] [95% confidence interval (CI), ]), the number of procedures per admission (OR 1.192; 95% CI, )] having major or extreme disease severity (OR 3.162; 95% CI, ), and fluid and electrolyte disorders as a comorbidity (OR 1.455; 95% CI, ) were all independently associated with inpatient mortality. After adjusting for major demographic and socioeconomic confounders, in-hospital mortality for patients with HE significantly decreased over the time period (OR 0.910; 95% CI, per year). Annual Dynamics in Resource Utilization Overall total charges for hospitalizations with HE increased from $4.677 billion to $7.254 billion nationwide. This corresponds to a 55.1% increase in 5 years after adjustment for inflation (Table 4). This increasing trend was steady with yearly increases ranging from 8.2% (in ) to 15.7% (in ). Similarly, the mean inflation-adjusted charge per patient also increased by 35.3% in 5 years. The largest proportion of the mean charge occurred between 2006 and 2007 (15.2%). The 2 major resource utilization parameters, namely, the length of stay and the average number of procedures per admission, also demonstrated an increasing trend over the study period (Table 4). Specifically, the average length of stay for hospitalization with HE increased from days in 2005 to days in 2009, P.019. Furthermore, the average number of procedures on the record for patients with HE also increased over time from to (P.0001). The rate of rise in charges was more rapid than that of length of stay or number of procedures. Similarly, the estimated total national cost increased only 23.8% over the same period of time (from $ billion to $ in inflationadjusted dollars) while the mean per-patient cost did not change ($16, in 2005 to $17, in 2009; P 0.169) (Table 4). The yearly dynamics for the cost and charge values are shown in Supplementary Figure 1. Independent predictors of the 3 resource utilization parameters are summarized in Table 5. The largest contributors

5 1038 STEPANOVA ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 10, No. 9 Table 3. Predictors of In-Hospital Mortality in Patients With HE OR (95% CI) Calendar year, per y ( ) Primary diagnosis of HE ( ) Age 45 y ( ) Neighborhood income, per quartile ( ) Medicare ( ) Medicaid ( ) Minor disease severity ( ) Major or extreme disease severity ( ) Presence of alcohol abuse as a comorbidity ( ) Presence of anemia related to chronic ( ) blood loss as a comorbidity Presence of fluid and electrolyte disorders ( ) as a comorbidity Number of diagnoses, per discharge ( ) Number of procedures, per discharge ( ) to the increased length of stay in patients with HE included elective admissions, major or extreme disease severity, and weight loss at admission. Hospitals in the Midwest, South, and West region reported significantly shorter length of stay for their patients with HE compared with the Northeast even after adjustment for major demographic confounders. Both Medicare and Medicaid were associated with a slightly longer length of stay. Similar results were seen with other characteristics of a hospital such as urban location and larger bed size. Finally, each procedure was found to be associated with 14.5% longer stay, and each diagnosis was also associated with 2.2% increase in the length of stay. Independent predictors of cost and charge increase were expectedly similar to those of the length of stay (Table 5). However, hospitals that are large or urban or located in the West region had an increase in charges out of proportion to increases in length of stay. The most notable difference in costs and charges were found for the West region (those hospitals report 22.0% shorter stay and similar costs compared with the rest of the country, and charge 17.9% more per admission) and urban location (9.9% longer stay compared with rural hospitals, with 5.0% higher cost and 52.8% higher charge). Elective admissions were found to be associated with longer stay while not contributing to higher costs or charges. Finally, while both length of stay and treatment cost of patients with HE demonstrated significantly decreasing trends over the study span (after adjustment for confounders), the yearly decrease was 3.1% ( 3.8% to 2.5%) and 2.8% ( 3.8% to 1.6%), respectively, the average charge steadily increased ( 1.8% [ 0.2% to 3.6%] per year). Discussion In this study, which used recent available data on inpatient admissions, we found that the inpatient incidence of HE remains stable at the level of approximately 100,000 cases per year nationwide, making up approximately 0.33% of all inpatient admissions in the United States. The socioeconomic portrait of a patient admitted with HE remained the same: 17% 19% of patients were in the Northeast region, 18% 22% in Midwest, 36% 37% in the South, and 24% 26% in the West. The proportion of patients covered by Medicare increased concurrently with the median age, while the proportion of Medicaid decreased from 23% to 20%. Patients with HE also tend to live in areas with lower income; 30% 32% of them come from the neighborhoods with the average income that fall in the lowest national quartile while only 17% 18% are from the top quartile neighborhoods. Furthermore, at least two-thirds of individuals with HE admitted to the US hospitals are covered by a governmentsponsored insurance such as Medicare or Medicaid. The higher rate of HE in the setting of lower income is, in fact, a concerning finding which may require a more accurate analysis of the potential socioeconomic obstacles in the access to treatment for patients with chronic liver disease before they eventually progress to cirrhosis and HE. The average length of stay for hospitalization of patients with HE increased modestly and remained at the level of 8 to 9 days per admission. The average number of procedures per admission increased from 1.91 to Admitted patients have also become older, more frequently female, and the typical severity of the disease evaluated both by the qualitative scale and by the number of concomitant diagnoses per record or the number of reported comorbidities increased significantly. The rates of most of studied comorbidities also increased with the most increased one being obesity which went up by 265% over 5 years. An indirect indicator of a more severe disease in patients hospitalized with HE could also be found in the distribution of where the patients were disposed to after receiving treatment in a hospital: the proportion of routine discharges decreased by 7% while 12.5% more patients were transferred to long-term care facilities in the study period. On the contrary, the inpatient mortality demonstrated a significant trend to decrease, especially after adjustment for the increased severity of the disease in the target population (OR 0.91; 95% CI, per year). Indeed, most of the parameters that reflect a more severe disease were found to be associated with increased mortality (Table 3). These data could potentially reflect improvement in treatment for patients with cirrhosis and HE. The only characteristics of a hospital which Table 4. Resource Utilization for Patients Hospitalized With HE in Parameter P Number of procedures Length of stay, d Average charge (2009 $) 46, , , , , Average cost (2009 $) 16, , , , , Total national charge, (2009 $) millions Total national cost, (2009 $) millions

6 September 2012 IN-HOSPITAL AND ECONOMIC BURDEN OF HE 1039 Table 5. Independent Predictors of Inpatient Resource Utilization in Patients With HE Length of stay, % (95% CI) Cost, % (95% CI) Charge, % (95% CI) Calendar year, per y 3.1 ( 3.8 to 2.5) 2.8 ( 3.8 to 1.6) 1.8 ( 0.2 to 3.6) Primary diagnosis of HE 10.9 ( 12.2 to 9.6) 14.4 ( 15.5 to 13.2) 14.4 ( 16.0 to 12.7) Age 45 y 1.4 ( 1.3 to 4.2) 3.4 ( 1.5 to 5.1) 2.6 ( 0.1 to 5.2) Female 4.7 ( 3.5 to 6.0) 3.5 ( 2.3 to 4.6) 3.4 ( 2.0 to 4.7) Neighborhood income, per quartile 0.9 ( 1.7 to 0.1) 2.5 ( 1.4 to 3.7) 1.1 ( 0.9 to 3.1) Elective admission 20.2 ( 16.4 to 24.2) 1.7 ( 5.2 to 1.8) 0.4 ( 4.2 to 5.2) Hospital Large 5.1 ( 2.4 to 7.8) 1.9 ( 1.5 to 5.4) 15.8 ( 9.4 to 22.6) Urban 9.9 ( 6.5 to 13.3) 5.0 ( 1.2 to 8.9) 52.8 ( 44.5 to 61.6) Teaching 3.1 ( 5.6 to 0.3) 6.4 ( 2.4 to 10.5) 3.0 ( 3.2 to 9.6) Location Midwest 16.0 ( 19.5 to 12.2) 13.2 ( 18.8 to 7.1) 18.8 ( 27.3 to 9.2) South 12.1 ( 15.1 to 9.1) 15.7 ( 20.5 to 10.6) 12.6 ( 21.6 to 2.6) West 22.0 ( 24.8 to 19.1) 2.3 ( 8.3 to 4.3) 17.9 ( 5.7 to 31.8) Medicare 6.7 ( 5.2 to 8.3) 0.2 ( 1.3 to 1.8) 1.1 ( 0.8 to 2.9) Medicaid 7.8 ( 5.7 to 10.0) 2.4 ( 0.6 to 4.4) 0.8 ( 1.8 to 3.5) Minor disease severity 12.6 ( 15.1 to 10.0) 11.6 ( 14.0 to 9.1) 11.7 ( 14.8 to 8.5) Major or extreme disease severity 26.6 ( 24.2 to 29.0) 31.1 ( 28.9 to 33.2) 31.0 ( 28.3 to 33.8) Comorbidity Alcohol abuse 1.6 ( 0.2 to 3.0) 1.1 ( 0.2 to 2.4) 0.1 ( 1.7 to 2.0) Chronic blood loss anemia 9.6 ( 11.7 to 7.4) 7.3 ( 9.5 to 5.0) 6.6 ( 9.3 to 3.8) Fluid and electrolyte disorders 6.6 ( 5.3 to 7.9) 8.3 ( 7.1 to 9.4) 8.0 ( 6.6 to 9.3) Weight loss 28.1 ( 25.7 to 30.6) 16.3 ( 13.9 to 18.8) 17.6 ( 14.5 to 20.8) Number of diagnoses, per diagnosis 2.2 ( 1.9 to 2.4) 2.5 ( 2.2 to 2.8) 2.4 ( 1.9 to 2.9) Number of procedures, per procedure 14.5 ( 13.9 to 14.9) 23.5 ( 22.6 to 24.2) 23.9 ( 23.0 to 24.7) was found to be associated with higher in-hospital mortality of patients with HE was the location in the Northeast region which was probably due to demographic characteristics of that region which were not accounted for in our study, such as race/ethnicity, which was not available for more than 25% of the sample. Between 2005 and 2009, an average per-case charge increased from $46,663 to $63,107. The total national charges also increased from $ million to $ million in 2009 dollars. Nevertheless, during the same period of time, the average cost per hospitalization remained the same after adjustment for inflation (P.169) while the total national cost increased proportionally to the increase in the incidence from $1651 million to $2044 million. Previously reported findings from estimated the total national charges for inpatient treatment of HE at the level of $930 million. However, this estimate was calculated for HE being listed as a primary diagnosis only. In our study, the primary diagnosis encompassed approximately 40% of hospitalizations with HE (from 42.25% in 2005 to 37.55% in 2009). Thus, after adjustment for this classification disparity and yearly inflation, we conclude that between 2003 and 2005, the increase in total charges was at least 60%, and it was followed by 55% more in the following 5 years. Table 6. Characteristics of the US Hospitals Participating in the NIS Data Collection Year P N Participating states Size (%) Small Medium Large Urban (%) Teaching (%) Location (%) Northeast Midwest South West In a health system (%) Na Na Median CCR Average CCR per patient Na, not available for the yearly cycle.

7 1040 STEPANOVA ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 10, No. 9 As could be seen, the average per-case charge increased by more than 50% in 5 years and so did the total national charge. However, this increase in charge was not accompanied by a comparable increase in the resource utilization, such as length of stay, number of procedures, and total cost of treatment. Therefore, we attempted to analyze the causes of such increase using the reported average CCRs which were available for most of the participating hospitals. The descriptive summary on the participating hospitals is given in Table 6. Indeed, the distribution of the participating hospitals remained nationally representative and did not change between 2005 and 2009: 45% 46% were small, 24% 25% medium, and 30% 31% were large based on their bed size (P.998), 60% were urban (P.999), 16% 19% were teaching (P.28) and 55% were members of health systems (P.97, the data were available starting from 2007). The geographic distribution also remained the same (P.97). At the same time, the median CCR for the participating hospitals decreased from 0.51 to 0.44 (P.0011). A more illustrative parameter, however provided that approximately two-thirds of patients with HE were treated in large hospitals and larger hospital size was found to be associated with higher charge is per-patient CCR. Indeed, we found that the per-patient CCR in the United States, as could be seen from the National Inpatient Sample, decreased from 0.42 to 0.36 (P.0001) between 2005 and Therefore, at least partially, the observed substantial increase in the total national and average charges for treatment of HE could be attributable to the causes associated with this decrease in the CCR nationwide which is a socioeconomic rather than medical phenomenon. Our study has several limitations. One of those is the lack of actual treatment cost data, which caused us to use the estimates only. As previously noted, we observed a trend in cost related to HE. Although this could be explained by a rapid use of newer treatment regimens, we do not have access to the list of medications administered to patients sampled for NIS, and thus could not provide any evidence to support this hypothesis. Furthermore, the total direct costs, which necessarily include physician fees and out-patient visits, together with indirect costs such as lost wages because of hospitalizations, cognitive dysfunction and overall impaired quality of life as well as shorter life expectancy, were not accounted for in our study. 19 That led to a significant underestimation of true total economic burden associated with the increasing incidence of HE. Furthermore, a recent study reported that an associated caregiver burden can also have a significant impact toward overall socioeconomic burden of the disease. 20 This indirect cost could further impact the total economic impact of HE. The dataset also had limited clinical and demographic data which did not allow us to explore the factors potentially associated with the outcomes and resource utilization more thoroughly. Finally, the used data were reported per admission, not per individual, so we could not distinguish between unique patients and those being readmitted with the same disease, which could potentially result in an overestimate in the observed incidence of HE as readmissions of the same patient are not separated in these data analyses. Nevertheless, the increasing number of admissions still poses a substantial economic burden. The key advantage of our study, however, is that using the NIS allows us to make conclusions and evaluate the longitudinal trends using consistently and accurately collected extensive administrative data which is also nationally representative with respect to geographic region, demographic characteristics of the patients, health insurance payers, and different types of acute care hospitals across the country. To conclude, the nationwide inpatient incidence of advanced liver disease is increasing as is the prevalence of its feared complication of HE. Furthermore, the socioeconomic and financial burden of this disease consistently increases due to both the observed trend toward a more severe disease and higher treatment costs. Our findings confirm a strong necessity for the implementation of more effective public policy together with preventive care strategies and interventions that could reduce the incidence and delay the progression of chronic liver disease to advanced liver disease and its complications. Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at and at doi.org/ /j.cgh References 1. Ferenci P, Lockwood A, Mullen K, et al. Hepatic encephalopathydefinition, nomenclature, diagnosis, and quantification: final report of the working party at the 11th world congresses of gastroenterology, Vienna. Hepatology 1998;2002: Cash WJ, McConville P, McDermott E, et al. Current concepts in the assessment and treatment of hepatic encephalopathy. QJM 2010;103: Mas A. Hepatic encephalopathy: from pathophysiology to treatment. Digestion 2006;73(Suppl 1): Polson J, Lee WM; American Association for the Study of Liver Disease. AASLD position paper: the management of acute liver failure. Hepatology 2005;41: Amodio P, Del Piccolo F, Pettenò E, et al. Prevalence and prognostic value of quantified electroencephalogram (EEG) alterations in cirrhotic patients. J Hepatol 2001;35: Romero-Gómez M, Boza F, García-Valdecasas MS, et al. Subclinical hepatic encephalopathy predicts the development of overt hepatic encephalopathy. Am J Gastroenterol 2001;96: Montgomery JY, Bajaj JS. Advances in the evaluation and management of minimal hepatic encephalopathy. Curr Gastroenterol Rep 2011;13: Lewis M, Howdle PD. The neurology of liver failure. QJM 2003; 96: Poordad FF. Review article: the burden of hepatic encephalopathy. Aliment Pharmacol Ther 2007;25(Suppl 1): Randolph C, Hilsabeck R, Kato A, et al. Neuropsychological assessment of hepatic encephalopathy: ISHEN practice guidelines. Liver Int 2009;29: Benhaddouch Z, Abidi K, Naoufel M, et al. Mortality and prognostic factors of the cirrhotic patients with hepatic encephalopathy admitted to medical intensive care unit [in French]. Ann Fr Anesth Reanim 2007;26: Udayakumar N, Subramaniam K, Umashankar L, et al. Predictors of mortality in hepatic encephalopathy in acute and chronic liver disease: a preliminary observation. J Clin Gastroenterol 2007; 41: Findlay JY, Fix OK, Paugam-Burtz C, et al. Critical care of the end-stage liver disease patient awaiting liver transplantation. Liver Transpl 2011;17: Bustamante J, Rimola A, Ventura PJ, et al. Prognostic signifi-

8 September 2012 IN-HOSPITAL AND ECONOMIC BURDEN OF HE 1041 cance of hepatic encephalopathy in patients with cirrhosis. J Hepatol 1999;30: Neff G. Pharmacoeconomics of hepatic encephalopathy. Pharmacotherapy 2010;30:28S 32S. 16. Nguyen GC, Segev DL, Thuluvath PJ. Nationwide increase in hospitalizations and hepatitis C among inpatients with cirrhosis and sequelae of portal hypertension. Clin Gastroenterol Hepatol 2007;5: Averill RF, Goldfield N, Hughes JS. All patient refined diagnosis related groups (APR-DRGs). Available at: ahrq.gov/db/nation/nis/apr-drgsv20methodologyoverviewand Bibliography.pdf. Accessed Feb 7, Cost-to-charge ratio files: 2009 nationwide inpatient sample (NIS) user guide. Available at: state/ccr2009nisuserguide.pdf. Accessed Feb 7, Sanyal A, Younossi ZM, Bass NM, et al. Randomised clinical trial: rifaximin improves health-related quality of life in cirrhotic patients with hepatic encephalopathy - a double-blind placebo-controlled study. Aliment Pharmacol Ther 2011;34: Bajaj JS, Wade JB, Gibson DP, et al. The multi-dimensional burden of cirrhosis and hepatic encephalopathy on patients and caregivers. Am J Gastroenterol 2011;106: Reprint requests Address requests for reprints to: Zobair M. Younossi, MD, MPH, Betty and Guy Beatty Center for Integrated Research, Claude Moore Health Education and Research Building, 3300 Gallows Road, Falls Church, Virginia zobair.younossi@inova.org; fax: (703) Conflicts of interest The authors disclose no conflicts. Funding This study has been supported in part by the Beatty Liver and Obesity Research Fund and by the Liver Disease Outcomes Fund of the Center for Liver Diseases at Inova Fairfax Hospital, Inova Health System, Falls Church, Virginia.

9 September 2012 IN-HOSPITAL AND ECONOMIC BURDEN OF HE 1041.e1 Supplementary Figure 1. Yearly dynamics ( ) in total national resource utilization for patients admitted to US hospitals with HE.