| |
HCV & Coinfection-Expected Increased Hospitalizations Rates: complications expected to quadruple
|
| |
| |
HCV has aptly been described as a sleeping giant: Our findings corroborate existing concerns over the aging of patients with HCV. As the patients continue to age and the disease burden progresses, suboptimal decisions regarding HCV treatments will bring increasing opportunity costs for the health care system and society.
"Trends in health care resource use for hepatitis C virus infection in the United Statesh
Hepatology
Dec 2005 Volume 42, Issue 6, Pages 1406-1413
William C. Grant 1, Ravi R. Jhaveri 2 3, John G. McHutchison 2 4, Kevin A. Schulman 1 5 *, Teresa L. Kauf 1 6
1Center for Clinical and Genetic Economics, Duke University Medical Center, Durham, NC
2Duke Clinical Research Institute, Duke University Medical Center, Durham, NC
3Division of Infectious Diseases, Department of Pediatrics, Duke University Medical Center, Durham, NC
4Divisions of Gastroenterology, Duke University Medical Center, Durham, NC
5Divisions of General Internal Medicine, Duke University Medical Center, Durham, NC
6Department of Medicine, Duke University Medical Center, Durham, NC
".....HCV infection represents a substantial health care burden.... future HCV-related complications will presumably increase and present further economic and other stresses on the health care system..... However, increasing rates of patients with disease progression are anticipated to raise the proportion of patients with cirrhosis from 25% in 2010 to 38% in 2040.....
....The primary outcomes of interest - hospitalizations, charges, hospital days, and physician visits - showed average annual increases in the range of 25% to 35%, indicating that the future burden of HCV infection will match and may exceed analysts' forecasts....
...From 1990 to 2015, the number of persons with long-standing complications is expected to at least quadruple...
....office visits by HCV patients increased at an average annual rate of 36%....
....Nearly 3 times as many HIV patients were hospitalized for liver-related reasons in 2001 than in 1994. Co-infected patients constituted 44.9% of HIV liver-related hospitalizations in 2001, up from just 9.1% in 1994.....
....From 1994 through 2001, HCV-related hospitalizations, hospital days, total charges, and deaths increased at average annual rates exceeding 20%, more than 3-fold higher than rates for all-cause hospitalizations.....
....Patients in their 40s and 50s also spent more time in the hospital, incurred greater costs, and died more frequently than patients in other age groups....
....Compared with hospital days for any reason, HCV liver causes for all ages accounted for almost 4 times as many hospital days in 2001 as in 1994 (371 vs. 98 per 100,000)....
....Racial/ethnic minorities constituted roughly half of HCV-related hospitalizations but less than 20% of physician visits....
....Vertical transmission rates drive the trends among children and adolescents, with active carriers born to mothers with HCV infection with 1% to 5% likelihood......data indicate that pediatric HCV trends merit further attention (HIV increases HCV MTCT rates)... Trends among pediatric HCV patients are relatively more volatile and less informative than for older patients...
HIV-HCV Co-infection.
(I extracted & put here the coinfection data since its so compelling. See other detailed results & Discussion below)
Patients with HIV-HCV co-infection were another important subgroup in the analysis. Because highly active antiretroviral therapy was introduced to many of these patients during the study period, we examined co-infection statistics for evidence of change. As shown in Table 5, increases in liver-related hospitalizations were approximately proportional to those of patients infected only with HCV. As a fraction of all HCV liver-related hospitalizations, hospitalization frequencies for co-infected patients were consistently between 2% and 3%. Both within and across age groups, hospitalizations among co-infected patients kept pace with rates of increase for mono-infected patients.
For the aggregate HIV population, however, HCV liver complications caused many more hospitalizations. Nearly 3 times as many HIV patients were hospitalized for liver-related reasons in 2001 than in 1994. Co-infected patients constituted 44.9% of HIV liver-related hospitalizations in 2001, up from just 9.1% in 1994. As a fraction of all HIV hospitalizations (not specifically liver-related), co-infected patients constituted 7.5 times as many hospitalizations in 2001 as in 1994, rising from 0.4% to 3.1%. The overall magnitude of the HIV burden remains much larger than that of HCV. Over the 8-year study period, HIV hospitalizations occurred at a frequency 3.4 times that of HCV liver-related hospitalizations. Between 1994 and 2001, total inpatient charges for HIV hospitalizations were 2.9 times higher than for HCV liver-related hospitalizations.
ABSTRACT
Chronic hepatitis C virus (HCV) infection affects approximately 3 million people in the United States (Brian Edlinfs study presented at AASLD Nov 2005 finds 5 million with HCV & 4 million with chronic HCV due to including homeless, incarcerated & other populations not reported in previous NHAHNES data) and places tremendous demands on the health care system.
As many observers have predicted, the disease burden continues to grow as the infected population ages.
In this study, we analyzed inpatient data from the Healthcare Cost and Utilization Project, outpatient data from the National Ambulatory Medical Care Survey, and drug data from the Verispan Source Prescription Audit.
We examined recent growth in the use of health care resources among HCV patients by age group and found average annual increases of 25% to 30% for hospitalizations, charges, hospital days, and physician visits. Corresponding time-trend coefficients were positive (P < .001). From 1994 to 2001, the HCV burden increased among patients aged 40 to 60 years, reflecting the natural history of disease progression.
In sensitivity analysis, HCV outcome growth rates remained significant, unless more than 3 out of 4 cases were initially underreported.
Also, patients co-infected with HIV and HCV in 2001 constituted 7.5 times as many hospitalizations and incurred 2.9 times the charges in 1994, relative to all HIV hospitalizations and charges.
Our findings highlight the urgency concerning HCV outcomes. In conclusion, as patients continue to age and disease burden progresses, suboptimal decisions regarding HCV treatments will bring increasing opportunity costs for the health care system and society.
Article Text
As many observers have predicted, the treatment burden of hepatitis C virus (HCV) continues to grow as the population of patients with HCV grows older. With approximately 27,000 hospitalizations and $500 million in hospital charges recently attributed to the disease in the United States,[1] HCV infection represents a substantial health care burden. Because of the long, asymptomatic disease course,[2] many patients with HCV infection are not diagnosed until the disease has progressed to some degree of liver damage. Available treatments are expensive, are associated with debilitating and sometimes dangerous side effects, and are effective in only approximately half of patients with genotype 1 HCV infection, the most common in the United States. The incidence of HCV infection has fallen sharply because of increased awareness of HIV and safer practices for injection drug use.[3-7] However, existing and undiagnosed cases represent a latent threat to public health. As the population of patients with HCV infection ages, future HCV-related complications will presumably increase and present further economic and other stresses on the health care system.
Seroprevalence data have allowed researchers to estimate the potential magnitude of HCV infection and thus the possible complications. In the sample population of the Third National Health and Nutrition Examination Survey, HCV antibodies and HCV RNA were observed with respective frequencies of 1.8% and 1.4% among surveyed individuals.[8][9] From these figures, it is estimated that 3 million Americans are chronically infected with HCV. Incidence-based models provide estimates of similar magnitude.[8] Because many of the infections occurred in the 1970s, a wave of complications during the 1990s was predicted as the duration of infections reached critical thresholds. Given the relatively low incidence of new infections, the total number of infected patients will eventually diminish because of deaths in the existing cohort. However, increasing rates of patients with disease progression are anticipated to raise the proportion of patients with cirrhosis from 25% in 2010 to 38% in 2040.[10]
Because of the potentially significant impact of HCV on public health and the health care system overall, updated expectations concerning HCV trajectories with quantitative observations from the recent past would be of great value. Therefore, we analyzed hospitalization, outpatient, and prescription data over 8 years (1994 through 2001) to better understand how HCV outcomes are evolving. Our objective was to provide longitudinal statistics concerning health care resource use by patients with HCV, with a particular focus on the relative distribution of HCV-related complications across age groups. We intended the analysis to inform important policy perspectives regarding treatment for patients with HCV in the near future.
Materials and Methods
Inpatient Trends.
We examined hospitalization trends using the Nationwide Inpatient Sample (NIS) from the Healthcare Cost and Utilization Project (HCUP).[11] The NIS approximates a 20% stratified sample of community hospitals in the United States and samples hospitals according to relevant characteristics (e.g., urban vs. rural, county vs. private) to produce a truly representative sample. After extrapolation to the national level, the NIS represents approximately 35 million hospitalizations per year and contains clinical, demographic, and economic variables.
Each NIS observation contains up to 15 ICD-9-CM diagnoses. If a patient had a primary or secondary diagnosis of HCV infection (ICD-9-CM codes 070.41, 070.44, 070.51, or 070.54), we defined the hospitalization as HCV-related. To better discern hospitalizations likely associated with HCV-related complications, we also identified liver-related hospitalizations. Based on criteria set forth by Kim et al.[1] we considered a hospitalization to be liver-related if (1) the principal diagnosis was HCV- or alcohol-related liver disease; (2) there was any diagnosis of cirrhosis, portal hypertension, other sequelae of liver disease, hepatic encephalopathy, ascites, hepatorenal syndrome, hepatocellular carcinoma, or combined hepatocellular carcinoma and cholangiocarcinoma; or (3) the patient underwent liver transplantation.
For all HCV-related hospitalizations and for liver-related HCV hospitalizations, we examined admittance frequencies, total hospital days, liver transplantations, nationwide charges, and deaths for the years 1994 through 2001. We constructed statistics for female patients with HCV, patients co-infected with HCV and HIV (ICD-9-CM codes 042.00 and 079.53), and 10-year age groups to better understand demographic changes in the burden of HCV infection. Annual magnitudes and annual group-specific shares for outcome variables provide insight concerning the HCV patient life cycle. To adjust for overall hospital population changes, we computed all statistics both in raw terms and relative to all hospitalizations. We adjusted for inflation using the US Consumer Price Index for Medical Care.
To assess the rate of change in outcomes, we used Poisson regressions for annual count data, treating each 10-year age group as the individual level of observation, observed in each of the 8 years from 1994 to 2001. For patients diagnosed with HCV, we ran separate regressions for the following dependent variables: hospitalization frequencies, liver-related hospitalization frequencies, total hospital days, and deaths. The regressor for each of these regressions was a simple time-trend variable for the years 1994 to 2001, the coefficient of which was used to assess the significance of the rates of change in outcomes over the period.
Outpatient Trends.
In addition to hospitalization trends, we examined changes in ambulatory care for patients with HCV and spending on prescription drugs for HCV. Physician visits were tracked for the years 1996 through 2002 using the National Ambulatory Medical Care Survey (NAMCS).[12] Based on randomly sampled visits during week-long reporting periods, NAMCS contains information reported by office-based physicians engaged in direct patient care. If HCV was indicated in any of the 3 diagnostic fields, we categorized the office visit as HCV-related. Stratified by 10-year age groups, the NAMCS data displayed fairly unstable numbers of HCV-related office visits. As a result, we focused on 3-year moving averages to smooth out year-to-year volatility in the data.
Drug Trends.
Aggregate annual spending on prescriptions for interferon-ribavirin combination therapy (ribavirin-interferon alfa-2b) from 1998 through 2000 was taken from the Verispan Source Prescription Audit,[13] which provides total national prescription payments collected from more than 35,000 retail pharmacies in the United States. Although interferon monotherapy was also used to treat HCV during these years, we have no way of identifying which prescriptions were intended to treat cancer or other conditions. Combination therapy, however, is only prescribed for HCV. Our interferon-ribavirin statistics describe the aggregate volume of prescription demand and do not contain patient-level data.
Analysis of Underreporting.
Because HCV antibody testing was introduced in 1991, it is unlikely that the dissemination of testing into practice was complete by 1994. The observed outcome frequencies probably understate the actual numbers of HCV outcomes by greater proportions during earlier years of our study period. To the extent that the rate of HCV testing and reporting increased over time, the growth rates observed in our data may overstate the increase in health care resource use for HCV. Therefore, we conducted a sensitivity analysis to consider the potential impact of underreporting.
As a baseline scenario, we hypothesized that 50% of cases were underreported in 1994 and that underreporting fell each year by some fixed amount, so that 0% of cases were underreported in 2001. For comparison purposes, we allowed the initial underreporting rate to vary between 25% and 75%. For each scenario, we maintained the assumption that underreporting fell each year by an amount resulting in 0% underreporting in 2001.
For purposes of comparison, we calculated hospitalization statistics for alcohol-induced cirrhosis of the liver (ICD-9-CM code 571.2). This condition was not as likely as HCV to be misclassified in the early years of the study period. Differences in statistics for HCV and alcohol-induced cirrhosis likely reflect both underreporting effects and differences in underlying disease burden.
Results
Tables 1 and 2 provide demographic profiles for hospitalizations and physician visits for the first and last years of each data set. Racial/ethnic distributions were stable, except for a moderate increase in the share of hospitalizations among Hispanic patients. However, the racial/ethnic profile for hospitalizations differed markedly from the profile for physician visits. Racial/ethnic minorities constituted roughly half of HCV-related hospitalizations but less than 20% of physician visits. Similar changes were seen with respect to patient sex.
Hospitalizations.
From 1994 through 2001, HCV-related hospitalizations, hospital days, total charges, and deaths increased at average annual rates exceeding 20%, more than 3-fold higher than rates for all-cause hospitalizations. Supplementary Table 1 reports annual frequencies by year for these variables. Table 3 reports year-over-year growth rates in these statistics, averaged over the 8 years. The growth patterns were most striking for patients in their 40s and 50s. Because these groups' shares of hospitalizations rose so much, other age groups saw their shares decline. Patients in their 40s and 50s also spent more time in the hospital, incurred greater costs, and died more frequently than patients in other age groups. From 1994 to 2001, patients aged 40 to 49 saw their share of liver-related hospital days increase from 32.3% to 37.6%. Similarly, the share of liver-related hospital days among patients aged 50 to 59 rose from 17% to 30.1%. Compared with hospital days for any reason, HCV liver causes for all ages accounted for almost 4 times as many hospital days in 2001 as in 1994 (371 vs. 98 per 100,000).
In the Poisson regressions of the rates of change in outcomes, the frequencies and lengths of stay increased significantly for all HCV-related hospitalizations and for liver-related HCV hospitalizations, relative to all-cause hospitalizations (P < .001 for all comparisons). We also conducted sensitivity analyses of our assumptions regarding underreporting of HCV infection in earlier years of the study period (Fig. 1). In the baseline scenario - in which 50% of actual cases were underreported in 1994 - the average annual increase in HCV liver-related hospitalizations was 11.1%. The increase in hospitalizations that we observed in the HCUP data would be entirely attributable to underreporting only if more than 77% of cases in 1994 were not reported.
Except among adolescents, the observed growth in HCV liver-related hospital days was entirely extensive. That is, more hospital days resulted from more frequent hospitalizations, as opposed to longer lengths of stay. For HCV liver-related hospitalizations, average length of stay declined to 6.9 days in 2001 from 8.5 in 1994. However, adolescents experienced both extensive and intensive growth. Average length of stay for HCV liver-related hospitalizations increased by 2.1 days, and the number of liver-related hospitalizations more than doubled (244 vs. 111).
Increasing hospitalizations and hospital days coincided with higher expenditures for HCV. For every $100,000 in nationwide charges from all hospitalizations, liver-related HCV hospitalizations resulted in $427 in charges in 2001, compared with only $145 in 1994. Once again, patients in the 40s and 50s age groups accounted for a large majority of the increase. Higher spending for HCV resulted both from more hospital days and higher charges per hospitalization, with 37% more charges per HCV liver-related hospital day in 2001 than in 1994.
Table 4 shows trends in HCV-related hospitalizations among women. Women aged 30 to 59 years accounted for much less than 50% of HCV hospitalizations; however, younger women and women older than 60 were hospitalized with HCV liver diagnoses more frequently than men. This phenomenon among women in their 20s disappeared after correcting for the spurious diagnosis of HCV in women whose primary reason for diagnosis was pregnancy related (ICD-9-CM codes 641-677).
Physician Visits.
Outpatient data corroborate the upward trend in HCV burden. In the 9 years of NAMCS data, office visits by HCV patients increased at an average annual rate of 36%. Between 1996 and 2002, the 3-year moving average rose by approximately 1 million annual office visits, from 449,800 in 1996 to 1.49 million in 2002 (Supplementary Table 1). Three-year averages were pulled up by spikes in 2002 office visits among patients aged 40 to 59. We did not pursue further subgroup analysis because of small and unstable numbers of women, children, and HIV-HCV-co-infected patients in the HCV outpatient population.
Prescription Drugs.
The Verispan Source Prescription Audit data[13] revealed the entry of the new ribavirin-interferon combination drug treatments for HCV. A total of $352.5 million was spent on 372,000 prescriptions for ribavirin plus interferon between 1998 and 2001, with no prescriptions observed before 1998. For every $100,000 in new prescriptions, spending for ribavirin-interferon rose from $78 in 1998 to $259 in 2000. Meanwhile, $297.9 million was spent on 578,000 prescriptions for interferon monotherapy between 1993 and 2000. Spending for interferon monotherapy increased every year before 1998 (when combination therapy entered the market) and decreased every year thereafter.
Discussion
This study documents accelerating growth in the use of health care resources by patients with HCV. The primary outcomes of interest - hospitalizations, charges, hospital days, and physician visits - showed average annual increases in the range of 25% to 35%, indicating that the future burden of HCV infection will match and may exceed analysts' forecasts. From 1990 to 2015, the number of persons with long-standing complications is expected to at least quadruple.[14] In the 8 years from 1994 through 2001, we found 3-fold to 4-fold increases. By 2015, long-standing infections may lead to even greater numbers of complications than have been predicted.
Our findings corroborate existing concerns over the aging of patients with HCV. Shatin et al.[15] estimated that the prevalence rate for chronic HCV infection is highest for the group aged 45 to 54 years (277 per 100,000). We found that the role of aging was highlighted by disproportionately high outcome shares attributable to patients aged 40 to 60 years, among whom average annual growth statistics above 30% and even 40% were common. As others have noted,[10][16] worsening health for patients with HCV advancing into middle and late middle age exacts increasing costs.
The possibility of underreporting in the mid-1990s poses a major limitation to this study. Real increases in HCV outcomes were almost certainly less than the observed increases because of underreporting. However, some limit to this distortion is evident from the age group variability. HCV liver-related hospitalizations occurred with nearly the same frequencies for the 40s age group and the 60s age group in 1994. We have little reason to expect that underreporting would be far greater for 40-year-olds than for 60-year-olds in 1994. Given that hospitalizations increased 4 times as much for the former group than for the latter, our overall findings are unlikely to be driven entirely by underreporting.
Nevertheless, our sensitivity analysis quantified the extent to which hypothetical underreporting mitigated the rise in hospitalizations. In every scenario, significant rises in hospitalizations were robust despite possible underreporting effects. Even if 50% of HCV liver-related hospitalizations were underreported in 1994, we found an increase in hospitalizations greater than 200% between 1994 and 2001. An initial underreporting rate exceeding 77% is necessary to completely nullify the increase in hospitalizations.
We found that growth rates were much lower for alcohol-induced cirrhosis than the corresponding growth rates for HCV, which may result in part from more underreporting of HCV in the early years of the study period. From 1994 to 2001, growth in liver-related HCV hospitalizations exceeded growth in alcohol-induced cirrhosis by a factor of 14. For growth in liver transplantations, total charges, and in-hospital death, HCV exceeded alcohol-induced cirrhosis by factors of 3.3, 5.7, and 56, respectively.
Differences in hospitalization rates between men and women provide additional insight into disease progression. Because infections are approximately twice as common among men than among women,[15] finding more frequent liver-related HCV hospitalizations among men was not surprising. However, our analysis revealed interesting sex-based variation in hospitalization rates according to age group, with women in their 60s being hospitalized for HCV liver-related reasons more frequently. The age group phenomenon may result in part from gender differences in the progression of HCV. In general, complications are more progressive in men than women, causing gender variations in times between peaks in infections and hospitalization shares. The rise in relative shares for women after age 60 may be a cumulative result of slower progression in women. Many of these late-life hospitalizations involve serious complications, including death, which are relatively more likely to have occurred before age 60 for men than for women.
Vertical transmission rates drive the trends among children and adolescents, with active carriers born to mothers with HCV infection with 1% to 5% likelihood.[17] Current guidelines do not involve HCV screening for newborns, leading to uninformative data concerning changes in childhood infection rates. This may help explain why hospitalization rates remained flat for children younger than 10 years, whereas simultaneously a 15% average annual increase occurred in HCV hospitalizations among adolescents. These patterns may reflect in part new infections during adolescence, but initial HCV presentation in vertically infected children likely accounts for part of the increase. Have vertical infections really remained flat? What are the implications for future changes in HCV outcomes among adolescents? Because newborns are not routinely screened for HCV, answers based on observational hospital data are not clear, but the data indicate that pediatric HCV trends merit further attention. A better understanding of these trends requires a thorough model of vertical transmission and childhood progression. For the near future, forecasts of HCV among adolescents will exhibit high variance. Transfusion-induced cases became extremely rare after 1992, but their frequency in prior years is uncertain. HCV prevalence in adolescents in 2005 is therefore uncertain.
HIV-HCV-co-infected patients are another important subgroup deserving further attention, because these patients are being hospitalized for HCV liver-related reasons with increasing frequency. Because of recent changes in treatments for HIV, HCV has taken on a larger role as a competing risk for co-infected patients. Mixed evidence concerning interactions between HIV therapies and HCV outcomes has complicated treatment decisions in the presence of competing risks for co-infected patients.[18] These interactions provide important topics for future research.
Our analysis has some limitations. We must emphasize that the trends found in this study reflect sample-based estimates, rather than comprehensive counts of actual events. In addition, we were unable to track individual patients over time and thus were unable to distinguish between changes in the numbers of hospitalizations per patient from overall hospitalization counts. Also, although HCUP uses rigorous sampling procedures to best represent hospitals throughout the United States, certain sectors are not included, such as the Veterans Affairs system. The data allow for casual observation of related trends; however, because we could not link patient identities between the prescription and HCUP data, we were unable to develop impact estimates. In addition, the prescription figures likely understate the use and costs of drugs for patients with HCV, because the standard of care now includes pegylated interferon.
HCV has aptly been described as a sleeping giant.[8] Our analysis provides descriptive statistics portraying the recent extent and pace of this giant's awakening. Hospitalization statistics indicate increasing takeoff in the use of health care resources by patients with HCV. Combined with predictions of seroprevalence-based models, our age group analysis provides reason to believe that resource use will continue to accelerate in the near future. For hospitalization frequencies, length of stay, charges, and deaths, we find that rates of increase are highest for patients in their 50s. The 40s age group involved the highest magnitudes of these variables, however, indicating an even greater growth potential as middle-aged individuals continue to age. Trends among pediatric HCV patients are relatively more volatile and less informative than for older patients. As a result, the burden of HCV is harder to forecast over the lifetimes of patients younger than 30. To refine our understanding of the HCV outlook, future efforts to quantify pediatric incidence and progression are also needed.
In conclusion, how to best care for the population of HCV-infected patients is an increasingly important topic. For patients who have already developed cirrhosis, combination therapy has improved outcomes, and by 2010, additional benefits may be realized for patients without cirrhosis.[16] Our findings highlight the urgency concerning HCV outcomes. Across the United States, health care providers are using tremendous amounts of resources for HCV care. As the patients continue to age and the disease burden progresses, suboptimal decisions regarding HCV treatments will bring increasing opportunity costs for the health care system and society.
|
|
| |
| |
|
|
|
