SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Authorship
  8. Acknowledgement
  9. References
  10. Supporting Information

Background

As baby boomers age, chronic hepatitis C (CHC) will become increasingly important in Medicare eligible group.

Aim

To evaluate trends in Medicare resource utilisation for CHC.

Methods

We analysed the Medicare in-patient and out-patient data from 2005 to 2010. For each patient, all claims with CHC as a principal diagnosis were added up and yearly CHC-related spending was calculated.

Results

A total of 48 880 out-patient claims for 21 655 CHC patients and 4884 hospital admission claims for 3092 patients were included. The number of in-patient (1.5–1.6/year) or out-patient (2.2–2.3/year) visits per patient did not change over time, nor did the demographic characteristics of the CHC population. The majority of this population was eligible for Medicare based on disability and the average number of diagnoses per in-patient claim (from 8.11 in 2005 to 8.60 in 2010) and per out-patient claim (from 2.18 in 2005 to 2.71 in 2010) increased (both P < 0.0001). The average total yearly spending per patient increased in the out-patient setting from $488 in 2005 to $584 in 2010 (P = 0.0132) and did not change in the in-patient setting (from $22 245 in 2005 to $23 383 in 2010, P = 0.14). In the multivariate analysis, the number of diagnoses and conditions per claim and the number of in-patient or out-patient procedures per year were the important independent predictors of increased resource utilisation.

Conclusions

Most Medicare beneficiaries with chronic hepatitis C who sought in-patient or out-patient care in 2005–2010 had received Medicare for disability. Although the total resource utilisation did not change, the proportion of patient's responsibility increased.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Authorship
  8. Acknowledgement
  9. References
  10. Supporting Information

The Centers for Disease Control and Prevention (CDC) estimates that approximately 3.2 million people in the United States are currently living with chronic hepatitis C virus (CHC) infection.[1] Hepatitis C is known to progress to advanced liver disease, cirrhosis and hepatocellular carcinoma (HCC).[2, 3] Furthermore, the rates of these complications have risen over the past years. Kanwal and colleagues recently studied a large cohort of patients from the Veterans Administration diagnosed with CHC and found that the prevalence of cirrhosis and decompensated cirrhosis had more than doubled from 1996 to 2006, and that the prevalence of HCC increased approximately 20-fold.[4] In another recent study, Davis and Roberts have postulated that as the individuals who are currently 50–65 years old (so called ‘baby boomers’) age, complications of CHC such as hepatic failure and hepatocellular carcinoma may become even more prevalent.[5]

The increase in disease progression has already led to a substantial increase in CHC-associated mortality.[6] Using U.S. census and multiple causes of death data, Wise and colleagues found that from 1995 to 2004, hepatitis C mortality rates increased 123% with the highest death rates noted for the time period from 1995 to 1999, among non-Hispanic blacks, males and those aged 55–64 years.[6]

In addition to mortality, patients with CHC suffer from substantial morbidity and poorer health-related quality of life (HRQL) when compared with healthy controls.[7-9] In a systematic review of HRQL, investigators reported that the vitality subscale of the SF-36 was the category most affected by patients living with HCV.[7] In a more recent study using the SF-36, a liver disease-specific instrument (Chronic Liver disease Questionnaire, CLDQ) and a fatigue questionnaire (Chronic Fatigue Screener, CFS) to assess features that may be associated with HRQL, investigators reported that, in fact, the main contributor to HCV patients' reported lower HRQL scores was extreme fatigue-related loss of activity.[8]

The 2008 Dartmouth Atlas of chronic diseases lists chronic liver disease (CLD) as one of the nine chronic diseases that account for the most Medicare dollars spent at the end of one's life.[10] Given that CHC is the most common cause of chronic liver disease[1], HCV-related complications will remain an important aspect of healthcare resource utilisation. The majority of the Medicare costs were sustained during the end-of-life in-patient care although over 50% of the patients wanted to die at home.[10] Value-based purchasing and reimbursement reduction, the phrases used for Medicare's new reimbursement method which started in October of 2012, links hospital payments to the quality of care that patients receive and penalises hospitals that do not deliver care that is on par with their peers.[11-16] These programmes are structured so that unnecessary care is not rewarded. Given this perspective, it is especially important to gain an understanding of Medicare's resource utilisation trends for patients living with HCV so that as Medicare's new reimbursement system expands, interventions can be developed to control spending and to eliminate unnecessary, inappropriate or duplicate services.[13, 16] The study purpose is to evaluate recent trends (2005–2010) and predictors of Medicare's major resource utilisation parameters including total charges, total payments, Medicare spending and the proportion of patient's responsibility for in-patient and out-patient visits for patients living with CHC.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Authorship
  8. Acknowledgement
  9. References
  10. Supporting Information

Data source

In this study, we performed a retrospective analysis of the Medicare in-patient and out-patient files with final action fee-for-service claims submitted by out-patient healthcare providers and by in-patient hospital providers for reimbursement of facility costs. For the purpose of this study, a 5% random sample of Medicare beneficiaries who received in-patient or out-patient care in 2005–2010 were provided to us by CMS, and included records with 05, 20, 45, 70 or 95 in a given position of the beneficiary's unique Medicare identifier.

In the out-patient file, each claim included a unique patient identifier (for de-identification purposes, it was different from the beneficiary's Medicare ID), the last day of services rendered to the beneficiary included in the claim (only quarter and year in the encrypted files used for this study), a principal diagnosis and up to nine secondary diagnoses, up to six out-patient procedures, total charges, Medicare reimbursement amount and payments from the patient and another insurance provider when applicable. The denominator file provided to us by CMS for the sample included Medicare eligibility category (age, disability and/or end-stage renal disease (ESRD), demographics (age, gender, ethnicity) and residence (state and county) data for each sampled participant.

Similarly, the in-patient file contained in-patient hospital claims, each accompanied by a unique patient identifier, up to six diagnoses and ten in-patient procedure codes, dates of service (quarter and year), length of stay in days, hospital charges, Medicare reimbursement amount, and payments from the patient and another insurance provider when applicable. All in-patient, out-patient and denominator files have been stripped of data elements that might allow identification of the sampled participants. The study was approved by Inova Institutional Review Board, and a data-use agreement with the CMS was signed.

Study population

In both the in-patient and out-patient parts of the study, the ICD-9 codes 070.41, 070.44, 070.51, 070.54, 070.70, 070.71 were used to establish the diagnosis of hepatitis C.[17] A claim was included in the study if principal or any of the secondary diagnoses for that claim was hepatitis C. If hepatitis C was listed as a secondary diagnosis, the claim was included in the study only if principal diagnosis indicated the presence of chronic liver disease (which might have been the diagnosis of advanced liver disease such as liver cancer or cirrhosis, or complications known to be specific to cirrhosis such as oesophageal varices or ascites; the complete list of ICD-9 codes used for establishing the diagnosis of CHC is included in Table S1). A patient might have had more than one in-patient or out-patient claim per year. In such cases, claims with a principal diagnosis other than hepatitis C or CHC were not included even if that patient had hepatitis C as could be seen from other claims. Therefore, only hepatitis C-related spending was studied.

Outcomes

In this study, we evaluated major resource utilisation parameters including total charges, total payments, Medicare spending and the proportion of patient's responsibility. The total per-claim payments were the sum of Medicare reimbursement amount, primary insurance payment, and the patient's responsibility which included all applicable co-payments, deductibles and coinsurance. For the in-patient portion, we also studied in-hospital mortality and length of stay (LOS), which was defined as the number of days, both chargeable and not chargeable to Medicare facility utilisation, of care in each claim. If more than one in-patient or out-patient claim was reported for a patient in a given year, then, for that patient in that year, the resource utilisation parameters were added up, and the total yearly resource utilisation, together with the average proportion of patient's responsibility in per cent of total payments, was calculated. All charges and total costs were adjusted to the 2010 dollars using Consumer Price Index values for the study years[18].

Statistical analysis

We calculated the demographic and clinical parameters of the study cohort and compared them across the study years using a nonparametric Kruskal–Wallis test for continuous variables (such as hospital charges or the proportion patient's responsibility) or chi-square test for ordinal, nominal and binary variables (such as gender or Medicare eligibility category). Here and below, P value of 0.05 or less was considered statistically significant.

Multiple linear regression was used to identify independent predictors of the total payments and patient's responsibility after logarithmic transformation of the respective outcomes. Beta coefficients with the respective 95% confidence intervals from these models were further exponentiated, and a percentage change in the outcomes associated with each predictor was calculated. Independent predictors of in-patient mortality and the risk of in-patient expenditures in recipients of out-patient care were studied using multiple logistic regression. All analyses were performed using SAS 9.1 (SAS Institute Inc., Cary, NC, USA).

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Authorship
  8. Acknowledgement
  9. References
  10. Supporting Information

Resource utilisation and mortality of CHC patients in the in-patient setting

A total of 4884 CHC-related in-patient claims for 3092 unique Medicare beneficiaries with hepatitis C were included for the six study years. The number of patients with at least one claim related to their CHC ranged from the minimum of 477 in 2008 to the maximum of 538 in 2007 and 2009, representing approximately 9500–10 800 Medicare beneficiaries with hepatitis C nationwide. The average number of claims per patient was between 1.53 and 1.64 (P > 0.05 for the yearly changes).

The proportion of CHC-related claims among all in-patient claims where the presence of hepatitis C had been recorded as the main reason for admission was 11.3–13.2% (P = 0.0065); thus, approximately 85–90% of claims for Medicare beneficiaries with hepatitis C were not related to their HCV infection. The most common primary diagnoses for the CHC-related claims included to the study were liver cirrhosis (32.15%), hepatic coma (37.03%), primary liver cancer (7.39%), the presence of HCV infection (8.25%), and ascites (3.76%). Of all in-patient claims, the primary or secondary diagnosis of liver cirrhosis was mentioned in 78.89%, primary liver cancer – 11.92%, hepatic coma – 49.02%.

The proportion of Medicare beneficiaries younger than 65 years of age increased from 66.2% to 73.4% (P = 0.0061), and similarly increased the proportion of Medicare individuals with disability (from 63.1% to 68.1%) while the proportion of male patients with hepatitis C and the racial/ethnic distribution did not change. Approximately 10% of patients were from the Northeast region, 18% from Midwest, 40% from South, 10% from West and 12% from California separately (Table 1). This regional distribution remained relatively stable during the study years.

Table 1. Demographic characteristics of Medicare beneficiaries with hepatitis C who sought liver disease-related in-patient care in 2005–2010
Parameter200520062007200820092010P
Number of patients512497538477538530 
Number of claims781778876730852867 
Proportion of claims for CLD among all claims, %11.2811.7312.4812.2413.1912.940.0065
Male, %63.6762.9863.9463.9763.5765.850.956
Age
Age <65, %66.2165.7966.9167.6773.7973.40.0061
Age 65–69, %14.8412.0715.2411.5510.9712.450.203
Age 70–74, %7.428.656.699.936.696.420.260
Age 75–79, %7.627.045.955.312.63.770.0019
Age 80–84, %2.343.623.163.934.281.890.204
Age 85 + , %1.562.822.041.621.672.080.722
Race/ethnicity
White, %69.9469.7670.717573.7472.430.614
Black, %18.2716.3315.314.2915.0816.730.738
Other race, %2.553.233.173.572.792.660.977
Asian, %1.772.822.431.191.681.520.621
Hispanic, %6.687.467.655.955.595.890.712
Native American, %0.790.40.7501.120.760.786
Medicare eligibility category
Aged only, %33.233.232.3429.7725.0926.60.0087
Aged + ESRD, %0.591.010.742.11.1200.0196
Disability only, %63.0961.9763.5763.169.5268.110.0486
Disability+ESRD, %1.172.212.043.562.043.210.141
ESRD only, %1.951.611.31.472.232.080.845
Location
Northeast, %19.4918.719.9626.192022.640.463
Midwest, %18.718.2916.7920.2419.4415.280.499
South, %41.1441.2639.1833.3338.537.550.609
West, %10.248.5411.018.338.7911.890.415
California, %10.4313.2113.0611.913.2712.640.748

The average total yearly payments for in-patient claims for patients with hepatitis C remained the same at the level of $19 800–24 000 (P > 0.05). However, the average proportion of patient's responsibility dropped in 2006 to the lowest 6.5% ($814 ± $864 per year in the 2010 dollars) and then increased to 8.88% ($995 ± $1849) in 2008 (both P < 0.05). The total charges submitted to Medicare by hospitals, however, increased from approximately $52 700 per year in 2006 to approximately $76 200 in 2010 (P < 0.0001) in inflation-adjusted dollars, while accompanied by no change in the total yearly length of stay (8.66–9.50 days, P = 0.54) or the number of in-patient procedures performed during that yearly stay (2.52–2.74, P = 0.84) (Table 2).

Table 2. Liver-related in-patient resource utilisation for Medicare beneficiaries with hepatitis C (mean ± standard deviation)
Parameter200520062007200820092010 P
Average number of claims per year1.53 ± 1.201.57 ± 1.301.63 ± 1.311.53 ± 1.241.58 ± 1.181.64 ± 1.380.601
Total payments, $22 245.12 ± 30 415.9920 364.51 ± 24 586.1121 602.99 ± 30 667.9219 846.29 ± 30 858.6024 036.94 ± 34 978.0223 383.12 ± 32 951.070.136
Total payment by patient, $975.25 ± 1714.72814.50 ± 863.97873.81 ± 1200.33995.26 ± 1849.72955.15 ± 1116.621082.71 ± 2180.10<.0001
Total payment by Medicare, $20 084.35 ± 28 979.2018 192.10 ± 21 355.3217 791.33 ± 22 079.8517 282.50 ± 26 181.3919 759.38 ± 28 959.1720 914.41 ± 28 644.060.1165
Total charge, $64 417.66 ± 136 481.3252 746.15 ± 66 107.2259 026.68 ± 74 749.6555 789.63 ± 84 867.3773 565.44 ± 131 738.4976 195.40 ± 120 197.570.0017
Patient's responsibility,%6.79 ± 6.516.50 ± 6.597.13 ± 9.028.88 ± 12.047.89 ± 9.497.38 ± 8.390.0245
Medicare's responsibility,%89.55 ± 18.4588.83 ± 20.9287.64 ± 22.6486.78 ± 22.3586.77 ± 22.6388.82 ± 19.690.0304
Average number of diagnoses per claim8.11 ± 1.368.17 ± 1.358.33 ± 1.338.40 ± 1.208.53 ± 1.178.60 ± 1.19<.0001
Average number of related conditions per claim1.37 ± 0.971.35 ± 1.061.36 ± 1.011.36 ± 1.001.48 ± 1.101.36 ± 1.040.488
Total number of in-patient procedures2.61 ± 3.322.52 ± 2.782.72 ± 3.042.55 ± 3.052.62 ± 2.902.74 ± 3.200.835
Total length of stay, days9.29 ± 11.748.66 ± 8.589.17 ± 9.768.42 ± 10.608.90 ± 9.899.50 ± 11.510.536

In multivariate analysis (Table 3), most of the demographic parameters with the exception of Hispanic ethnicity [payment increase (95% CI) = +25.17% (6.41–47.24%), LoS increase = +29.04% (14.54–45.37%)] were not independently associated with resource utilisation (Table 3). Longer in-hospital stay, but not higher costs, was associated with more diagnoses per record (+12.43% (9.90–15.01%) longer stay per each additional diagnosis). Overall, the number of related conditions per claim and the total number of in-patient procedures per year were the major predictors of the increased resource utilisation: +11.94% (7.63–16.42%) to total spending and +2.81% (0.11–5.81%) to the length of stay per each additional condition; +18.68% (17.14–20.24%) and +17.48% (16.36–18.61%) per each additional procedure. The proportion of patient's responsibility was higher in Midwest and South [+18.45% (2.85–36.42%) and +18.68% (5.21–33.87%) when compared with Northeast respectively] and lower in the setting of more conditions and in-patient procedures (−8.22% (−12.13 to −4.14%) and −6.94% (−8.28% to −5.58%) per condition and procedure respectively). All potential predictors and confounders for in-patient resource utilisation for Medicare beneficiaries with hepatitis C infection are listed in Table 3.

Table 3. Independent predictors of in-patient resource utilisation and in-hospital mortality for Medicare beneficiaries with hepatitis C
predictorTotal payments increase,% (95% CI)Proportion of patient's responsibility increase,% (95% CI)Total length of stay increase,% (95% CI)In-patient mortality
Calendar year−0.62 (−2.75 to +1.57)+1.61 (−0.81 to +4.09)−1.67 (−3.23 to −0.09)0.846 (0.783–0.914)
Age 65–69−8.47 (−37.80 to +34.68)−15.45 (−44.91 to +29.76)+9.23 (−17.73 to +45.03)0.358 (0.078–1.642)
Age 70–74−17.41 (−44.65 to +23.22)−24.81 (−51.75 to +17.19)+9.93 (−18.05 to +47.45)0.292 (0.061–1.394)
Age 75–79−7.39 (−38.61 to +39.72)−4.43 (−39.42 to +50.79)+6.12 (−21.53 to +43.50)0.583 (0.120–2.831)
Age 80–84−9.87 (−41.82 to +39.61)−27.78 (−55.55 to +17.33)+29.34 (−6.19 to +78.32)0.368 (0.067–2.028)
Age 85+−5.91 (−41.27 to +50.75)−34.91 (−61.41 to +9.77)+23.73 (−12.45 to +74.86)0.315 (0.051–1.962)
Male+9.44 (+0.65 to +18.99)+2.80 (−6.32 to +12.79)−11.51 (−16.78 to −5.90)1.328 (0.982–1.797)
Black+3.62 (−7.20 to +15.69)+0.46 (−11.10 to +13.52)−1.80 (-9.43 - +6.48)1.059 (0.731–1.536)
Hispanic+25.17 (+6.41 to +47.24)−9.38 (−24.31 to +8.49)+29.04 (+14.54 to +45.37)0.849 (0.480–1.502)
Asian+31.00 (−1.64 to +74.48)−20.51 (−42.15 to +9.22)−6.91 (−24.57 to +14.88)0.776 (0.297–2.028)
Disability−5.08 (−34.74 to +38.07)−14.19 (−43.36 to +30.01)+16.29 (−11.67 to +53.09)0.202 (0.045–0.910)
ESRD−36.05 (−49.75 to −18.62)−19.74 (−38.56 to +4.86)−18.76 (−31.94 to −3.04)0.256 (0.079-0.834)
Location: Midwest−6.18 (−17.40 to +6.56)+18.45 (+2.85 to +36.42)−10.48 (−18.47 to −1.71)0.776 (0.507–1.187)
Location: South−7.07 (−16.64 to +3.60)+18.68 (+5.21 to +33.87)−6.42 (−13.59 to +1.35)0.777 (0.540–1.117)
Location: West−8.71 (−21.76 to +6.51)+12.27 (−5.38 to +33.22)−21.43 (−29.84 to −12.02)0.829 (0.477–1.438)
Location: California

+1.37 to

(−12.03 to +16.81)

+2.32 to

(−12.57 to +19.74)

-17.97

(−26.08 to −8.98)

0.953

(0.601-1.511)

Number of diagnoses per claim, per dx.+1.36 (−1.73 to +4.54)−3.08 (−6.34 to +0.31)+12.43 (+9.90 to +15.01)1.477 (1.241–1.757)
Number of conditions per claim, per cond.+11.94 (+7.63 to +16.42)−8.22 (−12.13 to −4.14)+2.81 (−0.11 to +5.81)1.097 (0.962–1.250)
Number of procedures per year, per proc.+18.68 (+17.14 to +20.24)−6.94 (−8.28 to −5.58)+17.48 (+16.36 to +18.61)1.446 (1.344–1.555)
Total payment, per $1000NANANA0.999 (0.995–1.004)

In-hospital mortality was found to decrease with time: OR (95% CI) = 0.846 (0.782–0.914) per calendar year in comparison with the reference year of 2005 (Table 3). The presence of disability and ESRD were found to be associated with lower mortality; however, in the dataset used for this study, disability for individuals older than 65 was not reported. Of other potential predictors of in-patient mortality, only the total number of diagnoses per claim was found to be associated with higher mortality (OR = 1.477 (1.241–1.757) per dx), most likely being a surrogate marker of more severely diseased patients. Finally, more procedures, but not higher or lower spending, were associated with higher mortality: OR = 1.446 (1.344–1.555) per procedure and 0.999 (0.995–1.004) per $1000 respectively (Table 3).

Resource utilisation of chronic hepatitis C in the out-patient setting

Between 2005 and 2010, a total of 48 880 out-patient CHC-related claims for 21 655 unique Medicare beneficiaries with hepatitis C were included (from 7755 in 2006 to 8649 in 2010) representing approximately 67 000–80 000 Medicare beneficiaries with hepatitis C nationwide. The average number of claims per patient was between 2.16 and 2.32 (P > 0.05 for the yearly changes). Unlike in-patient, the proportion of CHC-related out-patient claims among all claims submitted for Medicare beneficiaries with hepatitis C increased significantly: from 49.30% in 2005 to 56.90% in 2010 (P < 0.0001) (Table 4). The most common primary diagnoses for these out-patient claims was HCV infection (84.53%), cirrhosis (5.92%), ascites (1.82%), primary liver cancer (1.23%) and history of liver transplant (1.16%).

Table 4. Demographic characteristics of Medicare beneficiaries with hepatitis C who sought liver disease-related out-patient care in 2005–2010
Parameter200520062007200820092010 P
Number of patients334733663507358438444007 
Number of claims777877557804831085848649 
Proportion of claims for CLD among all claims, %49.3050.5352.1652.3254.0956.90<0.0001
Male, %56.557.157.2356.2656.7458.30.5498
Age
Age <65, %69.6769.7970.0971.5571.3671.570.2116
Age 65–69, %13.3313.71313.3213.6813.60.9512
Age 70–74, %8.468.268.137.347.317.190.1659
Age 75–79, %5.084.95.254.554.474.440.463
Age 80–84, %2.392.442.482.032.292.350.8645
Age 85 + , %1.080.921.061.210.880.850.6253
Race/ethnicity
White, %68.7470.1168.526568.0868.660.3434
Black, %19.418.920.6924.2921.5922.110.002
Other race, %4.133.483.293.212.772.15<.0001
Asian, %2.341.822.322.142.352.10.6385
Hispanic, %4.374.54.122.864.123.830.6
Native American, %1.021.191.062.51.11.150.3668
Medicare eligibility category
Aged only, %3029.7729.428.0127.9727.680.0969
Aged + ESRD, %0.330.450.510.50.680.750.1572
Disability only, %65.7966.0765.9567.4967.2568.160.1612
Disability + ESRD,%2.061.92.542.542.392.120.3369
ESRD only, %1.821.811.61.451.721.30.3976
Location
Northeast, %22.322.5921.8321.4822.7422.050.9359
Midwest, %18.3118.919.9219.7220.2221.320.0265
South, %35.2435.634.2430.6332.6333.680.0556
West, %10.949.5810.099.5110.4610.110.5678
California, %13.2213.3313.9218.6613.9512.830.0917

Of demographic characteristics of the target cohort, the only significant change was the increased proportion of African American patients (from 19.4% in 2005 to highest 24.3% in 2008, = 0.0020) (Table 4). At the same time, the proportion of different age groups and Medicare eligibility categories remained the same (all P > 0.05) (Table 4). The proportion of patients from Midwest region increased from 18.3% in 2005 to 21.3% in 2010 (P = 0.0265), while balanced by a similar decrease in the proportion of patients from South (from 35.2% in 2005 to the lowest 30.6% in 2008, P = 0.0556) (Table 5). There was also an unexplained sudden change in the proportion of patients from California in 2008 (18.7%, vs. the lowest 12.8% in 2010 and similar proportions in other years).

Table 5. Liver-related out-patient resource utilisation for Medicare beneficiaries with hepatitis C (mean ± standard deviation)
Parameter200520062007200820092010 P
Average number of claims per year2.32 ± 2.682.30 ± 2.612.23 ± 2.542.32 ± 2.712.23 ± 2.632.16 ± 2.250.2963
Total payments, $488.06 ± 1027.65494.32 ± 866.32515.73 ± 1112.56542.49 ± 1162.73566.23 ± 1093.03584.38 ± 1485.250.0132
Total payment by patient, $104.63 ± 231.31100.17 ± 188.21108.35 ± 249.5996.95 ± 199.8399.70 ± 191.13103.93 ± 234.880.0218
Total payment by Medicare, $360.11 ± 782.03372.56 ± 657.74388.97 ± 857.39414.51 ± 893.05445.07 ± 901.64456.61 ± 1,248.160.0022
Total charge, $2,293.64 ± 6,041.302,420.06 ± 4,962.882,573.16 ± 5,857.452,693.79 ± 7,165.652,898.19 ± 5,712.233,144.72 ± 8,047.82<.0001
Patient's responsibility,%17.61 ± 17.4517.57 ± 17.2218.53 ± 17.8816.52 ± 17.1116.53 ± 16.3517.35 ± 17.67<.0001
Medicare's responsibility,%79.87 ± 21.6880.27 ± 20.9579.53 ± 21.1281.31 ± 20.9781.05 ± 20.8180.37 ± 21.49<.0001
Average number of diagnoses per claim2.18 ± 1.472.33 ± 1.542.42 ± 1.602.45 ± 1.612.63 ± 1.782.71 ± 1.85<.0001
Average number of related conditions per claim0.33 ± 0.540.34 ± 0.550.32 ± 0.550.36 ± 0.570.39 ± 0.580.39 ± 0.58<.0001

Of the resource utilisation parameters (Table 5), the most substantial change was observed for the out-patient service charges: from $2293 per year in 2005 to $3144 in 2010 (P < 0.0001). The total payments, however, changed by a much smaller magnitude (from $488 to $584, P = 0.0132), accompanied by moderate oscillations in patient's out-of-pocket share ($105 or 17.6% in 2005 to lowest $97–$100 or 16.5% in 2008–2009 and back to $104 or 17.4% in 2010, P < 0.05). However, the clinical presentation of a Medicare beneficiary with hepatitis C changed substantially: the average number of diagnoses per claim increased from 2.18 in 2005 to 2.71 in 2010, and the average number of related conditions per claim increased from 0.33 to 0.39 (both P < 0.0001).

In multivariate analysis, after adjustment for the changes in demographics, neither total costs of out-patient treatment of patients with hepatitis C, nor the proportion of out-of-pocket spending changed during the study span (P > 0.05). Older age was associated with both lower out-patient spending and smaller proportion of patient's share (Table 6). Ethnicity was associated with higher payments by both Medicare and the patient (Hispanics: +12.76% (2.44–24.13%) and +37.94% (23.79–53.71%); Asians: +24.89% (9.45–42.50%) and +37.65% (18.63–59.73%), respectively; African Americans – no association with spending, +36.13% (28.88–43.78%) to patient's share), all significant even after adjustment for location that might possibly suggest certain ethnic disparity. Lower spending accompanied by slightly higher patient's share was observed the in South region (−8.06% (−12.69 to −3.18%) and +6.52% (0.49–12.92%) respectively) while significantly higher proportion of patient's share from the reference Northeast region was found in the West region and California separately: +16.82% (7.73–26.67%) and +26.63% (17.58–36.37%), respectively. Finally, more diagnoses and conditions per out-patient claim were associated with higher costs from both Medicare and patient's side: +19.24% (17.89–20.60%) per dx and +22.04 (18.01–26.21%) per condition to total spending, +14.66 (13.19–16.14%) and +4.90% (1.00–8.95%) to patient's share respectively) (Table 6).

Table 6. Independent predictors of out-patient resource utilisation and the risk of hospitalisation for Medicare beneficiaries with hepatitis C
PredictorTotal payments increase,% (95% CI)Proportion of patient's responsibility increase,% (95% CI)OR of being hospitalised in a year of out-patient claim (95% CI)
Calendar year−0.49 (−1.50 to +0.53)−0.89 (−2.02 to +0.26)0.983 (0.950–1.017)
Number of out-patient claims in a yearNaNa1.119 (1.100–1.138)
Age 65–69+1.51 (−16.24 to +23.02)−1.96 (−21.07 to +21.78)0.735 (0.361–1.500)
Age 70–74−12.30 (−28.05 to +6.88)−2.11 (−21.69 to +22.37)0.726 (0.348–1.514)
Age 75–79−24.49 (−38.48 to −7.30)−16.46 (−33.71 to +5.28)0.807 (0.381–1.712)
Age 80–84−40.85 (−52.67 to −26.07)−32.65 (−47.63 to −13.38)0.594 (0.255–1.385)
Age 85+−42.25 (−55.88 to −24.42)−40.26 (−55.90 to −19.07)1.072 (0.426–2.697)
Male+1.63 (−2.22 to +5.64)+9.91 (+5.22 to +14.81)1.208 (1.059–1.377)
Black+3.58 (−1.33 to +8.72)+36.13 (+28.88 to +43.78)0.609 (0.505–0.735)
Hispanic+12.79 (+2.47 to +24.15)+37.94 (+23.79 to +53.71)1.554 (1.184–2.041)
Asian+24.89 (+9.46 to +42.50)+37.65 (+18.63 to +59.73)0.950 (0.606–1.490)
Disability+12.18 (−6.87 to +35.12)−5.39 (−23.31 to +16.71)0.910 (0.456–1.818)
ESRD−0.94 (−11.82 to +11.29)−18.46 (−28.49 to −7.02)0.727 (0.476–1.111)
Location: Midwest−2.94 (−8.45 to +2.90)−4.38 (−10.48 to +2.13)1.121 (0.912–1.377)
Location: South−8.06 (−12.69 to −3.18)+6.52 (+0.49 to +12.92)1.256 (1.048–1.505)
Location: West−0.09 (−7.01 to +7.34)+16.82 (+7.73 to +26.67)1.091 (0.857–1.388)
Location: California+5.70 (−1.03 to +12.88)+26.63 (+17.58 to +36.37)1.050 (0.835–1.320)
Number of diagnoses per claim, per dx.+19.24 (+17.89 to +20.60)+14.66 (+13.19 to +16.14)1.428 (1.387–1.469)
Number of conditions per claim, per cond.+22.04 (+18.01 to +26.21)+4.90 (+1.00 to +8.95)1.127 (1.015–1.251)

We also evaluated the risks of being hospitalised and incur in-patient spending in patients who sought out-patient care in the study years. The CHC patients were more likely to be admitted to a hospital at least once the same year they received out-patient care if they were treated as an out-patient more than once [OR (95% CI) = 1.119 (1.100–1.138) per additional claim], if they were male [OR = 1.208 (1.059–1.377)], Hispanic [OR = 1.554 (1.184–2.041)] but not African American [OR = 0.609 (0.505–0.735)], if they lived in the South [OR = 1.256 (1.048–1.505)] and if they had more diagnoses per claim in that year [OR = 1.428 (1.387–1.469) per each additional diagnosis] (Table 6).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Authorship
  8. Acknowledgement
  9. References
  10. Supporting Information

The study is the first to assess the impact of CHC on the Medicare population. We found that during the study years (2005–2010), there were approximately 9500–10 800 Medicare beneficiaries seen yearly in the in-patient setting for their CHC, as well as 67 000–80 000 CHC patients who received care in out-patient setting.

While gender and ethnic distribution of Medicare beneficiaries with CHC stayed relatively stable over time, the proportion of Medicare beneficiaries with CHC younger than 65 years old increased substantially. So far, individuals younger than 65 are eligible for Medicare only in case of disability, but as this cohort ages, this may have important implications for Medicare resource utilisation in the near future. Specifically, several reports suggested that as the baby boomers aged, there would be an increase in of HCV in the patient population 65 years old and greater.[5, 19] This supposition was based on the thought that the majority of patients with HCV acquired their infections in the 1970s when they were young adults; therefore, as they aged, there would be a corresponding increase in the complications of HCV-related chronic liver disease.[5, 19] As this group becomes Medicare eligible over the next decade, their impact on Medicare resource utilisation will become enormous.

Analysis of the CHC in-patient hospital costs demonstrated no trending pattern overall. The average total yearly payments stayed the same while the patients' monetary responsibility dropped in 2006, then rose to its highest point in 2008, then dropped again. The one consistent upward trend was the total charges submitted to Medicare from hospitals despite there being no changes in patient length of stay or in the number of in-patient procedures performed during the patient stay.

There was regional variation in the proportion of the patient's responsibility, which was higher in the Midwest and the South. Regional variation is not a new finding. In the Dartmouth Atlas of Chronic Illness project,[10] the investigators determined that there was a wide variation in Medicare spending across the country after matching similar patients. However, they found that only 4% of the variation in Medicare spending was accounted for by regional variation. Instead, the major contributor to the variation was due to the supply of the services deemed necessary for care. Furthermore, the care practitioners changed their practices based on the resources available, so areas with more resources had more procedures incurred and less resources led to fewer procedures. They found as well that more care did not necessarily improve patient outcomes and could actually increase their risk of dying.[10] Razavi, Menzin and Manos et al. also found that hospital stay costs were rapidly rising due to the use of more supplies and devices without necessarily better patient outcomes.[20-22] Menzin and Manos also found that Medicare costs varied by region and the number of comorbidities.[21, 22]

The multivariate analysis results determined that the number of related conditions per claim and the total number of out-patient procedures per year were the major predictors of increased resource utilisation. In addition, the total number of diagnosis per claim and more procedures performed per hospitalisations were associated with higher mortality (OR = 1.446; 95% CI: 1.3444–1.555). All these are not unexpected findings, but do help establish further that it is necessary to continuously review care procedures to determine which ones provide the most value to patients.[20-22]

Unlike in-patient, the number of CHC-related claims submitted to Medicare for patients with HCV who sought care in out-patient setting rose significantly from 49.3% of all out-patient claims in 2005 to 56.9% in 2010. The average number of out-patient claims (2.2–2.3/year) also exceeded the number of in-patient claims (1.5–1.6/year). This suggests that much of the care for the Medicare patient with CHC is being delivered in the out-patient setting.

The study is not free from limitations. Specifically, as the majority of those with CHC are younger than 65 years of age and have multiple disabilities, many of these patients could be considered dual eligible for both Medicare and Medicaid coverage. There have been a few reports that have found that in states with lower Medicaid spending, there was a corresponding higher Medicare spending[23, 24]. This spending pattern may, in fact, be biased due to differences in proportions of those underserved by their states' Medicaid programmes. Furthermore, we had limited access to patients' demographic characteristics and could not link them to respective spending, thus, we were unable highlight the directions for future dynamics of resource utilisation, given the ongoing changes in the U.S. population. Also, the nature of billing data warranted the use of bottom-up approach to estimating the total economic burden on Medicare associated with CHC. However, as a general limitation of this approach, it results in underestimation of that burden, because all costs that are not covered as in-patient or out-patient claims (such as costs of pharmacy, hospice, skilled nursing or other long-term care facility) were not accounted for in this study.

In summary, our study demonstrated that the number of Medicare patients living with HCV and the number of Medicare claims for CHC did not significantly change over the study period. The majority of patients in this cohort remain younger than 65 years and had multiple other disabilities. This indicates that the potential economic impact of CHC on Medicare has not been fully realised. Also, the total charges for managing CHC did significantly increase over time with the patients' payment responsibility increasing through 2008. This report of Medicare data supports efforts directed towards more aggressive identification and treatment of CHC patients who are younger than 65 years who are experiencing other disabilities. In addition, further study should be undertaken to understand why the total charges submitted to Medicare increased for in-patient stays over time without a subsequent increase in length of stay or number of procedures performed.

Authorship

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Authorship
  8. Acknowledgement
  9. References
  10. Supporting Information

Guarantor of the article: Zobair M Younossi, MD, MPH.

Author contributions: ZY designed the research study, MS collected and analyzed the data, AM and CV contributed to the design of the study, LH and SH contributed to the manuscript writing. All authors approved the final version of the manuscript.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Authorship
  8. Acknowledgement
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Authorship
  8. Acknowledgement
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
apt12485-sup-0001-TableS1.docxWord document13KTable S1. ICD-9 codes for the diagnosis of CLD.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.