|
Cost-effectiveness analysis of lamivudine for the treatment of chronic hepatitis B |
Crowley S J, Tognarini D, Desmond P V, Lees M |
|
|
Record Status This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn. Health technology Drug therapies for the treatment of the hepatitis B virus (HBV). Specifically these included interferon-alpha and lamivudine, a nucleoside analogue.
Economic study type Cost-effectiveness analysis; cost-utility analysis.
Study population A hypothetical, typical population cohort of patients in Australia with chronic HBV. All patients were positive for HBeAg and no patient had cirrhosis. All patients had raised alanine amino transferase (ALT) levels at least twice those of normal. Patients who had previously failed to respond to interferon-alpha treatment were excluded. 70% of patients were assumed to be male, and the average age at the onset of treatment was 30. In Australia the prevalence of HBV has been reported to be highest amongst Asian immigrants, Native Australians, intravenous drug users and homosexual men.
Setting Primary care and community. The economic analysis was conducted in Melbourne, Australia.
Dates to which data relate Short-term effectiveness data were collected from studies published in 1998. Longer term effective data and model transition probabilities were obtained from literature published between 1981 and 1998, as well as from a cancer register with data from 1977 to 1996 on file. Resource data were obtained from consultation with an expert panel, a questionnaire distributed as part of this study and literature published between 1991 and 1998. The base price year used does not appear to have been stated.
Source of effectiveness data Effectiveness data were derived from a review of the literature.
Modelling A decision analysis model was used to synthesise information from the literature and other sources on clinical outcomes (rate of seroconversion, chronic HBV, or development of compensated cirrhosis) and costs in the one year model, for the three scenarios A, B and C. A Markov state transition model was then used to estimate the lifetime costs (maximum of 69 annual cycles) and outcomes associated with the three scenarios. There were six possible health states in the Markov model; seroconverted, chronic HBV, compensated cirrhosis, decompensated cirrhosis, hepatocellular carcinoma and death. Initial probabilities of being in the three states seroconverted, chronic HBV, compensated cirrhosis, were based on the outcomes of the one year decision model. It was assumed that there would be no transition from seroconversion to decompensated cirrhosis or carcinoma, or from compensated cirrhosis to seroconversion or chronic HBV. For the first two years of the Markov model, seroconversion rates were taken from 3 year clinical trials. For patients in arms of the model without lamivudine treatment, an annual rate of natural seroconversion was applied, which was taken from a previously conducted clinical trial. This rate of seroconversion was applied to all arms of the Markov model from the third year of the model onwards. Based on published studies, a rate of 15% seroconversion reactivation in the year after seroconversion was applied.
Outcomes assessed in the review The primary outcomes measured were the efficacy rates of lamivudine, and interferon-alpha in the treatment of chronic HBV. Specifically HBeAg seroconversion rates and progression to cirrhosis were identified. In addition, information on seroconversion reactivation, mortality, development of hepatocellular carcinoma, and decompensated cirrhosis were all obtained from literature for use in the long term model.
Study designs and other criteria for inclusion in the review No specific study designs were reported, although only data from patients in the selected trials with an ALT at least twice the normal level were included. This was deemed to be more representative of the Australian chronic HBV population. Long-term reactivation rates were identified from studies following long term outcomes associated with interferon alpha therapy for patients with chronic hepatitis.
Sources searched to identify primary studies Criteria used to ensure the validity of primary studies Methods used to judge relevance and validity, and for extracting data The judgement criteria were not reported. A table with summary statistics for studies included in the analysis was presented by the authors.
Number of primary studies included 3 studies were identified to assess the efficacy of lamivudine and interferon alpha, all three being of 5 weeks duration. One was a multinational controlled study comparing lamivudine with interferon-alpha, whilst the remaining two studies compared lamivudine with a placebo. One of these was a trial based on the USA only, whilst the other used a three country population (Hong Kong, Singapore, Taiwan). One of these studies, in addition to one year outcomes, reported seroconversion outcomes at 104 and 156 weeks, which were used in the longer term model. 3 studies were identified which examined long-term outcomes associated with the use of interferon therapy. Abstract data from a placebo controlled lamivudine phase II and phase III trial were also used in estimating long term reactivation rates.
Methods of combining primary studies A Meta analysis was used to combine primary studies. The three studies were combined using a weighted average of means, which accounted for sample size and seroconversion rates.
Investigation of differences between primary studies Results of the review Pooling results of the three clinical trials;
the HBeAg seroconversion rates for patients receiving lamivudine were 29.7% (95% CI: 22.4 - 37.1), interferon-alpha were 23.8% (95% CI: 10.9 - 36.7), and placebo 8.6% (95% CI: 2.5 - 14.8).
Similarly the rates of progression to cirrhosis were 2% (95% CI: -0.72 - 4.56) for patients receiving lamivudine, 14% (95% CI: 1.32 - 27.25) for patients receiving interferon-alpha and 14% (95% CI: 4.18 - 22.74) for patients receiving placebo.
As no statistically significant difference was found in seroconversion rates between lamivudine and interferon-alpha, a mean weighted rate between the two drugs of 28.7% was used in the base analysis.
Seroconversion rates at 104 weeks for patients who had not converted at the end of one year were 18.7% and at 156 weeks for patients who had not previously seroconverted this was 39.6%.
The seroconversion rate was reported to be 15%.
Measure of benefits used in the economic analysis The benefit measures were additional seroconversion cases, cirrhosis cases avoided, life years gained and quality adjusted life years gained. A decision analysis model was used in estimating QALYs associated with treatment in the one year model. A Markov model was used to extrapolate short-term QALY data to the long term. Four members of the expert panel using a published Assessment of Quality of Life questionnaire estimated QALYs. The valuation tool used was not reported. Estimates of utility reported in the literature for patients with hepatitis B and C were used in sensitivity analysis.
Direct costs Costs were determined from the perspective of Australian health care providers and, for the one year model, included drug costs, outpatient and specialist room consultations as well as pathology and laboratory tests. Based on common practice in Australia, it was assumed that the interferon-alpha dose was 10 MU given 3 times per week for four months. Withdrawals due to adverse events, poor compliance and early seroconversion were taken into account in determining total dosage. It was assumed that patients receiving placebo would not receive any other medical treatment for the conditions and the costs of adverse events associated with either drug were not considered in the analysis. An expert panel of 6 hepatologists and a questionnaire sent to a further 30 hepatologists were used to identify expected resource use by patients based on current treatment practices and medical management in Australia. The panel and questionnaire were also used to estimate the costs associated with developing compensated cirrhosis. Eligibility to receive lamivudine or interferon-alpha were estimated by the expert panel. Prices of drugs were taken from the 1999 Pharmaceutical Benefits Schedule, and the 1998 Medicare Benefits schedule was used to estimate the cost of specialist consultations, with costs for outpatient specialist consultations taken from a 1996-1997 study in Victoria. The costs of long term hospitalisations were taken from the 1996/97 Australian National Hospital Data Collection. In the long term model costs and benefits were discounted at a rate of 5% per annum.
Indirect Costs Indirect costs were not included.
Currency Australian dollars (Aus$).
Sensitivity analysis A series of one and multi-way threshold sensitivity analyses was undertaken for both the long and short-term models to test for uncertainty in model assumptions. The parameters varied included the number of patients eligible for interferon or lamivudine therapy, the costs and resource use associated with therapy in the first six months, and the upper and lower rates of seroconversion. In the long-term model the assumptions varied included: the low and high values for state transition probabilities; rates of cirrhosis progression; the assumption of no additional seroconversion benefits in years 2 and 3; the high and low values for seroconversion rates; the costs, which were varied by 50% of base value; and the discount rates (0% and 10% discount rates were applied, as well as discounting costs only).
Estimated benefits used in the economic analysis Seroconversion rates in the three scenarios were 24.2% in A, 12.9% in B, and 9% in C and non progression to cirrhosis rates were 94.9% in A, 87.8% in B and 87.3% in C.
In the long term model the lifetime risk of compensated cirrhosis decreased by 5% and the risk of decompensated cirrhosis and hepatocellular carcinoma decreased by 11% for patients under scenario A compared with scenario C.
Average life expectancy under the three scenarios was 28.6 years in A, 24.7 years in B and 24.1 years in C.
The difference in quality adjusted life years gained between patients under scenario A and scenario B was 3.2 QALYs, 3.8 QALYs comparing A with C, and 0.6 QALYs comparing B with C.
Discounted at a rate of 5%, the QALYs gained were 0.98 when comparing A versus B, 1.17 when comparing A versus C and 0.19 B versus C.
Cost results It was assumed that in Scenario A, 90% of patients were eligible for lamivudine and 40% of these patients were eligible for interferon-alpha therapy. 50% of patients ineligible for lamivudine would receive interferon-alpha therapy. 60% of those patients eligible received lamivudine treatment, with 20% receiving interferon-alpha.
Total costs per patient estimates for the one year model were Aus$2,729 for patients treated with lamivudine, Aus$5,682 for patients treated with interferon-alpha and Aus$1,353 for patients treated with placebo.
In Scenario A the proportion of patients treated with lamivudine would be 65% with a further 12% receiving interferon-alpha therapy.
The mean cost per patient in the three scenarios respectively were Aus$2,900 for patients treated with lamivudine, Aus$2,595 for patients treated with interferon-alpha and Aus$1,673 for patients treated with placebo.
The mean costs per seroconversion in the three scenarios were Aus$11,999 for patients treated with lamivudine, Aus$19,512 for patients treated with interferon-alpha and Aus$18,588 for patients treated with placebo.
The mean costs per cirrhosis case avoided were Aus$3,055 for patients treated with lamivudine, Aus$2,875 for patients treated with interferon-alpha and Aus$1,917 for patients treated with placebo.
In the longer term model the annual costs of treatment for seroconverted patients were Aus$363, for those with chronic hepatitis B Aus$726, compensated cirrhosis Aus$2,989, decompensated cirrhosis Aus$13,647 and hepatocellular carcinoma Aus$14,111.
The undiscounted average lifetime costs per person, including first year costs were Aus$41,834 for scenario A, Aus$39,598 for scenario B and Aus$39,047 for scenario C.
Applying a discount rate of 5% these costs were Aus$22,996 for scenario A, Aus$22,276 for scenario B and Aus$21,801 for scenario C.
Synthesis of costs and benefits In the one year model the incremental cost per seroconversion gained in Scenario A compared with Scenario B was Aus$3,341, for B compared with C this was Aus$21,662, and for A versus C Aus$8,089.
Similarly the incremental cost per cirrhosis case avoided was Aus$5,272 for Scenario A compared with Scenario B, Aus$154,442 B compared with C, and Aus$16,003 for A versus C.
QALYs gained were not estimated for the one year model.
In the long term model baseline analysis, when comparing scenario A with B, the incremental cost per life year gained was Aus$633, comparing A with C this was Aus$899, and for B versus C this was Aus$2,470.
Similarly the incremental costs per QALY gained for each of these comparisons were Aus$735, Aus$1,024 and Aus$2,527 respectively.
In sensitivity analysis when comparing scenario A with B, decreasing the number of patients ineligible for interferon-alpha to 20% led to A becoming a dominant scenario with costs lower by Aus$249 per seroconversion;
reducing the number of patients eligible for lamivudine to 50% led to a reduction in cost-effectiveness with an incremental cost-effectiveness ratio of Aus$8,758 per seroconversion;
increasing the seroconversion rate for interferon-alpha to 42% and reducing the lamivudine rate to 18% would increase the cost per seroconversion to Aus$16,311;
inverting these rates would be more cost-effective than the baseline scenario with costs of Aus$ 2,953 per seroconversion gained.
Sensitive parameters in the Markov model included the seroconversion rates in years 2 and 3.
Reducing the one year seroconversion rates from 28.7% to 18% for both drugs led to much higher costs of Aus$41,359 per life year gained compared with Aus$633 in the base case analysis.
Authors' conclusions Lamivudine is a cost-effective treatment for chronic hepatitis patients compared with many currently funded interventions, which, because of its improved tolerability compared with interferon-alpha will allow a greater number of patients to undergo treatment, at reasonable cost. There are, though, many assumptions in the model, which require further examination when longer term outcome data on the drug becomes available.
CRD COMMENTARY - Selection of comparators Different potential scenarios rather than specific interventions were used for comparison, these were justified as they can be observed in real practice, where both lamivudine and interferon-alpha treatment may not be available. Furthermore, the model adopted reflected prescribing reality where a number of individuals would continue to receive interferon-alpha, or no treatment at all. Interferon-alpha is a well recognised current treatment for chronic hepatitis B.
Validity of estimate of measure of benefit The authors did not state that a systematic review if the literature had been undertaken. Effectiveness estimates from primary studies were combined, although the method of pooling does not appear to have been stated. Only data relating to patients with a similar clinical profile to those observed in Australia were used. Benefits were estimated using a decision analytic model for short term outcomes, and a state transition Markov model for lifetime health outcomes. Such an approach was appropriate.
Validity of estimate of costs All categories of cost relevant to the perspective adopted were included in the analysis, and costs were reported separately from quantities. Quantities of resources were obtained from published literature and by expert opinion. These parameters were varied in sensitivity analysis. Costs were obtained from published sources and a sensitivity analysis of these unit costs was conducted. Costs and benefits were discounted in the long term model, this was appropriate and the date to which prices relate was also reported. Both discounted and undiscounted costs were provided in the paper. As noted by the authors themselves, the study was limited, as only direct medical costs were included in the analysis, and other costs (which may be considerable) such as lost productivity, caregiver burden and a reduction in the rate of future infections should be included in future economic analyses.
Other issues The authors did not make direct comparisons with the results of other studies, as the study design adopted was different to that observed in other studies, which had made a direct comparison between lamivudine and interferon-alpha. The authors acknowledged that the study was specifically designed for the Australian setting, but considered that, as hepatitis B in Australia predominantly affects South East Asian immigrants, and treatment protocols in many countries are similar, the results of the study may be valid when substituting local costs for different settings. This is a well thought out study and the authors do recognise a number of limitations, namely that data on efficacy was taken from only 3 recent trials, which as yet do not have long term follow up data. The study may therefore have underestimated the benefits of lamivudine treatment as, in this model, no additional benefits are gained after 3 years, although treatment is continued. They also acknowledge that estimates of quality of life were based solely on the views of four hepatologists, and may not be representative of quality of life perceived by patients.
Implications of the study Within the caveats of this study, lamivudine is expected to reduce and delay the progression of chronic hepatitis B, increase life expectancy and quality of life, for a modest overall increase in healthcare costs.
Source of funding Funded by GlaxoWellcome Australia Limited.
Bibliographic details Crowley S J, Tognarini D, Desmond P V, Lees M. Cost-effectiveness analysis of lamivudine for the treatment of chronic hepatitis B. PharmacoEconomics 2000; 17(5): 409-427 Indexing Status Subject indexing assigned by NLM MeSH Antiviral Agents /therapeutic use; Cost-Benefit Analysis; Hepatitis B, Chronic /drug therapy; Humans; Interferon-alpha /therapeutic use; Lamivudine /therapeutic use AccessionNumber 22000008217 Date bibliographic record published 31/03/2001 Date abstract record published 31/03/2001 |
|
|
|