|A pharmacoeconomic evaluation of zuclopenthixol compared with haloperidol and risperidone in the treatment of schizophrenia
|Hansen K, Francois C, Toumi M, Lancon C
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.
Three treatments for schizophrenia were examined:
zuclopenthixol (ZUC), a neuroleptic belonging to the family of thioxanthenes, which are conventional agents;
haloperidol (HAL), another conventional agent; and
risperidone (RIS), a new combined two-serotonin and two-dopamine antagonist.
Economic study type
The study population comprised patients with a DSM-III or ICD-9 diagnosis of schizophrenia.
The setting was the community. The economic study was carried out in France.
Dates to which data relate
The effectiveness and resource use data were obtained from studies published from 1981 to 2002. The price year was 1998.
Source of effectiveness data
The effectiveness evidence was derived from several published studies. The authors also made some assumptions, which were used in the decision model.
A Markov model was constructed to assess the cost-effectiveness of the three treatments for schizophrenia over a period of 5 years. Each cycle in the model lasted 6 months. The 6-month cycle period was justified by the authors on the clinical grounds that, currently, it is accepted that any schizophrenic deterioration that occurs within 6 months of a relapse should be considered part of that relapse. Patients entered the model beginning with one of the three treatments. They could then experience extrapyramidal symptoms (EPS), compliance with the treatment, relapse, hospitalisation, switch to other drugs, death, and so on.
Outcomes assessed in the review
Several outcomes derived from the literature review were used as model inputs. These were the transitional probabilities for drug response, switch to opposite medication, use of depot form, tolerance without EPS, compliance, relapse, hospitalisation, outpatient care and intensive care. The remaining probabilities used in the model were calculated by subtraction. The result of the main analysis was the pooled odds ratio (OR) of clinical response.
Study designs and other criteria for inclusion in the review
All the primary studies included in the review were randomised controlled trials. The criteria for inclusion were that the study was randomised, double-blinded and conducted in patients with a DSM-III or ICD-9 diagnosis of schizophrenia, where the outcome measure was the percentage of patients with a minimum 20% reduction in scores on the Brief Psychiatric Rating Scale (BPRS). In addition, the study had to include the use of recommended dosages of ZUC (oral, 50 mg/day; depot, 200 mg/3 weeks) and HAL (oral, 20 mg/day; depot, 50 mg/4 weeks) had to be used.
Sources searched to identify primary studies
The primary studies were identified from the MEDLINE and EMBASE databases.
Criteria used to ensure the validity of primary studies
Randomisation and double-blind assessment were required to ensure the validity of the primary studies. Details on each trial (the number of patients and duration of the study) were reported.
Methods used to judge relevance and validity, and for extracting data
Number of primary studies included
For the comparison between ZUC and HAL, the effectiveness evidence was obtained from five primary studies. Only one relevant trial was found comparing ZUC and RIS.
Methods of combining primary studies
A meta-analysis, based on the combination technique proposed by Wolf (see Other Publications of Related Interest), was conducted.
Investigation of differences between primary studies
A test for heterogeneity was conducted to test for differences between the primary studies.
Results of the review
The following transition probabilities were derived from the review.
Drug response was 0.69 with ZUC, 0.61 with HAL and 0.46 with RIS.
Switch to opposite medication was 0.02 with ZUC, 0.21 with HAL and 0.52 with RIS.
Use of depot form given response was 0.71 with ZUC, 0.53 with HAL and 0 with RIS.
Tolerance without EPS was 0.23 with HAL (depot form) and 0.36 with ZUC (oral) and 0.12 with HAL (oral).
The compliance rate was 0.55 with depot form, 0.10 with oral form and EPS, and 0.53 with oral form and no EPS.
The relapse rate was 0.16 with depot form, 0.28 with oral form, 0.12 with compound and 0.42 without compound.
The hospitalisation rate was 0.13 with relapse and 0.01 without relapse.
The probabilities of outpatient care were 0.73 with relapse and 0.99 without relapse.
The probabilities of intensive care were 0.38 with relapse and 0.17 without relapse.
The results of the meta-analysis showed that the combined OR was 1.59. This indicated that patients taking ZUC were 1.59 times more likely (95% confidence interval: 1.11 - 2.70) to have had a clinically significant reduction in symptoms (20% or higher reduction in the BPRS from baseline to end point) than those taking HAL or RIS.
Methods used to derive estimates of effectiveness
The authors also made some assumptions to support the data used in the decision model.
Estimates of effectiveness and key assumptions
It was assumed that patients who did not tolerate the treatment continued the therapy with the aid of an anti-Parkinsonian drug. In addition, patients who did not relapse stayed in the same type of care over the whole study period. Conservative assumptions were also made in the decision model, as the same probabilities were assumed in each 6-month cycle over the 5-year period, although the authors stated that it is common knowledge that the risk of relapse increases with prior incidence of relapse.
Measure of benefits used in the economic analysis
The benefit measure used in the economic analysis was relapse-free time, which was derived by modelling. A 3.5% discount rate was also applied.
A discount rate of 3.5% was applied to all the costs. This was appropriate given that the time horizon of the study was 5 years. The unit costs were only reported separately from the quantities of resources for some items. The health services included in the analysis were study drugs, hospitalisation (day, night and complete), psychiatric visits, psychological visits, nurse visits and social worker consultations. The cost/resource boundary adopted was that of the French National Health Insurance. The costs were estimated from a study published in 1998, but the figures were then inflated to 1999 costs. The drug costs and daily doses were estimated using the Mediavidal source. The resource quantities were estimated from two complementary sources, the published study used to derive the unit costs and an expert panel of five French hospital psychiatrists. The total quantities and costs were derived using modelling. The price year was 1998.
Statistical analysis of costs
The costs were treated deterministically in the base-case.
The indirect costs were not included in the analysis.
Sensitivity analyses were performed on key variables in the decision model in order to assess the impact of each variable on the estimated cost-effectiveness ratios. The discount rate and the time horizon of the analysis were also varied. The type of analysis appears to have been univariate.
Estimated benefits used in the economic analysis
Relapse-free time was 55% of patient time with ZUC, 48% with HAL and 43% with RIS.
The overall costs were EUR 62,900 with ZUC, EUR 63,700 with HAL and EUR 64,000 with RIS.
Synthesis of costs and benefits
An incremental cost-effectiveness analysis was conducted to combine the costs and benefits of the treatments. The ZUC-based treatment was dominant over both HAL and RIS, as it led to lower costs and longer relapse-free time. This finding was robust to all variations examined in the sensitivity analyses. Only when the cost of ZUC was reduced from the base-case of 56 to 36% did the HAL-based treatment become equally cost-effective.
The zuclopenthixol (ZUC)-based treatment was a cost-effective strategy in comparison with haloperidol (HAL) and risperidone (RIS). It led to a longer period without relapse and cost-savings.
CRD COMMENTARY - Selection of comparators
The choice of the comparators appears to have been justified. The authors stated that ZUC, HAL and RIS represented alternative treatments for patients with schizophrenia. In particular, ZUC and HAL represented two conventional agents, while RIS was a new agent. You should decide whether they represent widely used treatments in your own setting.
Validity of estimate of measure of effectiveness
The analysis of effectiveness used a review of the literature. The conduct and methodology of the review were clearly reported. The inclusion criteria were given and details of the primary studies were presented. Some criteria were also used to ensure the validity of the primary studies. The method used to combine the primary study estimates (meta-analysis) was based on a sound methodology. Although not reported, it was likely that the authors adopted a weighting scheme to reflect differences in the sample sizes. The authors noted that the reliability of their final estimation of the effectiveness of the treatments was based on the validity of the primary studies. A further limitation was the fact that the time horizon of the analysis was five years, although all the primary studies analysed the short-term efficacy of the treatments. Several sensitivity analyses were conducted to test the robustness of the study, as the opinions of experts were also used in the analysis.
Validity of estimate of measure of benefit
The benefit measure used in the economic analysis was relapse-free time. This is a common outcome measure in studies examining patients with schizophrenia. Quality of life issues were not considered in the analysis, although the authors stated that the inclusion of specific factors (such as tolerance, galenic form and institutionalisation) in the decision model represented important factors in the assessment of compliance, quality of life and rehabilitation. The authors noted that a limitation of the decision model was that the method used did not allow for any variation in the rate of effectiveness over time, as would be expected.
Validity of estimate of costs
All the categories of costs relevant to the perspective adopted in the study appear to have been included in the analysis. The unit costs were reported, but the quantities of resources were not given for all item. Appropriate discounting was applied. The costs were treated deterministically in the base-case, but extensive sensitivity analyses were performed on the cost data. The indirect costs were not included in the analysis, but the authors noted that the employment rate (and consequently productivity loss) among schizophrenics is lower than in the general population. The authors also commented that this omission may underestimate the cost-savings achievable with ZUC. The costs were specific to the French Insurance System and, as a result, any extrapolation to other settings may be difficult.
The authors did not compare their findings with those from other studies, although they did compare their decision model methods. The issue of the generalisability of the study results to other settings was not explicitly addressed, but the authors stated that, due to heterogeneity in the cost data, caution should be used when extrapolating the cost results to other institutions. However, several sensitivity analyses were conducted and the model results appear to have been robust. The study referred to patients with schizophrenia and this was reflected in the conclusions of the analysis. The authors reported some limitations of the analysis. These were related to the structure of the decision tree and the use of effectiveness data from short-term trials.
Implications of the study
The main implication of the study was that, compared with HAL and RIS, ZUC proved to be a cost-effective treatment for patients with schizophrenia. Thus, ZUC should be the treatment of choice for these patients.
Hansen K, Francois C, Toumi M, Lancon C. A pharmacoeconomic evaluation of zuclopenthixol compared with haloperidol and risperidone in the treatment of schizophrenia. European Journal of Health Economics 2002; 3: 173-179
Other publications of related interest
Wolf FM. Meta-analysis: quantitative methods for research synthesis. Beverly Hills (CA): Sage Publications; 1986.
Subject indexing assigned by NLM
Date bibliographic record published
Date abstract record published