Comparing the cost-effectiveness of Haloperidol, Risperidone and Olanzapine in the treatment of schizophrenia using the net-benefit regression approachComparing the cost-effectiveness of Haloperidol, Risperidone and Olanzapine in the treatment of schizophrenia using the net-benefit regression approach
Faculty of Applied Economics
Faculty of Social Sciences. Sociology
Research group
Mathematics, Statistics and Actuarial Sciences
Publication type
Antwerp :UA, [*]
Source (series)
Research paper / UA, Faculty of Applied Economics , 2007:12
29 p.
Target language
English (eng)
University of Antwerp
OBJECTIVES: This study determines the cost-effectiveness of 3 antipsychotics for the treatment of schizophrenia in Belgium. METHODS: Data were retrieved from a prospective observational non randomized follow-up survey. Clinical investigators included 293 schizophrenic patients; 136 of those patients were assigned to Olanzapine, 129 to Risperidone and 28 to Haloperidol. Patients were followed for 2 years. Total health care costs were determined from the perspective of the public payer and calculated by multiplying resource use with official tariffs; effectiveness of the drugs was measured with EQ-5D. Several studies have already compared the cost-effectiveness of different antipsychiotics for the treatment of schizophrenia, most of them are however flawed by methodological issues. This study therefore uses a new method that was developed to address these limitations but is not widely used yet: the net-benefit regression approach (NBRA). We show its merits by performing a cost-effectiveness analysis of Olanzapine, Risperidone and Haloperidol. RESULTS: Models were checked for selection bias but drug choice was not endogenous; we therefore proceeded with simple OLS regressions. The results indicate that the drugs provide similar net monetary benefits to the patient (H vs O -4452.53 (p=0.645), R vs O 4439.54 (p=0.425), R vs H 8892.07 (p=0.366)). When we control for several patient characteristics R moves away further from H and O but the difference does not reach statistical significance (R vs O 5857.73 (p=0.332), R vs H 15233.53 (p=0.178)). Several important patient subgroups were also identified; they indicate that a drug performs better in a specific patient group. Numerous sensitivity analyses confirm the robustness of the results. CONCLUSION: We conclude by confirming that the NBRA is an important enrichment to the CEA methodology. As was demonstrated in this paper, it is often important to correct cost-effectiveness results for patient characteristics and to identify significant patient subgroups.
Full text (open access)