Clinical Decisions on Antipsychotic Treatment

Background: In real life, physicians frequently face complex clinical problems or even dilemmas in pharmacological treatment of patients. These problems may involve not only efficacy and safety, but also cost, quality of life and others. Cost-effectiveness analysis (CEA) and cost-utility analysis (CUA) may be helpful in this regard because these methods can take into consideration all relevant factors.
Objectives: We aimed to apply the CEA and CUA method to help solve two important clinical questions: (1) “Intramuscular (IM) olanzapine versus intramuscular haloperidol plus lorazepam for the treatment of agitation in schizophrenia: which one to choose?” and (2)“Atypical and typical antipsychotic agents for the treatment of agitation/ psychosis in dementia: which to prescribe or not to prescribe?”
Methods: We conducted 2 studies to answer the 2 questions. In study one, we implemented a randomized controlled trial (RCT) comparing IM olanzapine versus IM haloperidol plus lorazepam in the treatment of acute agitation of schizophrenia, and then performed Marcov decision model, CEA and CUA based on data from the RCT. In study two, we searched relevant information from literature regarding the antipsychotic treatment of agitation/ psychosis in patients with dementia, and performed Marcov decision model, CEA and CUA on 3 different strategies: typical antipsychotic treatment, atypical antipsychotic treatment and placebo (assumed to be equivalent to non-pharmacological supportive care).
Results: In study one, a total of 67 agitated patients with schizophrenia or schizoaffective disorder were enrolled. The RCT showed that both treatments were not significantly different in terms of efficacy and safety measurements. Regarding response, CEA showed the olanzapine treatment was dominated by haloperidol plus lorazepam treatment because the former had higher cost and marginally lower response proportion. However, in terms of quality, CUA showed olanzapine treatment had higher quality of life and the incremental cost was 4387.5 NTD per utility-gained compared to the haloperidol plus lorazepam treatment. Study two found that the 2 antipsychotic treatments were very similar in total survival life-years in one-year follow-up, but the survival life-years were shorter than that in the placebo group. However, the 2 antipsychotic treatment groups also spent much less money than the placebo group. CUA showed that both atypical antipsychotic treatments and placebo treatment were dominated by typical antipsychotic treatment. After one-year follow-up, the placebo group had worst quality of life and biggest expenditure.
Discussion: Study one found the 2 treatments are similar in efficacy, but IM olanzapine treatment seems to be better than IM haloperidol plus lorazepam treatment because of its relatively lower incidence of significant adverse drug reactions. However, the cost of IM olanzapine is about 9 times as that of haloperidol plus lorazepam treatment. When quality of life is considered, IM olanzapine treatment has higher utility and the incremental cost-utility ratio (ICUR) indicates one extra utility can be gained at the price of 4387.5 NTD for one time intervention. Depending on economical status and government policy, physicians of different countries may make different decisions. Study two showed the 2 antipsychotic treatments were similar in terms of one-year survival. This might be due to that facts that, compared with typical antipsychotic treatment, the atypical antipsychotic treatment was associated with slightly lower all-cause mortality, but also slightly higher risk of cerebrovascular accidents (CVAs). Compared with the 2 antipsychotic treatment groups, the placebo treatment group had the biggest expenditure and longest survival after one-year follow-up. When we took quality of life into consideration, both placebo and atypical antipsychotic treatment were dominated by typical antipsychotic treatment. This was demonstrated by the CUA. Although the placebo group had lower risk for CVAs, it also had higher risk for agitation (higher probability of “non-response”), which would impair the quality of life significantly. Therefore, the overall quality adjusted life-years (QALY) of the placebo group was still low. A sensitivity analysis on the assigned utility of “non-responder” also showed consistent findings that the placebo group had the lowest QALY during one-year follow-up.
Conclusions: These 2 studies provide support to the use of CEA and CUA in making clinical decisions. By using these methods, physicians can examine a clinical complex treatment issue in a more comprehensive manner, not only considering efficacy and safety, but also taking cost and quality of life into account.