Information about the cost-effectiveness of early intervention programmes for first-episode psychosis is limited.
To evaluate the cost-effectiveness of an intensive early-intervention programme (called OPUS) (trial registration NCT00157313) consisting of enriched assertive community treatment, psychoeducational family treatment and social skills training for individuals with first-episode psychosis compared with standard treatment.
An incremental cost-effectiveness analysis of a randomised controlled trial, adopting a public sector perspective was undertaken.
The mean total costs of OPUS over 5 years (€123,683, s.e. = 8970) were not significantly different from that of standard treatment (€148,751, s.e. = 13073). At 2-year follow-up the mean Global Assessment of Functioning (GAF) score in the OPUS group (55.16, s.d. = 15.15) was significantly higher than in standard treatment group (51.13, s.d. = 15.92). However, the mean GAF did not differ significantly between the groups at 5-year follow-up (55.35 (s.d. = 18.28) and 54.16 (s.d. = 18.41), respectively). Cost-effectiveness planes based on non-parametric bootstrapping showed that OPUS was less costly and more effective in 70% of the replications. For a willingness-to-pay up to €50,000 the probability that OPUS was cost-effective was more than 80%.
The incremental cost-effectiveness analysis showed that there was a high probability of OPUS being cost-effective compared with standard treatment.
Most health economic evaluations in mental care include outcome measures aimed at specific aspects of health, like symptom improvement, functional improvement and quality of life instead of generic preference based outcome measures. The health economic guidelines (NICE) recommend to include a generic preference based outcome measure, like EQ-5D, in health economic evaluations in order to allow for comparisons of health related quality of life of patient groups across different diseases, providing information particular useful to support health policy decisions and cost-effectiveness analysis. Although the EQ-5D is by far the most widespread outcome measure within the context of economic evaluations, its validity in psychiatric populations has not yet been established unambiguously. An increasing number of articles have tested the validity of the EQ-5D in comparisons with clinical measures in mental health, but only few studies have addressed the correlation between the EQ-5D and a condition-specific quality of life measure in mental health.
The aim of the article is to test for a potential correlation between the preferred generic outcome measure in health economic evaluations EQ-5D and Manchester Short Assessment of Quality of Life (MANSA) in order to assess to what extent quality of life dimensions measured by a psychiatric quality of life measure are captured in the EQ-5D in a population of patients with schizophrenia and cannabis abuse.
Data analysed is a part of a study of 103 patients with schizophrenia and abuse of cannabis participating in a randomized controlled trial testing a specialized addiction intervention during the period 2008-09. The correlation of the EQ-5D and scores of MANSA was assessed using the Spearman's correlation coefficient. In addition, we tested how the EQ-5D and MANSA correlated with PANSS, GAF and WHO-DAS in order to make comparisons with earlier studies.
We found moderate, statistically significant correlations between the EQ-5D index score and MANSA total score (rho = 0.358**). The dimensions 'Mobility', 'Self-Care' and 'Pain/discomfort' on the EQ-5D were overall not sensitive in this population, while the dimensions 'Usual activities' and 'Anxiety/depression' were moderately correlated with MANSA. The EQ-5D and MANSA both showed statistically significant moderate correlations with the clinical measures in the study PANSS, GAF and WHO-DAS.
Our results suggest that the EQ-5D and MANSA are complementary measures rather than substitutes.
Mental health interventions often seek to improve the patients' quality of life in a broader perspective, like improving the patients' relationship with family, friends and other network, financial situation, employment and accommodation. If the EQ-5D is used as a single outcome in health economic evaluations of e.g. mental health community interventions, these factors may be overlooked. Based on a relatively small sample, we therefore recommend applying the EQ-5D together with condition-specific quality of life measures in future health economic evaluations in mental health.