Multiple economic evaluations of Early Detection (ED) and Early Intervention (EI) for young people with first episode psychosis

Abstract

Background/Objectives: Recent studies show good evidence on the relationship between the length of the prodromal period and the risk of developing psychosis; the longer the duration of untreated psychosis, the higher the risk. It is suggested that early identification of at risk states before manifestation of severe disease can be effective in preventing mental disorders and can much reduce the economic burden to individuals as well as our societies. However, economic evaluation in the early psychosis field is lacking. The project aims to help policymakers to choose both clinically effective and cost-effective preventative strategies across various sectors including the health care service sector as well as the criminal justice system, education and employment. We explore what lessons can be leant from the findings of our models in which we also adapted some EI approaches used in UK contexts to the Australian setting. Methods: Four decision-analytical models are built to compare community-based multidisciplinary teams for Early Detection (ED) and Early Intervention (EI) in first episode psychosis with standard care. Data on costs and benefits in terms of vocational, educational, criminal justice system, and suicide related outcomes are based on empirical studies, literature review and national expert opinions. Sensitivity analyses are conducted to examine the robustness of different assumptions in the models. Results: In our first model on ED, although the intervention cost for usual care is the least costly, it incurs the greatest health care and societal cost (A$ 33,018) because of a higher proportion of people ending up with a transition to psychosis. Relative to usual care, ED costs A$ 8,288 per one psychosis event prevented with CBT, A$ 16,605 via medication regimen, and A$ 13,512 through combined therapy. The second EI model on vocational outcomes results in EI costs which are 32% lower. The expected costs of lost employment are A$ 14,165 for EI, and A$ 20,728 for standard care per year. In a sensitivity analysis, when the average wage is used instead of minimum wage, there is an increased difference in the costs for the lost employment between EI and standard care (A$ 25,441 vs. A$ 37,325). In our third EI model on homicide, the yearly cost for homicide per patient after standard care is almost 10 times higher than that of EI (A$ 16 vs. A$ 162). The fourth EI model on suicide showed the expected yearly costs of suicide with EI, which is only one third of that with standard care (A$ 2,104 vs. A$ 6,479. Therefore, overall EI services have substantially lower costs than standard care. Discussion/Conclusions: These modelling exercises do offer flexibility in estimating costs and effects, which can be adapted to local circumstances. Similar to the considerable cost savings found in UK settings, EI services in the Australian context suggest significant savings due to reduced lost employment at work, as well as decreased homicide and suicide events.