Do early intervention for psychosis services really save money?

To the Editor

In its 2011 budget, the Australian Government allocated $220m for the expansion of early intervention for psychosis services. One of the arguments in favour of this investment is that such services can actually save money (McGorry, 2012). It was therefore surprising to read the recent review by Amos (2012), which concluded that ‘The published literature does not support the contention that early intervention for psychosis reduces costs’ and the accompanying commentary by Carr (2012), which stated that ‘Clearly, the case for cost-effectiveness of EI [early intervention] for psychosis has not been made, and similarly the argument that cost savings from EI would be available for reinvestment in other services is not based on credible evidence’.

Reading these conflicting claims led me to examine some of the documents supporting the Government’s expansion of early psychosis services. A key document is the ‘Early Psychosis Feasibility Study Report’, which was prepared for the National Advisory Council on Mental Health and the Commonwealth Department of Health and Ageing (Orygen Youth Health Research Centre, 2011).

This report included a chapter on ‘The economics of early intervention’, which was based on some modeling done by economists in the UK. To me, as a non-economist, the model looks reasonable. However, the data on which the model draws are poor. In particular, the estimation of healthcare costs is based on a study of the Early Psychosis Prevention and Intervention Centre (EPPIC) service in Melbourne, which involved comparing 51 patients treated in 1993–1994 with 51 matched controls pre-EPPIC treated in 1989–1992. These costs from two decades ago were simply adjusted to 2009 figures. However, it is unclear whether service models that operated so long ago can be generalized to current practice for either early psychosis services or standard care. Even if they were relevant to current practice, the sample size is not a large one on which to estimate national costs. Furthermore, the use of historical controls is methodologically weak compared to the gold standard of randomized trial data.

The data used to estimate the costs of lost employment, suicide and homicide are similarly thin. For example, the data on suicide come from a single overseas study on attempted suicide and the assumption has been made that the difference in attempted suicide in one service in another country would also apply to completed suicide in Australia.

An interesting aspect of the report is the conclusion about how much the Government should allocate for the establishment of each new early psychosis service. Having concluded that early psychosis service costs are 65% of those of standard care, the authors take this cost (estimated as $20m per year for each metropolitan service) and ask for it as additional money. What the authors have done is to view the cost of a new early psychosis service as additional to the cost of standard care, rather than viewing it as an eventual substitute for more expensive standard care. Unless there is a longer-term reduction in expenditure on standard care, the total service cost might be closer to 165% than 65%.

My conclusion is that the economic argument for early psychosis services is indeed weak. However, despite this weakness, early psychosis services may still be valuable and could be justified on other grounds. Indeed, given that services for people with severe mental disorders are in general not well funded, any additional allocation for a more intensive form of care should be welcomed by everyone in the mental health sector. In rolling out these services nationally, it is to be hoped that the Government will allocate some of the funds for an independent evaluation of both the economic and therapeutic impact.