We would like to comment on the interesting editorial by Holloway (Reference Holloway2002) on outcome measurement in mental health, by reporting observations on psychogeriatric services in Italy.

Our country is going through a profound economic crisis, with consequences for health expenditure. In this framework regional governments, who have the duty to coordinate health service programmes, are induced to save money in the more ‘frail’ areas, such as mental health and geriatric services. One of the reasons for this attitude is the claim that evidence for the usefulness of these services is weak. Yet, at the same time, there are very few attempts to fund programmes devoted to acquiring such evidence. Only a small proportion of research grants, even from central government, are devoted to outcomes research in this area. We argue that the scarcity of health resources combined with devolution of health care from central to local governments support an urgent need for outcomes research implementation in the psychogeriatric field.

At a conservative estimate, <20% of the procedures adopted in psychogeriatrics are evidence-based and follow accepted guidelines. We agree about ‘the difficulty of conducting evaluation of the complex social interventions typically deployed within mental [and, we would add, geriatric] health services’ (Reference HollowayHolloway, 2002) but, paradoxically, it is in times of scarce resources that it is of most relevance to evaluate whether the allocation of money to psychogeriatric services leads to significantly improved outcomes. Moreover, the 21 regions of Italy are undergoing a process of autonomy. One of the risks of this is that each region will adopt different means of measuring the quality of procedures and outcomes. This is particularly relevant if we consider the fact that evidence-based medicine, which might be a standard reference, covers only a small proportion of interventions.

We do not have programmes similar to the UK Department of Health's ‘Mental Health Information Strategy’ nor do we collect data to compile a minimum dataset. The majority of the work in psychogeriatrics is done without quality controls and it is not possible to benchmark different services against each other. Furthermore, clinicians are deprived of the possibility of measuring outcomes of their interventions, particularly in areas where the data do not allow a direct transfer of information in everyday clinical practice.

Currently, the future of special care units (SCUs) in nursing homes in the Lombardia region is widely debated. In this region a network of 60 SCUs has been active for the past 8 years to treat patients with dementia with severe behavioural disturbances. The units are funded by the regional health system with 15 euros/patient/day more than regular nursing homes. Since the regional government has decided to optimise expenditures for geriatric and psychogeriatric services, it has been asked whether clinical results obtained in the SCU are worth the extra money. Unfortunately, the international literature analyses markedly different models, and Italian SCU researchers apparently have not performed adequate studies to measure outcomes. The only meaningful piece of evidence is an observational controlled study of 18 SCUs and 25 traditional nursing homes funded by the European Commission, which demonstrates that patients admitted to SCUs had behavioural disturbances of severity similar to patients cared for in traditional nursing home wards, but with significantly less physical restraints (Reference Frisoni, Bianchetti and PignattiFrisoni et al, 1999). Consequently, as physicians, we are unprepared to dispute the decisions of the government and cannot affect the future of SCUs.

This experience further supports the need to implement in our country a system of outcomes research. Although the promises are probably higher than the obtainable results, it is essential to start this process if we hope to improve the diffusion and the quality of psychogeriatric services in Italy.