The principle assumption to achieve this at current levels of resource is, that by investing now and incurring short term increase in cost due to the price of new medication, the net effect over time will be to see severe cases of dementia become less common. Consequently the need for institutional care in this group should diminish. The drugs cost on average £1000 per annum each and combinations are inevitable - perhaps meaning that up to £4000 per patient may eventually be required. Compared to £30,000 yearly for nursing home care it is easy to see how these predictions arise.But life is not that simple.

Drug budgets are historically low in mental health services, mainly because after many years of no new drugs, the budgets were sacrificed for efficiency savings in the early 1990s. The average spend across the UK is only 1.5 per cent of revenue - clearly insufficient. A central mandate to strategically restore budgets to a sensible level (at least 10 per cent revenue) using redevelopment and new funds would help clarify the financial situation.

Another problem is that most services are geared towards moderate to severe dementia and the infrastructure is not ready to treat mild illness - in terms of services or primary care awareness. NICE money compensates for some of this but care needs to be taken to ensure that it is not spent disproportionately in substituting for absent service costs at the expense of prescribing actual treatment to those who need it.

If the predictive models that say treating this patient group may be cost neutral are to work, money flow must be transparent in the system. The main saving is in social care (care home fees), the main initial expense is to health (drug cost). These have to match within the same system and they currently do not. Real dementia services need to be set up, either in one organisation (PCT), or across several, that ensure the resource for these patients flows within a truly joined-up system that also includes the private sector. Without this social services make all the savings - which then get diverted elsewhere.

The PCT should host such a service with the individual GP now able to opt in or out of caring for this patient group - changing the traditional model, where everything ends up in the GP surgery, and introducing a concept of specialist primary care. Finally, ageism abounds and may be difficult to eradicate. Transparency would be achieved by having dedicated budgets for the elderly that were adequate and protected. NICE has helped with postcode prescribing but postcode resourcing is now the issue. The average health spend per head of elderly population to provide the services in the Forget me not report was £86 per head. Many services do not have that.

In Swindon we tackled the drug budgets and infrastructure during repatterning of services, and worked closely with social services and the private sector to form good working relationships. Now being in a large mental health trust with competing priorities, finding intermediate care hijacked by acute services and discovering that formal links with social services brings new problems, getting appropriate care to our patients is yet again challenging. We aim to create a dementia action team across all providers and hosted by the PCT. If this is successful it will hopefully prove to be the model within which all future services can secure the best treatment for those unfortunate enough to suffer with dementia.