One common finding in analyses of health systems with respect to dementia is that there is a “diagnosis gap” in dementia with less than a half of those with dementia ever attracting a diagnosis of dementia. The service response to this has been to develop memory clinics and other services to enable good quality diagnosis of the syndrome of dementia and its sub-types.

This narrative review considers who memory assessment services are for and what they should aim to achieve. We will consider the evidence base and discuss “what good looks like.”

The sparsity of the evidence base for the provision of memory services is striking. There is a lack of studies that have evaluated the absolute and relative impact of different models of diagnostic services or the impact of diagnosis and stage of illness at diagnosis.

There remains genuine uncertainly about: the positive and negative impacts of receiving the diagnosis of dementia; the effects of receiving the diagnosis of dementia at an earlier or later stage; and how best to provide memory assessment services in terms of clinical and cost-effectiveness. We need applied health research designed to fill these important evidence gaps, resolving uncertainty, and allowing the development and delivery of efficient and effective services and policy to enable people to live well with dementia. The methodology that will be needed will be a challenge since, due to ethical and practical considerations, it is likely to have to be observational rather than experimental.