Assumptions used in estimating a population benchmark

The assumptions used in estimating a population benchmark rate for new referrals into a memory assessment service of 0.19%, or 190 per 100,000 population, per year are based on the following sources of information:

• epidemiological data on the prevalence/incidence of dementia
• current practice to establish the number of existing patients in contact with GP services
• published research and local audits on the diagnostic profile of people referred to a memory assessment service
• expert clinical opinion of the topic-specific advisory group, based on experience in clinical practice and literature review.

Epidemiological data

Incidence of dementia

The Medical Research Council Cognitive Function and Ageing Study^[1] used identical methodology in five diverse sites across England and Wales to assess the annual incidence of dementia by age and sex. The study found that rates rise with age, particularly above the age of 75, and that the rate of increase was marked in both sexes.

Application of the age- and sex-specific rates from the Medical Research Council Cognitive Function and Ageing Study to estimates for the population of England in mid-2005 suggests that the annual incidence of dementia in the population is 0.31%. This represents the proportion of the population that is likely to develop dementia within 1 year.

Currently, a large proportion of these people with dementia are likely to remain undiagnosed or will not have their condition diagnosed until the disease has progressed^[2].

Prevalence of dementia

A report to the Alzheimer's Society by the London School of Economics and Kings College London suggests that there are around 684,000 people with dementia in the UK, around 574,000 (84%) of whom are resident in England. This corresponds to a population prevalence for England of 1.1%. Currently a large proportion of these people are likely to be undiagnosed^[3].

Mild cognitive impairment

Some people are identified with symptoms of cognitive decline but do not meet clinical criteria for the diagnosis of dementia. These people are said to have mild cognitive impairment (MCI)^[4].

The definition of MCI used within the NICE-SCIE clinical guideline CG42 on dementia is a syndrome defined as cognitive decline greater than expected for an individual's age and level of education but which does not interfere notably with activities of daily life^[4]. MCI is heterogeneous in clinical presentation, and there is no clear strategy for diagnosing people with suspected MCI^[4].

A proportion of people with MCI have been shown to develop some form of dementia over time. The rate of conversion from MCI to dementia depends on the diagnostic criteria used, the type of cohort studied and the length of observation time from diagnosis of MCI^[4].

In people with MCI who are referred to memory assessment services and other specialist centres, the rate of conversion to dementia (generally Alzheimer's disease) has been estimated to be around 18% per year^[5].

Several different types of MCI have been proposed^[6,7], to include a range of cognitive impairments and deficits, including deficits in the non-memory domain. Studies that have been conducted using these definitions have estimated a population prevalence of MCI of between 5% and 25%^[8-10] in older people. The value depends on the specific diagnostic criteria used and the type of cohort studied^[6].

However, most people with MCI or subjective memory impairment - the perception of memory problems by people themselves or their carers - do not report any symptoms to their general practitioner (GP) or may present with symptoms other than cognitive impairment^[4].

Current practice

The General Practice Research Database (GPRD) holds data for over 3.4 million currently registered patients from a sample of general practices (GPs) in England, Wales, Scotland and Northern Ireland, and is a representative UK sample by age and sex.

Data on the incidence of dementia in people aged 60 years and over were extracted from the GPRD to determine age- and sex-specific rates, which were then applied to the population of England in mid-2005. The results of the analysis suggest that the annual incidence of diagnosed dementia is 0.10% in the population as a whole, and 0.49% in the population aged 60 and over.

The estimated prevalence of diagnosed dementia based on GPRD data is 0.40% in the population as a whole, and 1.89% in the population aged 60 and over.

Although the analysis using GPRD data is based on people with dementia who are aged 60 and over, the use of these data is unlikely to significantly underestimate the prevalence or incidence figures. This is because onset of dementia in older people (over 65) accounts for around 98% of all dementia cases, and in the population with early onset of dementia, people aged 60-65 account for around one-third of cases in this group^[3].

It is not possible to analyse GPRD data to estimate the number of people who may have MCI, subjective memory impairment or a related condition that would be suitable for referral to a memory assessment service. This is because many people with MCI do not report symptoms to their GP^[4].

The quality of data from general practice databases depends on the coding of conditions used by GPs. This means that there may be a discrepancy between data from GP practices and the actual incidence and prevalence of diagnosed dementia in primary care. The numbers of people diagnosed with dementia and the percentage of those reviewed by their GP are being recorded in primary care as part of the Quality and Outcomes Framework (QOF). For guidance relating to the technical requirements for correctly recording QOF information, see the QOF business ruleset.

Published research and local audits

Diagnostic profile of people referred to memory assessment services

Not all people attending a memory assessment service will have dementia. Some may have cognitive impairment that does not meet the criteria for dementia or conditions associated with cognitive symptoms, whereas others may have no objective cognitive impairment^[11].

Luce and coworkers (2001)^[12] examined the diagnostic profile of 100 consecutive referrals to a memory assessment service. They found that around 57% of the people attending had probable or possible dementia, with the remainder having other conditions or cognitive deficits that did not meet the criteria for dementia.

Banerjee and coworkers (2006)^[13] examined the diagnostic profile of people referred to a memory assessment service during an 18-month period. They found that 63% of people referred to the service had some form of dementia, with the remainder having either other conditions or no illness.

Other published research^[11] suggests that the proportion of people attending memory assessment services who have some form of dementia is 43%.

The diagnostic profile of people referred to a memory assessment service based on two service audits is given in the table below.

	People with condition (%)
	South Manchester memory clinic	Croydon memory service
Dementia	43%	65%
Mild cognitive impairment	37%	17%
Depression	14%	4%
No illness	3%	10%
Other	3%	4%

The figures from published research and service audits depend on the nature of local referral patterns and of the memory assessment service itself.

Taking the average of estimates from published research and audits suggests that around 54% of people attending memory assessment services are likely to have dementia, and 46% are likely to have either other conditions such as MCI or no illness.

Expert clinical opinion

The consensus opinion of the topic-specific advisory group was that where memory assessment services are available there is likely to be an increase in the referral of people with suspected dementia for assessment.

A benchmark determined by current activity (based on areas with and without memory assessment services) should be considered as a minimum expected rate of referral. Therefore commissioners should consider planning for increased activity to reflect current unmet need.

Conclusions

Based on the epidemiological data and other information outlined above, it is concluded that the proportion of the population suitable and identified for referral to a memory assessment service is 0.19% of the population as a whole and 0.91% of the population aged 60 years and over. This is based on the following assumptions:

• the annual incidence of diagnosed dementia for the whole population is 0.1%
• the proportion of people attending a memory assessment service who have conditions other than dementia is around 46%.

Increasing the estimate of 0.1% to take into account referrals of people without dementia gives a whole-population estimate of 0.19%, and an estimate of 0.91% of the population aged 60 years and over.

Hence a population benchmark for referral to a memory assessment service of 0.19% is considered appropriate. This represents a minimum expected rate of referral because a large proportion of people who develop dementia are currently likely to remain undiagnosed or will not have their condition diagnosed until the disease has progressed^[2]. Therefore commissioners may wish to consider planning for increased activity to reflect current unmet needs of their population.

Sensitivity analysis based on the differing age and sex structures within primary care organisations suggests that in around 90% of primary careorganisations variation around this population benchmark of 0.19% will be within ±0.07%, giving a range of 0.12-0.26%.

Use the memory assessment service commissioning and benchmarking tool to determine the level of service that might be needed locally and to calculate the cost of commissioning the service using the indicative benchmark and/or your own local data.

References

1. Matthews FC, Brayne C (2005) The incidence of dementia in England and Wales: findings from the five identical sites of the MRC CFA study. PLOS Medicine 2: 753-63.

2. Wilkinson D, Sganga A, Stave C et al. (2005) Implications of the Facing Dementia Survey for health care professionals across Europe. International Journal of Clinical Practice 59: 27-31.

3. Kings College London and the London School of Economics (2007) Dementia UK. A report to the Alzheimer's society on the prevalence and economic costs of dementia in the UK. London: Alzheimer's Society.

4. Gauthier S, Reisberg B, Zaudig M et al. (2006) Summary of the content of an expert conference convened by the International Psychogeriatric Association Jan 21-23, 2005 with objective of clarifying the diagnosis and management of mild cognitive impairment. Lancet 367: 1262-70.

5. Kluger A, Ferris SH, Golomb J et al. (1999) Neuropsychological prediction of decline to dementia in nondemented elderly. Journal of Geriatric Psychiatry and Neurology 12: 168-79.

6. Portet F, Ousset PJ, Visser PJ et al. (2006) Mild cognitive impairment (MCI) in medical practice: a critical review of the concept and new diagnostic procedure. Report of the MCI working group of the European Consortium on Alzheimer's disease. Journal of Neurology, Neurosurgery and Psychiatry 77: 714-8.

7. Winbald B, Palmer K, Kivipelto M et al. (2004) Mild cognitive impairment - beyond controversies, towards a consensus. Journal of Internal Medicine 256: 240-6.

8. Kumar R, Dear KBG, Christenesen H et al. (2005) Prevalence of mild cognitive impairment in 60- to 64-year-old community-dwelling individuals. Dementia and Geriatric Cognitive Disorders 19: 67-74.

9. Manly JJ, Bell-McGinty S, Tang M-X et al. (2005) Implementing diagnostic criteria and estimating frequency of mild cognitive impairment in an urban community. Archives of Neurology 62: 1739­-46.

10. Purser JL, Fillenbaum GG, Pieper CF et al. (2005) Mild cognitive impairment and 10-year trajectories of disability in the Iowa established populations for epidemiological studies of the elderly cohort. Journal of the American Geriatrics Society 53: 1966-72.

11. Hejl A, Hogh P, Waldemar G (2002) Potentially reversible conditions in 1000 consecutive memory clinic patients. Journal of Neurology, Neurosurgery and Psychiatry 73: 390-4.

12. Luce A, McKeith I, Swann A et al. (2001) How do memory clinics compare with traditional old age psychiatry services? International Journal of Geriatric Psychiatry 16: 837-45.

13. Banerjee S, Willis R, Matthews D et al. (2007) Improving the quality of care for mild to moderate dementia: an evaluation of the Croydon memory service model. International Journal of Geriatric Psychiatry 22: 782-88