Alzheimer's Disease (AD) has a major economic impact on the health service, local authorities, other public sector bodies, charities, families and sufferers themselves. The extent of this impact is not widely appreciated.
The gross annual cost of AD in the UK is between £7.06bn and £14.93bn.
Comparison of the costs of AD with similar calculations for stroke, cancer and heart disease - although hampered by methodological differences and data limitations - reveals the considerable absolute and relative economic impact of Alzheimer's Disease.
Research funding on AD is low, and especially so when set alongside research funding on other diseases. Research funding on AD is 60% of the amount spent on stroke, less than 10% of the amount spent on heart disease, and only 3% of the amount spent on cancer.
1. The purpose of this paper is to summarise evidence on (a) the costs of Alzheimer’s Disease in the UK; and (b) levels of research funding on this illness and its treatment. Comparisons are made in both respects with three other major causes of mortality in the UK – heart disease, cancer and stroke. This work was commissioned by the Alzheimer’s Research Trust, and undertaken by the Centre for the Economics of Mental Health, Institute of Psychiatry, King’s College London.
COST OF ALZHEIMER’S DISEASE
2. Cost of illness studies estimate the overall economic ‘burden’ of a disease. Such calculations endeavour to identify all of the services provided in order to prevent, diagnose, treat and otherwise cope with the disease in question. Attempts are also made to identify other economic consequences of the illness. The distinction is often made between direct and indirect costs, the former relating to goods and services which require a direct cash outlay, estimated by adding up the costs of the individual components, and the latter attaching values to resources which do not require a direct cash outlay, such as time lost from employment or premature mortality. Patient, health service or societal perspectives might be chosen, determining the breadth of coverage of the costing. Cost of illness studies are not evaluative, but they can usefully supplement prevalence, incidence and morbidity data; and they can provide useful background material when considering strategies for prevention and treatment and when discussing research priorities.
3. Estimation could follow either an incidence approach, measuring the lifetime costs associated with all cases of an illness diagnosed in a particular time period, or a prevalence approach, measuring the total costs of everyone with the illness at a point in time. Incidence-based estimates can feed useful data into cost-effectiveness discussions, while prevalence-based estimates are often employed when examining impacts on public and private expenditures.
4. Cost of illness studies have their limitations (Smith and Wright 1996), and it is not always straightforward to make comparisons between them. Some will be incidence-based and others prevalence-based, and incidence/prevalence may not be calculated in the same way across studies; they may cover different service ranges; unit costs may vary from study to study; inclusion and valuation of lost productivity may also vary; and some studies might estimate total costs associated with an illness while others estimate incremental costs over some normative amounts. Cost of illness calculations anyway reflect past resource allocation, so that disorders which have previously received low priority will have low direct costs that do not necessarily indicate current needs for support. Notwithstanding these difficulties, we believe that the comparative cost of illness data offered in this paper – for Alzheimer’s Disease (AD), cancer, stroke and heart disease - provides a useful platform for discussion.
5. A search was conducted on the University of York’s Centre for Reviews and Dissemination database and the Health Economics Evaluation Database using the following terms: cost of illness, burden of disease, Alzheimer’s, dementia, stroke, cancer and heart disease. We also contacted the Office of Health Economics and major charities and research centres related to each disease type (e.g. Imperial Cancer Fund, British Heart Foundation) to locate ‘grey literature’ on the cost of illness. A hand search of reference lists for all relevant publications was also conducted. Following a review of the available literature on the cost of illness of AD a re-working of secondary data was undertaken to make estimates complete and bring them up to date. Projections of future costs were also made based on these re-worked cost estimates.
Alzheimer’s Disease studies
6. We first present general findings from the literature, then summarise the UK findings on the costs of AD. A best estimate of the current cost of illness of Alzheimer’s Disease (in 2000) to the UK is presented based on these previous findings.
7. There is a wide range of estimates of the cost of dementia and AD across countries and over time. Nonetheless, several common results emerge. First and foremost, dementia is a costly illness regardless of how it is measured and what costs are included. Costs of formal care are dominated by institutional care, particularly in hospitals and nursing homes. However, for patients supported for at home the unpaid cost of informal care is typically much larger than the paid cost of formal care (Max 1998).
8. A number of UK studies have examined the cost of AD. The occurrence of AD is age-dependent, affecting few people aged under 60, but as many as 1 in 7 people aged over 80 (the Wellcome Trust estimates the figure as nearly 1 in 6). There is no evidence that the prevalence of AD is changing but the ageing population means the number of cases will increase. On current projections the costs of treating AD will substantially increase over the next two decades. A high proportion of the total cost of AD is the cost associated with informal care. Social and demographic change will likely impact on future informal care, and it is possible that people will be less willing to provide informal care in the future (Nuttall et al. 1993).
9. Current treatment for Alzheimer’s Disease tends to be palliative. As well as treatment directly related to the disease itself, many people with AD suffer from depression and many experience falls, accidents and infections - leading to increased acute hospital admissions.
10. International comparisons show that types of care and proportions of patients cared for at home rather than in institutions vary widely between countries (Knapp et al. 1998). In 1995 it was found that 7% of elderly people in the UK with dementia were ‘permanently’ treated in hospital and another 33% in residential and nursing homes (Kavanagh et al. 1995), while a Canadian study found 50% were in long-term care (6% of whom would probably have been there without dementia) (Æstbye and Crosse 1994).
11. Drug therapy costs for AD were until recently so low that they were not distinguishable by the NHS Executive in 1992/93. In contrast, net NHS pharmaceutical expenditure was reported for less common chronic diseases such as hypertension (9.04%), diabetes mellitus (2.67%). Pharmacotherapies now being introduced could ‘have three impacts of relevance to economic cost: they may permit a delay in the institutionalisation of the patient, may reduce the number of hours of informal caregiving required in the short run, and may therefore improve the quality of life for patients and caregiver’ P 204 (Max 1998). Studies are now being conducted that examine whether these potential effects are realised (see, for example, Wimo et al. 1998), although this is not the focus of the present paper.
12. A major issue in estimating the cost of Alzheimer’s Disease lies in valuing indirect costs, including the value of time lost from activities (both by the patient and by the informal carers), the value of lives lost and the physical and emotional suffering experienced by patients and their families. A further indirect cost lies with younger carers who might face a future of relative poverty as their time out of the labour force while caring makes them less attractive to employers.
13. Costing informal care is contentious. Obtaining time measures is a first hurdle. A survey in Lombardy found an Alzheimer’s Disease patient required 18 hours of non-medical services and 45 hours of personal care from a primary caregiver per week (Cavallo and Fattore 1997). A US study estimated the average weekly commitment to be 25.2 hours (Stommel et al. 1994), while another suggested it is between 69 and 100 hours per week (Rabins et al. 1982). A 1985 US survey found that severely demented elderly people receive an average of 56 informal care hours a week while mild and moderately demented patients receive 22 hours per week (Hu and Huang 1986).
14. Having established an estimate of the time commitment there is the issue of applying a suitable cost to this time. Valuing lost employment is one way, but many carers are retired (almost 50% of those caring for dementia sufferers in the UK are over 70; (Bosanquet et al. 1998). Another method is to use the replacement method - the cost of substituting a formal for an informal carer. As well as these time costs, caregiving may lead to stress, fatigue and poorer quality of life. Some studies have found that it leads to increased utilisation of health services.
15. A study in Northern California estimated the imputed value for informal care to be almost 75% of the costs for non-institutionalised patients and only 12% of the costs for patients in nursing homes (Rice et al. 1993). A study conducted in the mid-eighties in the US found that indirect costs accounted for 85% of the total cost of senile dementia (Huang et al. 1988). However only 42% of this was time lost by family members, with the remaining indirect cost component due to morbidity, disability and increased mortality associated with the disease. Another US study estimated that informal care was approximately 43% of the total cost of care of dementia (Manton et al. 1993), while an earlier study estimated that the indirect costs for community home care were more than twice the total direct costs and - by including the costs of premature death and lost productivity - were more than three times the total direct costs (Huang et al. 1988).
16. The 1996 annual costs (formal and informal) of caring for patients in a cross-sectional study of 679 Alzheimer’s Disease patients from 13 sites in the US were calculated (Leon et al. 1998). The costs for mild, moderate and severe AD patients were estimated as $18,406, $30,096 and $36,132 respectively. Increased institutionalisation with disease progression was the main reason for the direct association between cost increases and disease severity. On average it was estimated that 21% of total costs were for informal care.
17. A 1997 review of economic research in the United States concluded that ‘estimates of the national costs of Alzheimer’s Disease are open to a large margin of uncertainty’ (Ernst and Hay 1997). The estimates of the costs of Alzheimer’s Disease were found to vary widely across the ten studies assessed, which the authors attributed to four factors: the process of care varies across regions, there were marked differences in the levels of dementia in the studies, estimates were not standardised by socio-demographic factors or local availability of AD care, and there were differences in study dates. However, the studies consistently showed that the costs of unpaid caregiver time made up an especially large share of the overall costs. The ratio of indirect to total costs was far higher for those patients living at home than in other settings, but remains high overall because most AD patients live at home.
18. A feature of Alzheimer’s Disease is the tendency of both direct and unpaid caregiver costs to increase with the severity of the patient’s dementia or disability (Kavanagh and Knapp 1999). As might be expected, the positive association between unpaid caregiver costs and the degree of impairment seems to hold only for home-dwelling patients, and the major increase in direct costs occurs when the patient is institutionalised.
19. From figures produced by the NHS Executive, it was calculated that the direct health and social services costs of dementia to the public sector amounted to at least £850 million in England in 1992/93 (equivalent to £973 million in 1998/99) (Knapp et al. 1998).
20. Kavanagh and Knapp (1999) calculated the costs for elderly people with cognitive disability. This work was based on the OPCS disability surveys from the mid 1980’s, one survey for the general population and one for those living in communal establishments. No costs were attached to support by care-givers, and the study looked only at those over 65. For many services the porportion of people using them and the intensity of use increased with the degree of cognitive disability. For those living in households mean service costs rose with cognitive disability - (this relationship was statistically significant). The full costs of placement in communal establishments plus the costs of external services used varied between the sectors. The average mean cost per person per week was £199.63 (98/99 prices) for those with no cognitive disability, £278.86 for those with mild/moderate cognitive disability and £338.38 for those with severe cognitive disability.
21. A cost of care study for AD in the UK was based on in-depth interviews with 128 patients with the illness, 128 caregivers and 56 matched controls (Souêtre et al. 1999). The patients were all non-institutionalised with no major comorbidities. Based on ratings on the Mini Mental State Examination, patients were divided into those with mild (MMSE score >18), moderate (10-18), and severe (<10) AD. Direct medical costs for the three groups included hospitalisation, short-term institutionalisation (including respite episodes), medical consultations, paramedical services, laboratory and diagnostic tests and medication. Direct non-medical costs included community care centres, social services, personal expenses and transport, and equipment and home modification related to the illness. Indirect costs calculated were lost productivity due to caregivers giving up work and caregiver time spent with the patient (for previously employed caregivers, double counting was avoided). These hours of indirect cost were costed using average daily GDP in 1993 of £92.74.
22. It was estimated that the annual total costs (1993 prices) increased with disease severity, with averages of £26,464, £41,000 and £54,372 for the mild, moderate and severe patients respectively, compared to £1548 for the control group. Indirect costs represented 74.4%, 62.9% and 68.4% of the total cost for these three groups respectively. Indirect cost mainly comprised time spent with the patient (average of 47.6 hours per week, increasing with disease severity) rather than lost productivity (the average age of caregivers was 68.) There was a shift in the type of resources as the disease progressed, with hospitalisation and institutionalisation increasing directly with disease progression. Costs unrelated to AD (those apparently resulting from general ill-health in this elderly group) for AD patients were similar to costs for their caregivers and controls, suggesting that the study questionnaire made this distinction accurately.
23. Informal care may have been overestimated in this study as a number of caregivers of people in the severe group reported spending 12-24 hours per day in a carer capacity (although the average of 47.6 is not out of line with other studies). This study only looked at people living in the community, when the highest costs associated with AD are for hospital or institutional care.
24. The burden of illness of Alzheimer’s Disease among people aged over 65 in England was estimated in 1990/91 (Gray and Fenn 1993). Costs included hospital admission (inpatient and outpatient), GP visits, residential and home care services. Due to difficulties in calculating costs of informal care, this study looked only at the burden paid for through the Department of Social Security - attendance allowance (payable to anyone who requires care by another person due to mental or physical disability). Total cost was estimated at £1.04bn in 1990/91 prices (equivalent to £1.45bn in 1998/99), 94% of which was accounted for by people aged over 75. Long-term residential care accounted for more than 65% of the total.
25. Bosanquet et al (1998) also calculated the costs of AD in the UK. Two approaches extrapolated and combined available data to provide prevalence estimates of the illness in five-year age bands. Integrating these data with Government Actuary Department population estimates (which include an element for increasing life expectancy over time), current and future AD populations were estimated. The estimate of between 459,100 and 504,400 AD patients for 1996 is in line with the 500,000 estimate by Alzheimer’s Disease Society and by a recent York Centre for Health Economics study. The proportion of these patients with mild, moderate and severe illness was estimated using data from a UK study (O'Connor et al. 1991).
26. Bosanquet and coleagues distinguished four groups:
· AD sufferers under 65 years
· mild AD sufferers over 65 years
· moderate AD sufferers over 65 years
· advanced AD sufferers over 65 years
The proportion of the advanced and moderate over 65 groups looked after in the various settings was based on 1985/86 OPCS survey data (Office of Population Censuses and Surveys (OPCS) 1986). For mild AD and younger AD sufferers it was assumed that 21% lived alone and 79% were cared for in private households they shared with others.
27. Using the Hospital and Community Health Services Revenue indices, cost estimates by the Personal Social Services Research Unit (Kavanagh and Schneider) for various care settings were updated to 1996 levels (household living alone, household living with others, LA residential homes, private and voluntary residential homes, private and voluntary nursing homes, hospital care) (Schneider et al. 1993), (Kavanagh et al. 1995). The original costing was based on service use for those with advanced cognitive impairment in the 1985/86 OPCS survey. The validity of the data depends on the accurate powers of recall of the person interviewed as this was a retrospective survey. In many cases an informant provided these data, which may have further introduced bias. Informal care costs were based on work by Netten in 1989 using a survey of 285 carer households. These costs included direct financial expenditure (increased laundry and heating, changing house structure, travel expenses associated with caring). It was estimated that 11.1 hours per week (including travel time) was spent on caring when the elderly person lived separately and 12.4 hours when the carer shared the household. Waged time was costed at the value of lost production, and non-waged time at the private domestic rate. Lost earnings due to missed opportunities for carers who had taken time off work were also included. Although the methodology behind this estimate of informal care costs appears sound, the calculation is lower than in other studies, due to the low estimate of carer hours devoted to the AD patient.
28. The original cost study did not include younger or mild AD patients, who were assumed by Bosanquet to have less need for care (so that the cost of caring for them was adjusted downwards), based on a Canadian survey-based study which estimated that the net average cost of community care for moderate and severe patients is 2.27 times more than for mild patients (Østbye and Crosse 1994). A similar figure was found in the US in 1985 (Hu and Huang 1986).
29. The annual cost of Alzheimer’s Disease in the UK in 1996 was estimated by Bosanquet to be between £5.406bn and £5806bn. However, the care packages included all costs, for instance housing and personal consumption, some of which cannot be directly attributed to AD care. Costs were found to vary according to care setting and were highest in institutional care. Forty-two per cent of the total cost fell to the NHS and local authorities, and the remainder to the DSS, patients and families. From a purely NHS perspective, apart from patients residing in hospital the highest cost AD patients are those living at home (estimated to be around 80% of all AD sufferers). For those living alone, informal care was 23% of the total cost, and 28% for those living with others.
30. Bosanquet et al (1998) estimated that between 1996 and 2021 the number of cases of Alzheimer’s Disease is expected to increase to between 619,000 and 663,100 - an increase of between 23% and 44%. On the basis of the Bosanquet figures, the cost associated with AD patients would rise to between £7.20bn and £770bn over this period.
31. Table 1 summarises the UK and English evidence to date on cost of illness of AD.
Table 1. Summary estimates of the cost of Alzheimer’s Disease in England or the UK, inflated to £1998/99 prices (no studies include a component for lost life years)
Disease Cost 98/99 £billion
||Informal care as a
proportion of total cost
||Lost production by AD
||Only included attendance allowance
||England, >65 years only
||Combined costs with
Bosanquets ‘96 prevalence estimates. Only non-institutionalised
||5.8 - 6.2
||Max of 23% (those living with carer)
||Cost to the UK
Authors Total Alzheimer’s Disease cost 98/99 £billion Net cost Informal care as a proportion of total cost Lost production by AD patient included? Notes
Gray and FennSouêtreBosanquet 1.415.2 -17.25.8 - 6.2 noyesno Only included attendance allowance63-74%Max of 23% (those living with carer) nonono England, >65 years onlyCombined costs with Bosanquets ‘96 prevalence estimates. Only non-institutionalised patientsCost to the UK
32. Using Bosanquet’s estimates of prevalence in 2001 and updating the costs to 1998/99 prices, the total cost of illness of Alzheimer’s Disease currently would appear to lie between £6.17bn and £6.57bn. However, this includes an estimate of informal care which is lower than in most other studies - an estimated 11.1-12.4 hours of contact per week, compared to between 22 and 100 in the other studies cited above.
33. We therefore reworked our updated Bosanquet estimates under different assumptions (table 2). We estimated costs assuming that either 25.2 hours (as in Stommel 1994) or 45 hours (as in Cavallo and fattore 1997) per week is spent on informal care for those living in households. These hours were costed at both the minimum wage of £3.50 per hour and the cost of face-to-face contact with a home care worker of £9.39 per hour (Netten et al. 1999). We also included an average financial opportunity cost of caring of £2.56 per week (1998/99), covering costs of transport, extra heating, laundry, etc (Netten 1989). In the Netten study, 10.2% of carers identified missed training or job opportunities through being a carer, which were costed at £29.23 per week - equivalent to an average of £2.98 (1998/99) per AD patient with an informal carer (Netten 1989). As in the Bosanquet study, we reduced the costs of care for those less than 65 and those with mild AD by a factor of 2.27 relative to those with moderate and severe AD for all components of cost except informal care. The resultant estimates for 2000 are presented in table 2. It should be noted that these calculations still do not include the full costs of AD: there is no allowance for lost productivity due to AD patients leaving work before they would otherwise have retired, or for loss of life. On the other hand, these estimates do not subtract the cost that would be associated with these individuals in the absence of AD, that is they are gross costs. These estimates suggest that the cost of AD to the UK is between £7.06bn and £14.93bn.
Table 2. Total annual cost of Alzheimer’s Disease in 2000 under different assumptions about informal care volume and value (£1998/99 prices, billions)
hours of informal care per week
hours of informal care per week
The same costing exercise was conducted with the 98/99 prices and the predictions of future prevalence in 2021. It was found that assuming 25.2 hours of informal caring at a cost of £3.50 per hour and the lower 2021 prevalence estimate the cost of AD in the UK would be £10.7bn. However if we assume 45 hours of formal care at £9.39 per hour and the high 2021 prevalence estimate the cost of AD to the UK would be £20.7bn.
34. No cost of illness studies for cancer in the UK could be located. Here we therefore present the findings from studies that cover part of this disease area.
35. One recent study looked at direct costs for 20 patients with advanced colorectal cancer from two centres in each of five European countries: Belgium, France, Germany, Italy and the UK (Neymark and Adriaenssen 1999). Information on the medical resources used for managing these patients was collected by means of a retrospective examination of patient hospital charts. To make inter-country comparisons the authors used Belgian unit costs across all countries. Average imputed cost per patient varied from 570,000 BF to almost 1 million BF. Within each hospital, there was a large spread between the patients with the highest and the lowest imputed costs.
36. The direct treatment costs in Sweden for patients with cancer were calculated for the period 1985 to 1996 (Tennvall and Karlsson 1998). The total number of patients classified in ICD-9 diagnosis codes 140 - 239 treated in inpatient care remained almost constant over the period. Total inpatient costs in 1996 were 4,970,662 SEK in 1995, outpatient costs 794,328 SEK (1995) and annual drug sales for inpatient and outpatient care 102,186 SEK and 702,098 SEK respectively.
37. A report in 1997 calculated that the annual direct health care costs of prostate cancer in England and Wales was £45 million (Chamberlain et al. 1997).
38. A study in 1997 looking at inpatient episodes in Glasgow (population of 890,000) with a principle diagnosis of cancer found almost £23 million was spent. Of this the top 15 cancer sites made up 76% of the total cost. (Walker 1997).
39. A costing of malignant cancers in the Netherlands in 1988 suggested a total cost estimate of 1894 million Dutch guilders (dfl) (Koopmanschap et al. 1994). This represented 4.8% of total health care costs, with in-hospital care accounting for 60% of this total.
40. Cancer costs in a US HMO (covering 8.5 million members) were estimated from diagnosis until death or 15 years. Long-term costs to the HMO attributable to cancer (net costs) were: breast $35,000; colon $42,000; rectum $51,000; lung $33,000; ovarian $64,000; prostate $29,000 and non-Hodgkin’s lymphoma $48,000 (Fireman et al. 1997). These figures were based on resource use in the period 1987-1991. After subtracting non-cancer costs, cancer care for these seven cancer sites comprised 4.9% of all costs. This overall cost estimate is similar to that found in two earlier studies, one in California which estimated that the costs of cancer (all sites) was 5% of all spending on medical care (Rice, Hodgson et al. 1989), and one in 1990 for the US as a whole which estimated costs for cancer accounted for 4.8% of national health care spending (Brown 1990).
41. A consistent finding from these Dutch and US studies is that cancer appears to account for between 4.8% and 5.0% of direct health care costs. If this same proportion held for the UK, the total cost would have been £1.58bn to £1.65bn in 1998/99. We have not found it possible to include cost estimates for mortality or lost productivity.
42. The incidence of stroke increases with age, with more than 70% of stroke patients aged over 65. Someone having a stroke before age 65 has a one in three chance of never working again. Lost productivity thus represents a large economic burden. Estimates of the total cost of stroke vary widely - mainly due to different calculations of the indirect costs associated with mortality and disability. An estimate in 1993 pitched the total annual cost of stroke at $30bn, of which $13bn (43%) was indirect lost productivity cost (Dobkin 1995). Indirect costs were estimated to account for 58% of the total costs of first ever stroke in USA in 1990 (Taylor et al 1996), and 24% of total costs in Sweden (Terent et al. 1994). Studies from England, Scotland and the Netherlands estimated that stroke alone accounts for 3% to 4% of the direct costs of the health care sector (Porsdal and Boysen 1997).
43. A 1998 study estimated that £150m was being spent on day support and social care for stroke patients by social services in 1995-96. The total costs of stroke to the NHS and social services in England were calculated to be £2.49bn (updated to 1998/99 prices) (Bosanquet and Franks 1998). It was estimated that by 2023 the incidence of strokes would increase by 30%, with costs to the NHS and social services in England rising to £3.23bn (1998/99 prices) (assuming current care arrangements remain unchanged). The study found that stroke patients on average required 74 hours of care each week (based on the mid-1980s OPCS disability surveys), 7% of which was provided by district nurses or local authority home help. This translates to 192 million hours of informal care provided annually. At a minimum wage of £3.50 this would cost £672m. It was also estimated that 2100 carers would have to give up work to care for a relative following a stroke, at a cost of £30.5m per annum. Including these informal care costs, the total cost of stroke would amount to £3.20bn (1998/99 prices).
Heart disease studies
44. A recent study calculated the direct prevention, treatment and rehabilitation costs of coronary heart disease in 1996 to be £1.58bn (equivalent to £1.70bn at 1998/99 prices),(Maniadikis 1997). The indirect costs associated with informal care were £2.19bn (equivalent to £2.35bn at 1998/99 prices), with the costs of lost production due to mortality and morbidity being £1.63bn (£1.74bn) and £4.76bn (£5.10bn), respectively. The method used in the calculation of lost production costs assumed that employees would not be replaced. The total cost of coronary heart disease was therefore estimated to be £10.1bn (equivalent to £10.8bn at 1998/99 prices).
45. The available evidence clearly shows that AD is a high cost disease in the UK. Substantial costs are borne by the NHS, local authorities, social services and patients and their families. Putting a single figure on these costs is not easy, as there are many ways to calculate them. The difficulties are compounded when we try to compare across disease areas. With these caveats in mind, we summarise the best available figures on the costs of illness of AD (both highest and lowest cost from table 2), cancer, stroke and heart disease in figure 1. The costs are broken down where possible into direct health and social care costs and informal care (including estimates of lost productivity to the carer). The coverage of costs is not the same across the four diseases.
46. Although it is difficult to make comparisons between studies, Figure 1 reveals some interesting differences
‘Within the next 20 years, AD will likely surpass heart disease and cancer as the most costly disease in America…. Medical interventions are becoming more promising, and with them comes the possibility that the economic burden of the disease may be lessened and, more important, clinical benefits realised. By improving cognitive function with pharmacotherapy, it is possible to reduce caregiver time, delay nursing home placement, and improve quality of life' (Meek et al. 1998, p.72)
47. An aim of the work underpinning this paper was to identify the overall level of R&D funding in relation to each of four major causes of mortality in the UK: Alzheimer’s Disease, cancer, heart disease and stroke. We sought funding data on all research beginning in 1999. We assumed that this would provide an approximate measure of the annual volume of research funding, on the assumption that funding on research in these disease areas remains fairly stable across neighbouring years. Where it was not possible to find a figure for research funding which began in 1999, we tried to establish the most recent figure for annual expenditure on research. We initially hoped to distinguish basic from health services research (although this proved impossible to achieve).
48. We asked all major funders in the UK to provide us with details of projected funding for all research beginning in 1999. Contact was made with the Department of Health, all NHS regional executive bodies, major grant providers (such as the MRC, Wellcome Trust and Nuffield Foundation) and all relevant charities. A contact list of charities which fund research in these disease areas was established via communication with known bodies and a search of the web (in particular the Association of Medical Research Charities, and CaritasData which ranks the top 500 UK charities). A list of all bodies contacted is shown in Appendix A. In most cases initial contact was made via a letter with a follow-up call within the next two weeks, although in some cases initial contact was via a telephone call. With all non-government organisations we checked whether any of their funding for research came from external grants (such as from the DH or MRC) to ensure that we did not double count. We also undertook a search of the latest version of the National Research Register (NRR) which seeks to provide a systematic record of R&D projects funded by, or of interest to, the NHS. We did not attempt to establish sums of money that trusts themselves spend on research. However the majority of trusts' ‘own account’ research is funded from support fund monies and is therefore covered in our analysis of breakdown of support fund spend.
49. Where funding for research was on dementia rather than specifically on Alzheimer’s (this was the situation for one of the charities and the MRC) we assumed that 60% of the funding could be attributed to Alzheimer’s based on estimates that 60% of dementia cases have Alzheimers.
50. The Department of Health spends around £500m per annum on research (Department of Health 2000). Of this, £360m is on service support for NHS providers (managed via the NHS Regional Executives) involved in R&D. This support funding is intended to cover the cost of hosting R&D activities (staff time, use of facilities and equipment), rather than to cover the basic costs of treating patients involved in R&D which will often be funded by other bodies. This funding comes as either portfolio funding or task-linked funding (Department of Health 2000).
51. Attributing support funding for R&D to different disease areas is particularly complex. Large proportions of this funding are provided in lump sums to NHS Trusts which then allocate resources to individual projects. However, based on local knowledge and reports of individual providers it is possible to attribute 46% of 1998/9 R&D support money to five disease areas [personal communication from R&D funding branch]. Table 3 summarises these figures.
Table 3. R&D support spending in England (1998/99)
||% of total
52. Of the 46% of support funding that can be attributed, 18% is spent on cancer research and 13% on CVD/stroke research. Spending on AD comes within the mental health category. It is not possible to tell exactly how much of this 9% is for AD. However between April 1999 and March 2000 the South London and Maudsley NHS Trust was allocated £18.8m. This Trust specialises on mental health and neurological conditions and so the majority of this money would fall in the category of Mental Health in table 4. Of this £18.8m, only £1.2m (6.3%) was spent in the category of ‘neurodegeneration and the elderly’, of which only a proportion would be for Alzheimer’s Disease. If we assume that all of the South London and Maudsley NHS Trust support money is for ‘Mental Health’ and that all of the ‘neurodegeneration and the elderly’ money is for AD, this would suggest that £2.07m (0.6%) of support money nationally is for AD. This is very likely to be an overestimate due to the second assumption made.
53. These figures suggest that in terms solely of support funding, research support for AD is less than 1/30 of the support funding for cancer research and less than 1/20 of that received by CVD and stroke research combined.
54. Around £70m of the total Department of Health R&D spend is on the NHS R&D programme where commissioning is largely undertaken through specific programmes of research (these include separate programmes on cardiovascular disease and stroke, cancer and mental health). However, these separate programmes will shortly be recast into three more general programmes. The mental health programme was launched in 1991, with a total of £5m committed to 32 commissioned projects, all of which were completed by the end of 1998 (and so falls outside our inclusion criteria). The CVD and stroke programme was launched in 1993 and funded £11.8 million of projects. The cancer programme was launched in November 1993. Some spending on research continued into 1999 for projects in CVD, stroke and cancer and so were included in our figures.
55. NHS regional executive offices also have responsive research funding schemes, and NHS Trusts fund research. Time did not allow us to contact individual Trusts. However, the majority of trusts’ ‘own account’ research is funded from support fund monies and is therefore already included in the analysis reported above. For the remainder of Trust spend on research we have assumed that their pattern of research funding does not differ significantly from that conducted at the National and Regional level.
56. The total level of research funding identified for 1999 is shown in table 4 and figure 2. It can be seen that the amount spent on research into Alzheimer’s Disease is substantially lower than amounts in the other three areas, as was found when we assessed the R&D support funding expenditure.
Table 4: Information on levels of funding on R&D in 1999 (£ million)
DoH, NHS (national and regional) and MRC
57. The total amount spent on research in AD is only 57% of the amount spent on stroke, 9.6% of the amount spent on heart disease and 3% of the amount spent on cancer. Looking at medical charities’ funding of research, spending on cancer is 82 times greater than AD, heart disease 23 times greater than AD, and stroke over twice that for AD.
58. Our method was to choose one year to assess spend on R&D. Naturally there are peaks and troughs in terms of spending on R&D in different areas and it is not possible to tell whether the 1999 spend was high or low for any of our four categories relative to an average year.
59. We also hoped to calculate the amount spent on R&D by pharmaceutical companies. Contact with the Association of the British Pharmaceutical Industry and the Centre for Medicines Research International suggested that this information is not centrally held for the UK, although the Centre for Medicines Research International plans to examine this in the future (personal communication). The only figures we were able to find came from the Cancer Research Campaign which estimated that in 1997/98 £114.03m was spent by the pharmaceutical industry on cancer research.
60. To assertion the amount spent on research for each person with the disease in the UK the total research funding for each disease was divided by prevalence figures. Prevalence is the number of people with the disease alive on a given date (affected by both the incidence rate and disease survival rate), and varies according to methodology used in its calculation. The prevalence figures used in this instance and their source are shown in table 5. The estimated annual level of research funding per person with the disease is shown in table 6.
Table 5: Number of people with disease in the UK, 1999
||`Estimated no with disease (000's)
||Source of prevalence data
||The mid-point of 2 prevalence estimates by Bosanquet (Bosanquet, 1998)
||The General Practice Research Database, covers
2.1 million of the population of England and Wales, with an age
distribution similar to the total population.
Prevalence of stroke in 1996 was based on diagnosis by ICD9
430-436, excluding transient ischaemic attacks (TIAs), recorded in
the last 12 months (Office of National Statistics, 1998)
||The General Practice Research Database, covers
2.1 million of the population of England and Wales, with an age
distribution similar to the total population.
Prevalence of coronary heart disease in 1996 was based on
diagnosis of CHD (ICD9 410-414) ever-recorded and treatment in the
12 months (Office of National Statistics, 1998)
||Office of National Statistics compiled validated
and complete information from all regionally based cancer registries
in England and Wales (Office for National Statistics, 1999). Found 1.1 percent of the population of England and Wales who
were alive on 1 January 1993 had been diagnosed with cancer
(excluding non-melanoma skin cancer) in the previous ten years.
Note: All prevalence estimates were applied to mid-1999 UK population statistics as supplied by the office of national statistics to obtained the estimated number living with the disease.
61. Table 6 shows that less is spent on research for each person living with AD than stroke, heart disease or cancer. In terms of government and MRC spend, £7 is spent on research per person with AD compared to £8 for heart disease, £44 for stroke and £51 for Cancer. The differences become even greater when we look at spend on research from all sources, with research spend per person with heart disease being 2.5 times greater than AD, stroke over 7 times greater than AD and cancer over 26 times greater. Care needs to be taken in interpreting these figures. They do not necessarily imply that AD is under-funded as there is no clear rationale that states that the ratio of research spend to prevalence should be equal across diseases. Any statement about the appropriate ratio of research spend per measure of disease morbidity will necessarily involve a value judgement.
Table 6: Research spend per individual living with disease
level of research funding per person with disease
62. We have estimated that the gross cost of AD in the UK is between £7.06bn and £14.93bn. This figure does not include any allowance for lost productivity or for the costs of mortality. Costs are highest where patients are cared for in institutional care. Less than half (42%) of the total cost is borne by the NHS and local authorities, with the remainder falling to the social security branch of government, and patients and their families.
63. AD ranks alongside better known chronic disorders such as stroke and diabetes in terms of mortality yet for the population aged over 80 its prevalence is much greater. The prevalence of AD and the associated costs are set to rise substantially over the next 20 years. We have estimated that assuming care patterns continue as at the present time the cost of AD in 2021 will rise to between £10.7 - £20.7 bn (in 98/99 prices). These costs will fall mainly to the NHS, social services and families of sufferers.
64. Comparison of the cost of illness of AD, cancer, heart disease and stroke, although hampered by the data limitations and methodological differences, shows the considerable absolute and relative economic impact of AD in the UK.
65. Expenditure on R&D is significantly lower for AD than it is for stroke, heart disease or cancer. These differences are formed both when looking at total research expenditure and at its main component sources. The difference between research on AD and other diseases is especially marked in relation to charity spending. The situation in the UK would appear to mach that in the US:
‘Given the magnitude of potential savings - including savings to public health care programs - if AD could be prevented or arrested at an early age, a substantially increased federal commitment to AD research is clearly warranted. Since private research spending on AD is unknown….there are few signs that public or private research is being undertaken at a scale warranted by the magnitude of the disease’s costs’[Hay, 1987 #15]
We are very grateful to the Alzheimer’s Research Trust for funding of this work, to the HEED team for considerable support, and to R&D departments at the DH and the NHS Regions and within charity organisations for their time helping us gather this information
Appendix A: List of organisations contacted for information on funding on Research and Development
NHS HTA programme*==
NHS funding on support for providers*=
Department of Health*=
National R&D Programme Cancer*=
National R&D Programme Heart Disease*=
National R&D Programme Stroke*=
National R&D Programme AD*
NHS executive Northern and York*=
NHS executive London regional office*=
NHS executive South East*=
NHS executive North West
NHS executive West Midlands*=
NHS executive Trent*=
NHS executive South and West*=
NHS executive Eastern *=
Scottish Executive health department*
DHSS Northern Ireland
Medical Research Council*=
Joseph Rowntree Foundation*=
Economic Social Research Council*
Alzheimer's research Trust*=
Alzheimer Scotland - Action on Dementia*=
Mental Health Foundation*=
Psychiatric Research Trust*=
Cancer Research Campaign*=
Imperial Cancer Research Fund*=
British Heart Foundation*=
The Stroke Assoication*==
Breast Cancer Research Trust*=
Richard Dimbleby Cancer Fund*
North of England Children’s Cancer Research Unit*=
Breast Cancer Campaign *=
Marie Curie Cancer Care*=
CHEST Heart and Stroke - Scotland*=
Leukemia Research Fund *=
National Heart Research Fund*=
Yorkshire Cancer Research *=
World Cancer Research Fund *=
Breakthrough Breast Cancer*=
Quest Cancer Research
Prostate Research Campaign UK *=
Association for International Cancer Research *=
Action Cancer *=
Northern Ireland - chest heart and stroke *=
Cancer Prevention Research Trust *=
Roy Castle lung cancer foundation
Bob Champion Cancer Trust *=
Clatterbridge Cancer Research Trust *=
Colon cancer concern
Ulster Cancer foundation *=
North West Cancer Research Fund *=
Orchid Cancer Appeal *=
Royal Marsden Cancer Fund
Gynaecology Cancer Reserch Fund
War on Cancer*=
Ludwig Institute for Cancer Research*=
British Society for Research into ageing*=
*Response rate - all organisations we heard back from
=- Figures for research found - included in table 4
Bosanquet, N. and P. Franks (1998). Stroke care: reducing the burden of disease. London, The Stroke Association.
Bosanquet, N., J. May, et al. (1998). Alzheimer's disease in the United Kingdom. Burden of disease and future care, Imperial College of Science, Technology and Medicine, University of London.
Brown, M. (1990). The national economic burden of cancer: An update. Journal of the National Cancer Institute 82: 1811-1814.
Cavallo, M. and G. Fattore (1997). The economic and social burden of Alzheimer's disease on families in the Lombardy region, Italy. Alzheimers Disease Association Disorders: in press.
Chamberlain, J., M. S. Moss, et al. (1997). Report prepared for the health technology assessment panel of the NHS executive on the diagnosis, management, treatment and costs of prostate cancer in England and Wales. British Journal of Urology 79(Suppl.3): 1-32.
Department of Health (2000). Department of Health/NHS Research - Organisation of DH R&D: An Overview, www.doh.gov.uk.
Department of Health (2000). R&D Support funding for NHS providers ("Culyer" Funding), www.doh.gov.uk.
Dobkin, B. (1995). The economic impact of stroke. Neurology 45(suppl. 1).
Ernst, R. and J. Hay (1997). Economic research on alzheimer disease: A review of the literature. Alzheimer Disease and Assoiciated Disorders 11(Suppl. 6): 135-145.
Fireman, B., C. Quesenberry, et al. (1997). Cost of care for cancer in a health maintenance organization. Health Care Financing Review 18(4): 51-76.
Gray, A. and P. Fenn (1993). Alzheimer's Disease: The burden of the illness in England. Health Trends 25(1).
Hay, J and R. Ernst (1987). The economic costs of alzheimers disease. American Journal of Public Health 77(9).
Hu, T. and L. Huang (1986). Evaluation of the costs of caring for the senile dememted elderly: a pilot study. Gerontologist 26: 158-63.
Huang, L., W. Cartwright, et al. (1988). The economic cost of senile dementia in the United States, 1985. Public Health Reports 130(1): 3-7.
Kavanagh, S. and M. Knapp (1999). Cognitive disability and direct care costs for elderly people. The British Journal of Psychiatry 174: 539-546.
Kavanagh, S., J. Schneider, et al. (1995). Elderly people with dementia: Costs, effectiveness and balance of care. The Economic Evaluation of Mental Health Care. K. MRJ. Aldershot, Arena.
Knapp, M., D. Wilkinson, et al. (1998). The economic consequences of alzheimer's disease in the context of new drug developments. International Journal of Geriatric Psychiatry(13): 531-543.
Koopmanschap, M. A., L. van Roijen, et al. (1994). Current and future costs of cancer. European Journal of Cancer 30A(1): 60-65.
Leon, J., C. Cheng, et al. (1998). Alzheimer's disease care: Costs and potential savings. Health Affairs 17(6): 206-216.
Manton, K., L. Corder, et al. (1993). Estimates and projections of dementia-related service expenditures. Forcasting the Health of the Oldest Old. S. B. Suzman R, Manton K.
Max, W. (1998). Cost of illness of dementia. Health economics of dementia. J. B. Wimo A, Karlsson G, Winblad B. Chichester, UK, Wiley.
Meek, P., E. McKeithan, et al. (1998). Economic considerations in alzheimer's disease. Pharacotherapy 18(2 Pt 2): 68s-73s.
Netten, A. (1989). An approach to costing informal care. Discussion Paper 637, PSSRU, University of Kent, Canterbury.
Netten, A., J. Dennett, et al. (1999). Unit costs of health and social care 1999. Kent, PSSRU, University of Kent at Canterbury.
Neymark, N. and I. Adriaenssen (1999). The costs of managing patients with advanced colorectal cancer in 10 different European Centres. European Journal of Cancer 35((13)): 1789-95.
Nuttall, S., R. Blackwood, et al. (1993). Financing long-term care in Great Britain. Journal institute of Actuaries(121): 1-68.
O'Connor, D., P. Pollitt, et al. (1991). Does early intervention reduce the number of elderly people with dementia admitted to institutions for long term care? British Medical Journal 302: 871-75
Office for National Statistics (1999). Health Statistics Quarterly. London, Office for National Statistics: 59-63.
Office of National Statistics (1998). Key Health Statistics from General Practice 1996. London, Office of National Statistics.
Office of Population Censuses and Surveys (OPCS) (1986). Survey of disability among adults in private households, 1985. London, OPCS.
Østbye, T. and E. Crosse (1994). Net economic costs of dementia. Canadian Medical Association Journal(43): 108-112.
Porsdal, V. and G. Boysen (1997). Cost of illness studies of stroke. Cerebrovascular Disease 7: 258-263.
Rabins, P., N. Mace, et al. (1982). The impact of dementia on the family. Journal American Medical Association 248: 333-5.
Rice, D., P. Fox, et al. (1993). The economic burden of alzheimer's disease care. Health Affairs 12: 165-176.
Rice, D., T. Hodgson, et al. (1989). The economic burden of cancer, 1985: United States and California. Cancer Care and Cost. S. R. e. a. eds., Ann Arbor, MI. Health Administration Press.
Schneider, J., S. Kavanagh, et al. (1993). Elderly people with advanced cognitive impairment in England: Resource use and costs. Ageing and Society 13: 27-50.
Smith, K. and K. Wright (1996). Costs of mental illness in Britain. Health Policy 35: 61-73.
Souêtre, E., R. Thwaites, et al. (1999). Economic impact of alzheimer's disease in the United Kingdom. Cost of care and disease severity for non-institutionalised patients with alzheimer's disease. British Journal of Pyschiatry(174): 51-55.
Stommel, M., C. Collins, et al. (1994). The costs of family contributions to the care of persons with dementia. Gerontologist (34): 199-205.
Taylor, T., P. Davies, et al. (1996) . Lifetime cost of stroke in the United States. Stroke 27:1459-1466.
Tennvall, G. R. and G. Karlsson (1998). “Cancer treatment in Sweden - costs of drugs, inpatient and outpatient care from 1985 to 1996 and cost effectiveness of new drugs.” Acta Oncologica 37(5): 447-453.
Terent, A., L.-A. Marke, et al. (1994). Costs of stroke in Sweeden. A national perspective. Stroke 25: 2363-2369.
Walker, A. (1997). The economic burden of cancer. Journal of Managed Care 1: 72-75.
Wimo, A., E. Witthaus, et al. (1998). Economic impact of introducing propentofylline for the treatment of dementia in Sweden. Clinical Therapeutics 20(3): 552-566.