Societal Costs of Vascular Cognitive Impairment in Older Adults
Background and Purpose— The construct of vascular cognitive impairment (VCI) includes many whose care is or will be costly. Nevertheless, estimates of these costs are not well described. We therefore set out to estimate the societal costs of VCI in elderly people.
Methods— In a secondary analysis of the Canadian Study of Health and Aging, a representative cohort study, Canadian dollar costs using a societal perspective were estimated by standard methods.
Results— The total annual per-patient societal costs for VCI by severity were $15 022 for those with mild disease, $14 468 for those with mild to moderate disease, $20 063 for those with moderate disease, and $34 515 for those with severe disease. The most expensive component per individual was the cost of institutional long-term care. Although severe impairment was associated with higher costs, the extent of institutionalization at all levels of severity and less drug use among those more severely impaired mitigated a severity-cost gradient.
Conclusions— The societal costs of VCI are not inconsiderable. In contrast to Alzheimer disease, there is no clear gradient relating cost to severity. Unpaid caregiver costs are an important aspect of societal costs, even in those with only mild impairment.
In recent years, the idea of vascular cognitive impairment (VCI) has been advanced as an alternative to the more narrowly construed notion of vascular dementia.1,2⇓ Compared with vascular dementia, VCI includes patients with mixed vascular and neurodegenerative dementias, as well as patients with cognitive impairment of vascular origin who do not meet the traditional dementia criteria.2,3⇓ The latter patients include 2 groups: those with important cognitive and functional impairment who nevertheless do not meet the criteria for dementia (eg, hemiparesis and aphasia, followed left middle cerebral artery stroke) and a group, akin to people with amnestic mild cognitive impairment, who potentially have a “predementia” syndrome.4
Patients with VCI comprise a spectrum that includes many whose care is or will be costly. Nevertheless, to date there appear to be no published estimates of the cost to society of VCI. We report here population-based Canadian data on the societal costs of providing care for people with VCI.
Materials and Methods
To provide context for these estimates and to aid comparisons with Alzheimer disease, we followed the cost estimation protocol of Hux and colleagues.5 They estimated the costs of Alzheimer disease from a societal point of view, taking into account direct care costs and estimating unpaid costs of time by informal caregivers.
The data are drawn from the Canadian Study of Health and Aging (CSHA), a representative cohort study of older (>65 years) Canadians undertaken chiefly to estimate the prevalence and incidence of dementia6,7⇓ and risks for Alzheimer disease and vascular dementia.8,9⇓ Briefly, in 1991 to 1992, community-dwelling subjects were sampled in a 2-stage sample, clustered by area, and stratified by age. These people were screened with the Modified Mini-Mental State Examination.10 Those who screened positive (≤77 of 100) and a sample who screened negative were invited to a detailed clinical11 and neuropsychological examination.12 In addition, a similar sample of institutionalized residents was examined.11 Follow-up interviews were carried out after 5 years and included interviews with knowledgeable informants, even for those participants who had died in the interval.3,7⇓
The present analysis relies on cost figures produced for the 1998 analysis of Hux et al.5 A number of assumptions are noteworthy. The costs are in Canadian dollars. (As of February 1, 2002, Can $1=US $0.62US, Eur $0.73, and 0.44£.) The estimated cost of institutional care is based on the cost of providing care in Ontario, the most populous province in Canada. Because this included no estimate of the cost of institutional care for those with mild impairment, we were obliged to extrapolate that number from the relationship between cost and severity for the other levels of severity for which costs were known by using a regression equation and noting that the fit of the line was high (R2=0.995). Institutions included nursing homes and other facilities providing domiciliary care with personal assistance. The cost of community care came from estimates of the Ontario Ministry of Health, supplemented by additional data from the City of Hamilton Homecare Placement Day-Center Program. The cost of unpaid caregiver time was assigned a value equivalent to the 1998 Ontario minimum wage of Can $6.85/h.
Medication was coded by use of the World Health Organization’s Anatomical Therapeutic Chemical Classification (ATC) Index,13 which also lists defined daily doses for medication. A defined daily dose is “the assumed average maintenance dose per day for a drug used for its main indication in adults.” For products not found in the ATC index, the Compendium of Pharmaceuticals and Specialties14 was consulted, and the defined daily dose was based on the recommended maintenance dose for adults. Because we did not know the doses prescribed, we assumed that drugs were prescribed in defined daily doses. The cost per defined daily dose was calculated by use of the suggested retail price, excluding pharmacist’s dispensing fee, listed in the 1996 Atlantic Pharmaceutical Services Incorporated Pricing Guide.15 The price was calculated using the least expensive product because we did not know which manufacturers for multisource products were used. If the drug was not supplied in a dosage equal to the defined daily dose, it was calculated by either adding 2 dosage forms together (eg, to calculate a defined daily dose of 5 mg, the prices of the 1- and 4-mg dosages were added) or dividing the price by the appropriate denominator (eg, to calculate a defined daily dose of 5 mg, the price of the 10-mg dose was divided by 2). This price was then multiplied by 365 to obtain price per year. Only drugs in the following 4 categories were included: anticoagulant or antiplatelet, antihyperlipidemic, antihypertensive, and psychotropic (ie, antidepressants, antipsychotics, anxiolytics) medications.
To estimate severity of illness, we aimed to take into account that the construct of VCI is likely to have stages distinct from the usual cognitive staging in Alzheimer disease. Most staging measures comprise estimates of functional and cognitive impairment.16 In Alzheimer disease, such impairment generally precedes in parallel, but the same assumption appears less tenable in VCI. Folstein et al17 based their severity estimate on the Mini-Mental State Examination Score (MMSE). Our estimate reproduces their cut points but also includes the global clinical measure of severity used in the CSHA clinical consensus diagnosis.11 When the clinical impression of severity and MMSE accorded, we used the Folstein et al17 MMSE cut points. When the clinical assessment of severity and the MMSE cut points did not accord, we used the higher (less severe) score.
To estimate costs associated with the use of community services and caregiver time, we used data from the CSHA Caregiver Survey. This survey was undertaken with all people identified as “primary caregivers” of participants with cognitive impairment, as well as a comparison group who were caregivers, or potential caregivers, for elderly persons who were cognitively normal. Estimates in this report came from the caregivers of those identified as having VCI. Details of the caregiver interview, which included information on functional impairment, caregiver health and burden, and use of formal and informal services, are outlined elsewhere.5,18⇓ Informal caregivers were asked about direct care activities such as assistance with activities of daily living and supervision in the case of those who could not be left alone in the house. These elements were costed separately.
Table 1 presents the proportion of people with VCI in the CSHA who were in long-term residential care and the expected cost of care per subject by level of disease severity. On a per-person basis, costs increase with disease severity. A significant proportion of those with mild (41.4%) and mild to moderate disease (43.1%) are in institutional long-term care.
Table 2 presents the costs of monthly medication use. For community-dwelling people, there is no clear relationship between disease severity and the cost of either vascular or psychotropic medications. In contrast, among institutionalized respondents, the costs of vascular medications decrease with severity, whereas the cost of psychotropic medications increases as impairment becomes more severe.
Table 3 reports the use and costs of community support services. Costs vary substantially by stage. Those with the most severe impairment have the highest costs, but otherwise there is no evident gradient relationship between severity and cost. Use of respite care, caregiver support, and counseling was reported only infrequently.
Table 4 presents the results of estimating the cost of otherwise unpaid caregiver time. Unpaid costs in the community appear to be substituting for paid costs in institutions. Nevertheless, unpaid caregiver costs are still incurred for those who are in institutions.
The total effect of these various costs is summarized in the Figure, which portrays the mean annual cost of care for people with VCI by disease severity. From this and given the CSHA estimate, the societal costs (sum of direct care costs and unpaid informal caregiver time) by level of severity are $15 022 for patients with mild disease, $14 468 for those with mild to moderate disease, $20 063 for those with moderate disease, and $34 515 for those with severe disease.
We have estimated societal costs of vascular cognitive impairment, which appear to be considerable. The data suggest that treatment of VCI and, better still, its prevention will be important in managing illnesses associated with the aging of the population. Our data, however, are subject to some important caveats. The costs are derived from an era before symptom- or disease-modifying treatments for dementia became available. The costs do not include the cost of diagnosis. The perspective is simple and reflects societal costs, without attendant benefits being estimated. In addition, the costs are driven in large measure by the valuation of the costs of informal care. The best way to do this is not clear.19 Briefly, 2 methods predominate: All informal care can be costed at the rate that would be required to substitute paid formal care, or a value representing the opportunity cost of the earnings forgone to provide the care can be estimated. Our method, which is a compromise between these 2 perspectives, was to limit total informal costs to those needed to substitute formal care in a long-term care institution for informal care provided at home. One component of societal costs is not measured in this study, ie, market productivity losses attributable to VCI onset and progression. Such costs are probably small, however, because VCI onset typically occurs in older persons already retired from the labor force.
Compared with Alzheimer disease, in which staging is well understood and has a determining impact on disease costs,20 staging of VCI remains controversial. How best to summarize across subtypes of VCI (eg, VCI, no dementia, vascular dementia, mixed Alzheimer disease with vascular dementia) is not known. Our method was to compare the cognitive cut points of Hux et al5 with the clinical global severity measure.11 When they accorded, that level was used, with the cognitive cut point used to discriminate between mild and mild to moderate disease severity. When they were discrepant, we scored the patient at the higher level of function and lower level of disability. Although this is a more conservative means of estimating costs, it likely contributed to our finding of a less evident gradient relationship between severity and cost. Underlying this pattern (essentially of greater-than-expected functional impairment) is chiefly the motoric effects of stroke. Other factors appear to operate as well, including aspects of cognitive impairment that necessitate supervision but that are not well captured by cognitive measures that emphasize the predominance of memory impairment that is seen in Alzheimer disease.4 These considerations likely underlie the striking comparison to the report of costs of care for people with Alzheimer disease, in which no subjects with mild dementia were in institutions, with our finding that 41% of those with mild VCI were in institutional long-term care. In general, our cost estimates for VCI are slightly lower than the estimates of Hux et al5 for Alzheimer disease at all levels of severity. The 1 exception is for mild disease, for which the cost estimates for VCI are much higher. Specifically, the estimates contrast as follows: for mild disease: VCI, $15 022; Alzheimer disease, $9451; for mild to moderate disease: VCI, $14 468; Alzheimer disease, $16 054; for moderate disease: VCI, $20 063; Alzheimer disease, $25 724; and for severe disease: VCI, $34 515; Alzheimer disease, $36 794. One of the limitations of this work is that the comorbidities in these patients (eg, heart disease, diabetes) were not compared.
As has been noted,20 staging appears to have an impact on prescribing practices in VCI. Especially for those in institutional long-term care, pharmacotherapy of vascular risks is used less frequently as the impairment becomes severe.21 Nevertheless, the cost offset of medications is minor compared with the high prevalence of institutionalization (93%) among those with the most severe impairment.
The data reflect the Canadian situation in 1996 to 1997, when community services were more underdeveloped than today. For example, very few caregivers availed themselves of respite care or caregiver counseling and support. A third wave of the CSHA is now underway, and a comparison of the present with the earlier pattern will be of interest in this regard.
At all levels of severity, the largest direct cost of VCI is institutionalization. Such observations in other causes of dementia sometimes prompt a desire to see a larger proportion of patients maintained in the community. The societal cost perspective, however, allows us to see that such policies simply shift the focus of costs from formal to informal providers and even then do not account for costs to providers in terms of health and quality of life.
In our view, the chief merits of the present study are providing a baseline understanding of the costs of untreated VCI and drawing attention to the not inconsiderable societal costs associated with cognitive impairment in general. As we come to grips with this new construct, it is important that such information be known so that VCI as the subject for further study can be appreciated.
Dr Rockwood is supported by a Canadian Institute of Health Research Investigator Award and by the Dalhousie Medical Research Foundation as the Kathryn Allen Weldon Professor of Alzheimer Research. Dr Sketris is supported by a Canadian Institute of Health Research/Canadian Health Services Research Foundation and Nova Scotia Health Research Foundation Chair in Health Services Research. The data reported in this article were collected as part of the CSHA. The core study was funded by the Seniors’ Independence Research Program through the National Health Research and Development Program of Health Canada [project 6606-3954-MC(S)]. We thank Chole Campbell, BSc, for assistance with pharmacy costs. The VCI Investigators of the CSHA are Ian McDowell, University of Ottawa; David B. Hogan, University of Calgary; Réjean Hébért, Université de Sherbrooke; Christine Wolfson, McGill University; and René Verreault, Centre Hospitalier Saint-Augustin.
The Vascular Cognitive Impairment Investigators of the Canadian Study of Health and Aging are listed in the Acknowledgments.
- Received September 4, 2001.
- Revision received February 5, 2002.
- Accepted February 28, 2002.
- ↵Rockwood K, Wentzel C, Hachinski V, Hogan DB, MacKnight C, McDowell I. Prevalence and outcomes of vascular cognitive impairment. Neurology. 2000; 54: 447–451.
- ↵Wentzel C, Rockwood K, MacKnight C, Hachinski V, Hogan D, Feldman H, Ostbye T, Wolfson C, Gauthier S, Verreault R, McDowell I. Progression of impairment in patients with vascular cognitive impairment without dementia. Neurology. 2001; 57: 714–716.
- ↵Hux MJ, O’Brien BJ, Iskedjian M, Goeree R, Gagnon M, Gauthier S. Relation between severity of Alzheimer’s disease and costs of caring. Can Med Assoc J. 1998; 159: 457–465.
- ↵Canadian Study of Health and Aging Working Group. Canadian Study of Health and Aging: study methods and prevalence of dementia. Can Med Assoc J. 1994; 150: 899–913.
- ↵Canadian Study of Health and Aging. The incidence of dementia in Canada. Neurology. 2000; 55: 66–73.
- ↵Canadian Study of Health and Aging Working Group. Risk factors for Alzheimer’s disease in Canada. Neurology. 1994; 44: 2073–2080.
- ↵Hébert R, Lindsay J, Verreault R, Rockwood K, Hill G, Dubois MF. Vascular dementia: incidence and risk factors in the Canadian Study of Health and Aging. Stroke. 2000; 31: 1487–1493.
- ↵World Health Organization Collaborating Centre for Drug Statistics Methodology. Anatomical Therapeutic Chemical Classification (ATC) Index With Defined Daily Doses (DDD). Oslo, Norway: World Health Organization; 1999.
- ↵Gillis MC, ed. Compendium of Pharmaceuticals and Specialties. 31st ed. Ottawa, Canada: Canadian Pharmaceutical Association; 1996.
- ↵Atlantic Pharmaceutical Services Incorporated Pricing Guide. Halifax, Canada: Atlantic Pharmaceutical Services; 1996.
- ↵Rockwood K, Morris JC. Global staging measures in dementia.In: Gauthier S, ed. Alzheimer’s Disease: Diagnosis and Treatment. London, UK: Martin Dunitz; 1996: chap 9.
- ↵Canadian Study of Health and Aging Working Group. Patterns of caring for people with dementia in Canada. Can J Aging. 1994; 13: 470–485.
- ↵Koopmanschap MA, Brouwer WBF. Indirect costs and costing informal care.In: Wimo A, Johnson B, Carlson G, Winblad B, eds. Health Economics of Dementia. New York, NY: John Wiley & Sons; 1998.
- ↵Ostbye T, Crosse E. Net economic costs of dementia in Canada. Can Med Assoc J. 1994; 151: 1457–1464.