Alzheimer's disease (AD), the most common form of dementia, is the fifth leading cause of death in the United States (US) among persons aged 65 years or older [1
]. Approximately 5.3 million Americans have AD, 5.1 million of whom are over the age of 65 years [2
]. The financial burden of AD and other dementias is considerable, with an estimated annual total cost of $148 billion in the US in 2005 [3
]. The annual cost per patient in the US was found to be three times higher for persons with AD and other dementias relative to persons without AD or other dementias (i.e., $33,007 relative to $10,603 in 2004) [3
]. Other studies conducted in the US or elsewhere similarly concluded that AD costs were high and likely to rise over time due to the aging of the population [4
Although several medications have been approved to treat AD (e.g., cholinesterase inhibitors, memantine), some jurisdictions (e.g., United Kingdom [8
], Ontario, Canada [9
]) limit reimbursement of these drugs (e.g., reimbursement in the United Kingdom is limited to persons with moderate AD only). The most prominent reasons for limiting reimbursement are grounded in the fact that the medications treat the symptoms of AD only (they are not a cure) [10
] and cost between $2.90 and $6.80 per pill (US figures) [11
]. In a cost-containment environment, economic evaluations are playing an increasingly important role in pricing and reimbursement decisions, [12
], especially for relatively expensive medications that have modest efficacy.
In AD, cognitive decline continues when persons with the disease take cholinesterase inhibitors or memantine [10
]. Consequently, an important outcome to consider when assessing these drugs in economic evaluations is health-related quality of life (HRQoL). In economic evaluations, HRQoL is usually expressed as part of a quality-adjusted life year (QALY). A QALY is a composite measure of outcome where utilities for health states (expressed on a 0-1 scale where 0 corresponds to death and 1 to perfect health) act as qualitative weights to combine the quantity and quality of life. Utilities can also be negative as some health states may be considered worse than death.
The primary means of eliciting utilities in AD is to administer a HRQoL questionnaire to patients or their caregivers (acting as patient proxies) and convert the resulting ratings into health utility scores using a scoring algorithm. At least two instruments, the quality of life-Alzheimer's disease (QoL-AD) [13
] and dementia quality of life (DEMQOL) [14
], have been developed to assess AD patients' HRQoL. Both instruments have been tested in caregiver proxies and baseline norms for a nondiseased population have been developed for the QoL-AD [15
]. However, neither instrument has a scoring algorithm that can be used to convert HRQoL ratings into health utility scores. For these reasons, generic utility instruments such as the EuroQoL (EQ-5D) [16
] and the Health Utility Index (HUI) [17
] have also been used to elicit HRQoL for AD patients. These instruments have scoring algorithms that allow HRQoL ratings to be transformed into health utility scores. These generic measures are often preferred to disease-specific measures in economic evaluations for policy making to facilitate comparisons across diseases by using benchmarks (e.g., $50.000 per QALY gained).
Several studies have shown that the EQ-5D [18
] and the HUI [21
] may be used to reliably elicit utilities in mild-to-moderate AD patients. However, a number of studies [18
] also reported that some mild-to-moderate AD patients would rate their health as perfect (i.e., utility score of 1) which could be due to patients' lack of insight about their impairment [22
]. In addition, many AD patients do not consider themselves sick or suffering from AD [21
]. To avoid the challenges of patient-measured HRQoL in AD, caregivers have been used as proxies to measure HRQoL for persons with AD. In such cases, caregivers complete instruments like the EQ-5D or the HUI on behalf of patients. Comparisons suggest caregivers' mean proxy health utility scores for patients are lower than patients' self-reported health utility scores on both the EQ-5D and HUI [19
]. These differences may be due to the influence of caregiver-specific issues such as burden, which may be especially important when caring for moderate AD patients. One study of the EQ-5D and HUI reported that caregivers' proxy HRQoL ratings were associated with their own levels of burden, rather than patients' self-reported HRQoL ratings or patients' scores on the Mini-Mental State Examination [29
] for cognitive impairment [24
]. Conversely, the authors of the same study wrote that caregivers' proxy ratings on the EQ-5D or HUI are at least as reliable as patients' own ratings on these scales.
The equivocal results in patients and caregivers suggest the need to search for other viable options that may be used to obtain estimates of AD patients' HRQoL and health utility. One option that has not been explored in the literature to date is to use the general public as a proxy to elicit HRQoL for AD patients. This alternative may be especially relevant for healthcare systems financed in whole or in part by public taxation (e.g., United Kingdom, Canada).
The first step in assessing the usefulness of the general public as a proxy is to determine whether the general public's HRQoL ratings would be sensitive to AD as a disease entity. That is, would the general public assign different HRQoL ratings to an AD disease state relative to a non-AD disease state? We had the opportunity to conduct a pilot study to address this question by using data collected for a project investigating Canadians' level of support for a tax increase to fund unrestricted access to AD medications [30
]. In the taxation study, we conducted a telephone survey of 500 randomly-selected members of the Canadian general public who did not have AD. Part of the survey involved the administration of the EQ-5D. Participants were asked to complete the EQ-5D for their current health state and again for a health state of moderate AD. Although it is possible to derive utilities from the general public through direct measurements using either a standard gamble (SG) approach (iteratively find the probability p in which the individual is indifferent between living with an hypothetical “moderate AD” forever or taking a drug which can cure or kill him with the probability p) or a time trade-off (TTO) approach (hypothetically live with “moderate AD” forever or live shorter in a better state of health); these methods are more resource intensive and complicated to implement than the administration of a preexisting questionnaire (e.g., EQ-5D). For these reasons, respondents were asked in this pilot study to answer the EQ-5D under a hypothetical “moderate AD” health state.
Findings suggesting that the general public can act as a proxy rater for AD patients' HRQoL have important implications for research. Recruitment of a random sample of the general public is easier than recruiting a sample of AD patients or caregivers. Regional or national rosters of patients or caregivers do not exist, so research involving these persons often faces the challenge of recruiting subjects from a patchwork of advocacy organizations, support groups, and medical practices. If the general public can be used in place of patients or caregivers, then there would be a clear efficiency gain for researchers who wish to determine HRQoL and health utilities for AD patients. Such a gain would be important for analysts conducting economic evaluations in response to rapidly changing policy environments.