This paper reports the economic consequences to both private and public funds of a best-bet intervention aimed at patients with mild AD and their primary caregivers. The intention of this work was to examine the extent to which long-standing intervention involving not only patients but also caregivers requires direct and indirect copayment, which ultimately may affect equity in access to health care. The main finding was that intervention in mild AD may come at a relatively low cost to public funds but, at the same time, it requires significant copayment from private parties, which can be summarised in a cost-sharing ratio between private and public pockets of almost 1:1.
The concurrent and detailed measurement of patient and caregiver time required for the successful intervention in mild AD could have been performed in other intervention studies, yet this seems to have been neglected in studies reported in the literature. This is in contrast to a strong economic theoretical recommendation to include patient and caregiver time as essential resource utilisations: costing from a societal perspective should include all resource utilisation irrespective of who invests it. The absence of such practice could be due to a lack of consensus regarding a valid and reliable framework for costing informal care [23
] or to the more pragmatic issue of whether the required effort of measurement is worthwhile relative to the benefit.
From a theoretical point of view, a patient is 'occupied' when he or she is receiving treatment and thus benefits are sacrificed elsewhere i.e. the person cannot gain utility from leisure time or financial earnings (if in paid employment). The minimum value of benefits forgone is, using the opportunity cost approach, approximated by the individual's net (if leisure time is used) or gross wage (if time off from work is used) [26
]. However, as most AD patients are not in paid work, a hypothetical wage rate (we used the age- and gender-matched national average) may not reflect the true value of patients' time. We conducted sensitivity analysis using a lower rate, in order to be flexible about using the average wage rate to value patient time. However, only a moderate impact on the overall cost-sharing ratio was demonstrated.
The term 'concordance' was used instead of the more conventional term 'compliance'. This was in consensus with recent criticism that 'compliance' denotes the extent to which the patient obeys the clinician rather than takes an active role in the doctor-patient relationship [27
]. In this context, concordance refers to the extent to which the collaboration between counsellor and participant results in participation in individual components of the intervention. Relating to external validity, the high rates of participation achieved in the current trial were higher than expected and may be explained by the fact that participants seemed to be socioeconomically advantaged relative to the average citizen and, not least, that the trial was undertaken by highly dedicated professional staff. Lower rates of participation may be expected in routine practice, although that would not be expected to significantly influence the ratio of private to public costs.
There are some uncertain variables in relation to local infrastructure - both of the health care sector (e.g. whether relevant professionals are available at the nearest institution) and the private sector (e.g. what is the average distance from participants' homes to the institution). We observed significant variation across the five county districts of the trial, partly due to the free choice of participating centres to conduct counselling in the patients' homes or at health care institutions. Similarly, some centres were able to run highly efficient courses with all places filled, while others suffered low volume due to location in rural areas. Such variation affects not only total intervention costs but also shifts costs back and forth between private and public pockets.
Several issues are of note for community health planners who might consider adaptation of the intervention described here. First, the study population represents a relatively resourceful group as they are at an early disease stage and are socioeconomically advantaged. While the study is one of the few to describe an intervention targeted to this population, the results are not likely to be generalisable to more severe stages of AD. Second, favourable resource utilisation and costs are irrelevant if there is no evidence for a clinical effect and the most useful approach would be a synthesis of information as a cost-effectiveness evaluation. These analyses are currently in progress. A third remark relates to budget impact, which should be estimated by multiplying not only the n of a target population with the intervention cost, but should also take into account that a certain proportion of eligible dyads reject the invitation to participate and a further proportion drop out after initially agreeing to participate. We have no estimates on the former proportion, whereas the latter appeared to be 0.12. Consideration should also be given to recruitment of participants, that is, whether existing regimens can accommodate recruitment or whether extra resources are required to establish a regimen for screening and inviting candidate participants. Such costs are not included in the present analysis. Finally, start-up costs of e.g. recruiting and educating staff, developing informational packages etc. should be expected (these were not included in the current study).
It has long been argued that disparities in the use and quality of care are particularly evident in long-term care, but the first attempt to summarise this evidence was reported only recently [29
]. Much of the evidence arises from the USA and focuses on racial disparities, where Blacks generally appear to have a lower health care use than Hispanics. One study, however, shows that individuals with low socioeconomic status generally underuse health care; explanatory variables appear to include organisational and geographic characteristics of care provision [30
]. This could be synonymous with a hypothesis that requirements of considerable participant time input and high transportation costs pose a restriction to health care access. However, the direct question of whether a 1:1 cost-sharing ratio between private and public finances restricts access to care for some groups relative to others remains unanswered.