Annual Costs of Illness versus Lifetime Costs of Illness and Implications of Structural Change

Abstract

Annual, or prevalence, costs of illness typically measure the costs during a year (the base period) of manifestations of disease that may have had its onset during or any time prior to the base year. Lifetime, or incidence, costs of illness include costs incurred from onset until cure or death, which can extend over many years. While annual costs estimate the burden of illness during a year, lifetime costs more appropriately measure the benefits of changes in incidence of disease, as might result from prevention or successful intervention during the course of disease. There are two important conceptual dktinctions between annual and lifetime costs: (I) the liming of costs, or the time period(s) during which costs are counted, and (2) the influence of time on the underlying structure that generates costs of illness. From these conceptual differences follow important empirical differences. Annual costs require observing costs during the recent past, irrespective, for example, of the present stage of disease in an individual and its influence on costs. Lifetime costs, on the other hand, require knowledge of the natural history of disease and the concomitant use of medical care and projections of future costs over the course of the disease. Although conceptually quite distinct, lifetime costs are frequently estimated from current cross-section profiles of costs for patients at various stages in the disease process and may be derived directly from annual costs using fewer data and with little loss in accuracy. Since estimates of lifetime costs typically involve projecting future costs, possibly decades away, it is important to consider the impact on future costs of changes in parameters that determine costs. One of the most important of these, and usually neglected, is technological change, which has had a pervasive and significant influence on medical care costs. In times of rapid change in medical care technology and its diffusion, estimates of lifetime costs may be seriously biased if potential structural change is not taken into account.