Disease-staging for modelling current and future health-care impact of disease: illustrations for diabetes mellitus and AIDS.

This paper addresses the economic relevance of appropriately staging progression of disease for cost assessment and projection. Illustrations are drawn from two Dutch cost-of-illness studies for Diabetes Mellitus (DM) and AIDS. Our disease-staging specifies a separate late stage to capture terminal high-intensity care for end-stage complications. Data are used from a registration of national hospital resource utilization for DM and from a detailed database on AIDS hospital care and costs. In particular, differences in average length of hospital stay for several potential DM end-stage complications are compared with non-DM patients. Neglecting the separation of a specific late stage is estimated to possibly miscalculate 15-20% of the hospital bed needs for DM and AIDS in the Netherlands. In particular, confidence intervals of future projections for AIDS hospital beds--using respectively staged and non-staged models--do not overlap. AIDS hospital costs are overprojected by almost 11% in the non-staged model. Our estimation of DM hospital bed needs raises the percentage in total Dutch hospital beds from 1.7% (neglecting a specific late stage) to 2.0%. For DM and AIDS an appropriate disease-staging averts structural biases in estimations of bed needs and costs, and therefore benefits the planning of hospital care facilities. Obviously, gathering comparable information and developing similar methodology on other diseases, such as respiratory disease, cardiovascular disease and cancer, is needed and could benefit planning in these fields.