| Eur J Health Econom 4 · 2003 304 Countries all over the world are facing important issues of health care organiza- tion, cost, financing, and quality, one be- ing the appropriate level of per capita fi- nancing for defined populations. While age and gender are important predictors of future health care needs, these predic- tions can be improved by considering the previous health problems of the popula- tion so that the various insurance plans do not have incentives for selective enroll- ment of healthy individuals and avoiding those with health problems [1]. For in- stance,in the United States the Center for Medicare and Medicaid Services is using principal inpatient (PIP) diagnostic cost groups (PIP-DCGs) [2], a measure that is defined by clinical conditions in the pre- vious year, to account for patient risk in the Medicare health maintenance organi- zation population. While the health districts in Italy are geographically based and not subject to selective enrollment, it is nevertheless im- portant that the criteria for financing the health districts adequately account for population differences so that care pro- grams may be equitably planned and im- plemented to meet specific population needs. The Italian National Health Service guarantees universal access to health care to 57 million Italians. Each of the 20 Ital- ian regions has the responsibility for pro- viding care to residents of the region. Funding flows from the central govern- ment to the regions with an increasing emphasis on regional taxation. The re- gions in turn have delegated major re- Original Papers Elaine J.Yuen 1 · Daniel Z. Louis 1 · Paolo Di Loreto 2 · Joseph S. Gonnella 1 1 Thomas Jefferson University, Philadelphia, Pa., USA · 2 Regione Umbria, Italy Modeling risk-adjusted capitation rates for Umbria, Italy sponsibility to geographically defined Lo- cal Health Units. As a part of the decen- tralization of responsibility for providing health care, the Local Health Units have been further divided into geographically defined health districts, typically with a population of 50,000–100,000. A major challenge with the creation of health dis- tricts to assure that appropriate financing is provided to meet the needs of the pop- ulation [3]. Currently the allocation of re- sources from the regions to the Local Health Units is based on a formula which takes into account age, gender, and mor- tality data. Methods and materials Study objective The purpose of this project with the re- gion of Umbria was to study the effect of refining per capita financing by risk ad- justment in an Italian population.Adjust- ment on the basis of age and gender as well as severity of illness was studied. Dis- ease Staging risk adjustment measures were used. We had two major study ob- jectives: (a) to collect and compile data from an entire population in the region of Umbria and (b) to estimate and test risk adjustment factors using models based upon Disease Staging in this population. Data sources Data consisted of 2 years of hospital data (1997–1998), 2 years of outpatient phar- maceutical data (1997–1998), and 1 year of demographic data (1998) from the region of Umbria. Hospital data were from the Italian hospital discharge abstract (Sche- da di Dimissione Ospedaliera, SDO) and included inpatient admissions as well as day hospital episodes. Pharmacy data were from the paid pharmacy claims and cap- tured information at the individual’s pre- scription level.As data were provided by the regional health services agency in Um- bria which was responsible for payment of health services for the entire Umbrian population, we obtained information data on Umbrian residents who sought care in other regions as well as Umbria residents hospitalized in Umbria. Data on hospital SDO files included di- agnosis-related groups (DRGs) and DRG- based tariffs, ICD-9 CM codes (which had been converted in Italy from the Interna- tional Classification of Diseases, 9th edn.), and demographic information. Hospital payments included both facility and phy- sician reimbursements. The pharmacy files included individual prescriptions, drug codes, tariffs, and copayments. For both hospital and pharmacy records the costs were estimated by tariffs, the amount reimbursed by the region for health care services, as there is no cost accounting system in place in Italy at the present time. Tariffs were reported in Italian lira in the data, but for purposes of reporting these have been converted to euros. Demographic data were from region- al files kept for all Umbrian residents and included unique identifiers such as health Eur J Health Econom 2003 · 4:304–312 DOI 10.1007/s10198-003-0193-2 Online publiziert: 15. August 2003 © Springer-Verlag 2003