Cyclic fluctuations in hospital bed occupancy in Roma (Italy): supply or demand driven?

The objective of this study was to assess hospital bed occupancy both by planned and unplanned cases, and to assess how supply and demand affect bed occupancy. Data was obtained from the Lazio Hospital Information System (HIS) dataset on all hospital discharges from July 1998 to June 2001. Using Diagnosis Related Groups (DRG) as the reason for hospital stay, admissions were classified into four categories: 'planned stay', 'presumed planned stay', 'presumed unplanned stay', and 'unplanned stay'. Time series analysis of daily bed occupancy by category of stay was performed. Generalized Additive Models (GAMs) were used to asses the effect of weekdays and holidays on bed occupancy. Fluctuations in daily occupancy were observed in all categories of stay-in general, bed occupancy decreased over weekends, on national holidays, and during the major holiday season of August. In comparison with unplanned stays, the largest fluctuations were observed for planned stays while presumed planned and unplanned stays showed lesser fluctuations. It is possible to distinguish planned and unplanned hospital stays by using DRG grouping. Cyclic rigidities in the supply of services rather than the availability of beds or demand for beds seem to dictate hospital use in Roma so that restrictions in services hamper any reallocation of beds for 'planned stay' when demand for 'unplanned stay' beds declines.