Objective: To calculate per-case hospital costs for patients with cystic fibrosis under routine conditions from a healthcare provider's perspective; identify the impact of different cost categories; investigate whether cases with cystic fibrosis can be grouped into homogenous cost groups according to defined severity levels; and determine the value of specific factors as predictors of hospital cost variations.
Methods: All data were collected from cases (n = 131) admitted to an inpatient cystic fibrosis unit under routine conditions during a period of 6 months in 2004. All costs were calculated for the year 2004 and divided into categories with high and low impact on variation in hospitalisation costs between patients. Staff costs for patient care, laboratory costs and drug costs were defined as categories with high impact, thus the individual resource utilisation for each case was measured. Cost categories that were classified as having a low impact were measured as overhead costs. Cases were classified according to two different severity models; within each model, patients were classified according to three severity levels. The diagnosis-related model classifies patients with pulmonary hypertension and global respiratory insufficiency as having severe disease, patients with Pseudomonas aeruginosa as having moderate disease, and patients with no colonisation of the lungs as having mild disease. The lung-function-related model differentiates patients as having mild, moderate and severe disease when patients have forced expiratory volumes in 1 second (FEV(1)) that are > or =70%, between > or =40% and <70%, and <40%, respectively. Analysis of variance tests were performed to investigate the differences of mean costs between the groups. Ordinary least squares regression analysis was used to determine predictors for cost variation.
Results: The mean total costs per case were 7326 euro. Almost one-third of the total mean costs were attributable to drug costs (28% of total costs), while shares of staff costs for patient care and laboratory costs (both 9% of total costs) were relatively small. Most of the difference in costs between severity levels was attributable to the variation in overhead costs and drug costs. For both severity models differences in mean total costs of mild and severe cases were statistically significant (p < 0.01 and p < 0.05, respectively) when compared with the mean costs of non-mild and non-severe cases. However, in moderate cases, significant differences compared with cases that were not of moderate severity were only seen for certain cost categories. In the multiple regression model the variables 'diagnosis-related severity' and 'FEV(1)' explained 31% of the variance of 'Ln (total costs per case)' between severity levels (p < or = 0.01).
Conclusion: This study shows that to a large extent hospitalisation costs for patients with cystic fibrosis vary according to the severity of their disease; drug costs play a major role in these differences. In the light of this variation it seems plausible to create separate reimbursement rates for two or three severity groups. Diagnoses as well as FEV(1) seem suitable criteria for such a classification.