Costs in rheumatology: results and lessons learned from the 'Hannover Costing Study'.

The objective of this study is to review the concept of the 'Hannover Costing Study' and to present and discuss the major insights generated during the course of the project. The costing study was performed in conjunction with a randomized controlled prospective trial assessing the effectiveness of a disease management module in rheumatoid arthritis (RA). A full set of clinical and cost data both from patient-reported and payer-derived cost data was developed. In particular the study included (1) the development of a matrix of cost domains which might be used as a common taxonomy in costing studies, (2) the descriptive analysis of payer derived cost data, (3) the analysis of cost data in patients with uncertain diagnosis; (4) the development and validation of a patient-reported costing instrument, and (5) an assessment of productivity costs. The following are the results (1) the developed matrix of cost domains included 16 separate cost domains: 7 outpatient, 3 inpatient, 4 other disease related, and 2 productivity domains; (2) the micro-costing analysis showed total direct costs of <euro> 3,815 per patient-year (standard error of mean, SEM: <euro>267) and RA-related direct costs were <euro>2,312 per patient-year; (3) in patients with uncertain diagnosis of RA and no treatment with 'Disease Modifying Antirheumatic Drugs' (DMARD) costs were significantly lower; (4) the comparison of patient-reported with payer-reported cost data generally supports the use of highly aggregated items to assess health care utilization in RA; (5) productivity costs in patients that are gainfully employed and in patients who receive RA-related retirement payments exceed RA-related direct costs. Furthermore, RA-patients reported their productivity losses adequately. The study added some additional insights to the following questions: What costs should be collected, what level of detail is required for that task, what patients should by analyzed, and what data sources should be used in further studies in RA.