Real world data in health technology assessments in kidney transplants in Germany: use of routinely collected data to address epidemiologic questions in kidney transplants in the AMNOG process in Germany.

Introduction: It is discussed whether real world data can be used in health technology assessment. Following it is of interest whether routinely collected data for quality assurance (QA) in the hospital sector is feasible to address epidemiologic questions in kidney transplantation in the AMNOG process in Germany. Objectives: To investigate the proportion of kidney transplants classified as from so-called standard criteria donors (SCD) and from expanded criteria donors (ECD) in Germany and to study the age distribution. Methods: After granted use by the Federal Joint Committee (G-BA), the data analysis was carried out by the AQUA institute, and a SPSS code was developed. Special challenge was the complex definition of SCD/ECD criteria that, in addition to donor age, takes into account combinations of donor diagnoses, creatinine, and cold ischemia time. Results: Age analyses could be performed in all patients. Median age of the adult transplant recipients in Germany was 54 years in 2012 as well as in 2013, range 18-85 and 18-82 years and a mean (SD) of 53 (14) and 52 (14) years, respectively. 63.5% (2012) and 62.5% (2013) of recipients were male. Classification in SCD/ECD transplants could be performed for 2,083 of 2,461 patients (85%; 2012) and for 1,795 of 2,079 patients (86%; 2013). Of all classifiable transplants 61.4% (2012) and 66.5% (2013) were SCD transplants. Total project time from the request to results was <6 months. Conclusions: The use of data routinely collected for QA in the hospital sector is feasible to address epidemiologic questions in kidney transplantation in the AMNOG process in Germany, which is basically following the systematic of an HTA process. All patients with kidney transplants are represented thus avoiding sampling error. Limitations include the availability of all necessary data in the QA data set. Within <6 months' time with reasonable resources it was possible to meet timelines. The analyses were accepted by the authorities.