OBJECTIVES The lack of accessible, comprehensive sources of medical and quality of life data in Germany has partially hindered the ability to research diabetes clinical practice. The aim of this study was to build a prospective, national, multi-centre Type 2 diabetes mellitus (T2DM) registry using an innovative data collection methodology to better understand the disease specific epidemiology, treatment patterns and patient reported outcomes (PRO). METHODS From the physician universe of the IMS^®Disease Analyzer (1,500 diabetologists and 57,500 general practitioners), a sub-set of 101 physicians have agreed to participate. Pre-programmed inclusion criteria (i.e. ≥18 years old with T2DM diagnosis) triggers consecutive patient selection for further data collection. A custom software-based electronic case report form (eCRF) is completed once per quarter by the physician to capture further diabetes information. Patients complete a validated set of PROs at physician office every 6 months. Anonymised data from the electronic medical record (EMR), eCRF and PRO are linked through a unique process to ensure patient confidentiality.  RESULTS A total of 1,029 T2DM patients are contributing to the registry (at June 2014). Preliminary analysis of the DIAREG cohort has shown comparable representativeness to IMS^®Disease Analyzer for several patient characteristics: age, gender, distribution of T2DM treatment and physician type. The DIAREG improves research use of the EMR data by filling in missing information for weight, blood pressure and HbA1C (e.g. completeness of HbA1C records improved from 42.3% to 83.3% of patients). The DIAREG enhances breadth of information by collecting data that is not normally captured in EMR sources, such as physician rationale related to treatment choice or switch, target HbA1C, hypoglycaemic events, occurrence of micro or macrovascular events, and patient reported information (e.g. quality of life, treatment satisfaction and depression). CONCLUSIONS DIAREG enables unique capability to research T2DM in Germany through multiple linked longitudinal data sources.