The reference model models clinicaltrials.gov

The Reference Model for disease progression combines multiple existing risk equations to describe progression of diabetes. It uses publicly available data sources to collect models and validation data. There are two main data types of interest, baseline population data used for initialization of simulation, and outcomes data used for validation. Collecting these data elements from the literature was extremely labor intensive, error prone, and inefficient. It involved humans reading the paper and required reasoning and interpretation of data presented. The accumulation of population data was therefore slower than desired. However, a few years ago the NIH created a public database where such global information is stored and is accessible via the internet - ClinicalTrials.Gov. The Reference Model has been adapted to interface with this database. This paper presents these new capabilities and their benefits and provides new results indicating how well models describe newly imported populations.