Fig. 1From: Rule-based and machine learning algorithms identify patients with systemic sclerosis accurately in the electronic health recordDevelopment of algorithms to identify patients with systemic sclerosis (SSc) in the electronic health record (EHR). At least a 1-time count of the SSc ICD-9 code (710.1) or ICD-10-CM codes (M34*) was applied to the 3 million subjects in Vanderbilt’s Synthetic Derivative, which resulted in 1899 potential SSc cases. Of these 1899 potential SSc cases, 200 were randomly selected for a training set to develop and test algorithms with various combinations of the SSc ICD-9 and ICD-10-CM codes, keyword search for Raynaud’s phenomenon, and positive ANA (≥ 1:80). The highest performing algorithm was internally validated in a set of 100 subjects who were not part of the original training setBack to article page