Fig 1.
Overview of multi-stage process for SSc ICD-10 code validation and algorithm testing and validation. Patients were selected from healthsystem databases if the ICD-10 code was present at least twice in encounters, billing codes or the problem list. Of the 2138 potential patients 1183 were excluded and a random selection of 100 patients underwent code validation. The 955 patient cohort was divided in half for testing of two algorithms using disease manifestation keywords and internal validation of highest performing algorithm. When applied to the entire cohort, 788 patients were identified as likely cases. ICD-10 = International Classification of Diseases, Tenth Revision.
Table 1.
Characteristics of entire cohort by randomization for baseline code validation subset n = 955.
Table 2.
Characteristics of patients based on case classification N = 100 (baseline validation cohort).
Table 3.
Performance characteristics of ICD-10 code and algorithms.
Table 4.
Select comparisons of patients identified by Algorithm 2 compared to patients with ICD-10 code use at least two times.