Using EMR Data Collections to Outline SCID Clinical Phenotypes
Monday, March 7, 2016: 2:00 PM
Room 408A (Convention Center)
Shradha Agarwal, MD, FAAAAI, ,
Rationale:  SCID is a genetic disorder caused by mutations in cellular and humoral immunity leading to heterogeneous presentations.  Early identification and transplantation reduces morbidity and mortality as demonstrated by improvement in immune function.  We compared three patient registries to dissect clinical phenotypes of SCID.

Methods:  Using ICD code 279.2, three registries were queried for patients with SCID.  USIDNET (1997-2014) included specialists entered data; SPARCS (2005-2010) collected hospital data from NYS; HCUP (2009) collected data from 103 US hospitals.  Demographics and clinical data for subjects under the age of 20 were extracted utilizing databases’ reporting tools.

Results:  In all data sets there was predominance of males with SCID (63%, 51%, 60%) and Caucasian subjects (53%, 54%, 41%). Pneumonia was the most common diagnosis associated with SCID (17.8%, 18.9%, 11.1%).  Other diagnoses across registries:  skin infections (10.5%, 10.9%, 0.73%), sepsis (4.0%, 5.1%, 6.8%), fever (7.3%, 2.3%), viral (4.5%, 2.9%, 5.8%), sinusitis (7.4%, 1.7%, 1.7%), otitis (9.1%, 1.1%, 0.87%).  While USIDNET collected selected fields, SPARCS and HCUP collate all ICD codes, including numerous gastrointestinal, neurologic, allergic, nutritional, failure-to-thrive, and endocrine codes.  Within HCUP, California had the most SCID related discharges followed by Ohio, Texas, New York.  47% of admissions occurred by age 2 (0yr-25.6%; 1yr-12.1%; 2yr-9.2%).

Conclusions:  Data sets of EMRs are powerful resources allowing the capture of global clinical data about rare diseases. Contrasting the collection of selected data fields (USIDNET) with other global medical collections (SPARCS, HCUP) produces a more rounded clinical phenotype of rare immune defects, across medical boundaries.