Abstract: PO1872
Validating a Computable Phenotype for Nephrotic Syndrome in Children and Adults Using PCORnet Data
Session Information
- Glomerular Diseases: Clinical, Outcomes, and Trials - 2
October 22, 2020 | Location: On-Demand
Abstract Time: 10:00 AM - 12:00 PM
Category: Glomerular Diseases
- 1203 Glomerular Diseases: Clinical, Outcomes, and Trials
Authors
- Oliverio, Andrea L., University of Michigan Medical School, Ann Arbor, Michigan, United States
- Marchel, Dorota, University of Michigan Medical School, Ann Arbor, Michigan, United States
- Tran, Cheryl L., Mayo Clinic Minnesota, Rochester, Minnesota, United States
- Ayoub, Isabelle, The Ohio State University Wexner Medical Center, Columbus, Ohio, United States
- Almaani, Salem, The Ohio State University Wexner Medical Center, Columbus, Ohio, United States
- Greco, Jessica M., The Ohio State University Wexner Medical Center, Columbus, Ohio, United States
- Denburg, Michelle, The Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
- Gipson, Debbie S., University of Michigan Medical School, Ann Arbor, Michigan, United States
- Mariani, Laura H., University of Michigan Medical School, Ann Arbor, Michigan, United States
Background
Primary nephrotic syndromes (pNS) are rare diseases which can be a barrier to adequate sample size for observational patient-oriented research. A computable phenotype may be powerful in identifying patients with these diseases for research while leveraging data from millions of patients in the PCORNet Common Data Model (CDM).
Methods
A comprehensive algorithm of ICD-9 and ICD-10 codes indicative of pNS was developed based on prior work in the University of Michigan Health System. Cases of pNS were defined as subjects that were seen for at least one encounter with more than 1 NS code, and did not have codes for diabetes mellitus, systemic lupus erythematosus, or amyloidosis. Non-cases were individuals not meeting case criteria who were seen in the same calendar year and within 2 years of age of a case. The algorithm was executed against the PCORNet CDM at 3 institutions from Jan 1, 2009 to Jan 1, 2018, where a random selection of 50 cases and 50 non-cases were reviewed by a nephrologist, for a total of 150 cases and 150 non-cases reviewed. The classification accuracy (sensitivity, specificity, positive and negative predictive value, F1 score) of the computable phenotype was determined.
Results
The algorithm identified a total of 2,708 patients with NS from 4,305,092 distinct patients in the CDM at all sites. For all sites, the sensitivity, specificity, PPV, and NPV of the algorithm were 99.1%, 81.0%, 76.7%, and 99.3%, respectively. The accuracy of the algorithm was 88.0% with an F1 score of 86.5%. The most common cause of false positive classification was secondary FSGS (17/35), followed by class V lupus nephritis (9/35).
Conclusion
While prior computable phenotypes for glomerular diseases have used IMO and SNOMed codes, this computable phenotype had good classification in identifying both children and adults with pNS utilizing only ICD-9 and ICD-10 codes, which are universally available. This may facilitate future screening and enrollment for research, however further refinements to the algorithm or addition of natural language processing may help better distinguish primary and secondary FSGS.
2 x 2 Computable Phenotype Classification Table
True Status | |||
Primary NS | Not primary NS | ||
Predicted Status | Primary NS | 115 | 35 |
Not primary NS | 1 | 149 |
Data from all 3 health systems
Funding
- NIDDK Support