Validating a Computable Phenotype for Nephrotic Syndrome in Children and Adults Using PCORnet Data

Abstract

Background: Primary nephrotic syndromes are rare diseases which impedes adequate sample size for observational patient-oriented research and clinical trial enrollment. A computable phenotype may be powerful in identifying patients with these diseases for research across multiple institutions. Methods: A comprehensive algorithm of inclusion and exclusion ICD-9 and ICD-10 codes to identify patients with primary nephrotic syndrome was developed. 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 (individuals not meeting case criteria seen within the same calendar year and within five years of age of a case) 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 nephrotic syndrome from 4,305,092 distinct patients in the CDM at all sites from 2009-2018. For all sites, the sensitivity, specificity, and area under the curve of the algorithm were 99% (95% CI: 97-99%), 79% (95% CI: 74-85%), and 0.9 (0.84-0.97), respectively. The most common causes of false positive classification were secondary FSGS (9/39) and lupus nephritis (9/39). Conclusion: This computable phenotype had good classification in identifying both children and adults with primary nephrotic syndrome utilizing only ICD-9 and ICD-10 codes, which are available across institutions in the United States. This may facilitate future screening and enrollment for research studies and enable comparative effectiveness research. Further refinements to the algorithm including use of laboratory data or addition of natural language processing may help better distinguish primary and secondary causes of nephrotic syndrome.

Standardizing electronic health record ventilation data in the pediatric long‐term mechanical ventilator‐dependent population

“…Multiple efforts to standardize data or advance the use of CDMs have been described in pediatric populations. Some examples include studies of children with multiple sclerosis, seizures, sickle cell disease, sleep‐disordered breathing, and rare kidney disease 16–19 . In some of these studies, the authors sought to ensure that systems could accurately identify children with certain conditions, 16,18 and in other studies, understanding treatment approaches for rare conditions was the goal 17 .…”

Influenza Vaccine Administration and Effectiveness Among Children and Adults With Glomerular Disease

“…Forward-backward mapping 18 using general equivalence mappings 19 were used to identify ICD-10-CM codes for GD from a published list of ICD-9-CM GD codes. 20 In a sensitivity analysis, inclusion or exclusion codes were adapted from a previously validated computable phenotype for GD 21 yielding a smaller GD cohort presumed to have a higher specificity for GD ( Supplementary Table S1 ). To restrict to only new-onset GD, patients were required to have 1 inpatient or 2 outpatient billing codes ≥30 and ≤365 days apart 20 following at least 1 year, and up to 5 years, of continuous enrollment in a fee-for-service insurance plan with prescription drug coverage with no GD billing codes present.…”

Optimizing the Electronic Health Record for Clinical Research: Has the Time Come?

“…Oliverio et al developed a "computable phenotype" to identify patients of interest from within the EHR, which aims to improve the sensitivity and specificity of these searches (2). A computable phenotype is a definition of a condition that is solely on the basis of routinely collected and stored data elements from the EHR.…”

“…Patients and nonpatients were reviewed individually by a nephrologist to confirm the diagnosis of primary nephrotic syndrome the old-fashioned way-through individual chart review. Sensitivity was a remarkable 99%; however, specificity was only 79%, with false positives primarily due to patients with secondary FSGS and membranous lupus nephritis (2). Interestingly, specificity varied by institution, suggesting there may be some variability among coding practices in various centers that may limit the application of this computable phenotype more broadly.…”

“…In this edition of Kidney360 , Oliverio et al . (2) provide a much-needed framework and proof of concept for optimally using EHR data to identify patients for nephrology research.…”