Evaluating phecodes, clinical classification software, and ICD-9-CM codes for phenome-wide association studies in the electronic health record

Abstract

ObjectiveTo compare three groupings of Electronic Health Record (EHR) billing codes for their ability to represent clinically meaningful phenotypes and to replicate known genetic associations. The three tested coding systems were the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes, the Agency for Healthcare Research and Quality Clinical Classification Software for ICD-9-CM (CCS), and manually curated “phecodes” designed to facilitate phenome-wide association studies (PheWAS) in EHRs.Methods and materialsWe selected 100 disease phenotypes and compared the ability of each coding system to accurately represent them without performing additional groupings. The 100 phenotypes included 25 randomly-chosen clinical phenotypes pursued in prior genome-wide association studies (GWAS) and another 75 common disease phenotypes mentioned across free-text problem lists from 189,289 individuals. We then evaluated the performance of each coding system to replicate known associations for 440 SNP-phenotype pairs.ResultsOut of the 100 tested clinical phenotypes, phecodes exactly matched 83, compared to 53 for ICD-9-CM and 32 for CCS. ICD-9-CM codes were typically too detailed (requiring custom groupings) while CCS codes were often not granular enough. Among 440 tested known SNP-phenotype associations, use of phecodes replicated 153 SNP-phenotype pairs compared to 143 for ICD-9-CM and 139 for CCS. Phecodes also generally produced stronger odds ratios and lower p-values for known associations than ICD-9-CM and CCS. Finally, evaluation of several SNPs via PheWAS identified novel potential signals, some seen in only using the phecode approach. Among them, rs7318369 in PEPD was associated with gastrointestinal hemorrhage.ConclusionOur results suggest that the phecode groupings better align with clinical diseases mentioned in clinical practice or for genomic studies. ICD-9-CM, CCS, and phecode groupings all worked for PheWAS-type studies, though the phecode groupings produced superior results.

Genome‐wide and Phenome‐wide Approaches to Understand Variable Drug Actions in Electronic Health Records

“…ICD codes are published and maintained by the World Health Organization for classification of diseases and services for reimbursement of medical services. While the phecode groupings have been shown to better align with clinical diseases in practice, other methods of phenotype classification also are effective for PheWAS studies . Hebbring et al .…”

“…While curation of a single (or few) phenotypes for a GWAS is manageable, this is much more difficult for the thousands of potential phenotypes used in a PheWAS. Phecodes have shown efficacy for PheWAS analyses; however, they do not align precisely with clinical diseases and may not have adequate granularity or specificity for some phenotypes . Phecodes currently use ICD‐9‐CM codes as their sole source of information.…”

“…While the phecode groupings have been shown to better align with clinical diseases in practice, other methods of phenotype classification also are effective for PheWAS studies. 14 Hebbring et al reported a PheWAS method using individual ICD-9 codes and parent ICD-9 three-digit groups as phenotypes. 15 They not only replicated a known association of an HLA www.cts-journal.com class II allele, HLA-DRB1, with multiple sclerosis, but also replicated associations of HLA-DRB1 with erythematous conditions and benign neoplasms of the respiratory and intrathoracic organs, found to be significant in a prior study by Denny and colleagues.…”

Processing of Electronic Medical Records for Health Services Research in an Academic Medical Center: Methods and Validation (Preprint)

“…ICD codes have good utility for clinical research where the researcher needs the granularity for identification and attribution of pathology at an individual level [22,52]. However, for health services research, broader classification and coding methods such as the Clinical Classification Software (CCS) [53], have demonstrated utility as there is sufficient granularity at a population level, yet reduces sparsity [54][55][56].…”

“…While the standardization of diagnosis codes of different nosologies or different versions of the same nosology has been reported previously [22,23], the completeness and validity of such mapping is rarely reported. This lack of transparent sharing of code set definitions, construction process and validity is a barrier to rapid scaling of health services research [24] given its importance and widespread relevance.…”