(184) The Challenges with Identifying Pediatric Recurrent Wheezing in Real World Data

Publication Number: 1229

Background: Recurrent wheezing (RW) is an important source of pediatric burden and has been identified as a potential risk factor for asthma. However, it is a challenging outcome to evaluate in pharmacoepidemiology studies because it is inconsistently defined clinically and there are no standardized diagnosis codes for identifying children with RW.

Objectives: To characterize the identification of pediatric RW in healthcare databases using International Classification of Disease (ICD) diagnosis codes.

Methods: We conducted a systematic literature review within Pubmed and Embase to identify studies examining pediatric RW within healthcare databases published between January 1, 2000 and August 3, 2021. ICD codes used to identify RW were abstracted from the publications. The search was restricted to manuscripts where RW was the outcome as identified by ICD-9 or ICD-10 codes in a pediatric (≤ 18 years old) study population.

Results: We identified 18 studies that utilized ICD coding algorithms for identifying pediatric RW. 14 studies used electronic health record data either in isolation or in tandem with other data sources, whereas only 8 studies used claims data in some capacity. 12 of the studies used ICD-9 codes, 9 studies used ICD-10 and 3 studies used both. There was a trichotomy of RW coding patterns observed in these studies. The first group of studies (n= 4) identified RW with the codes of 786.07 “Wheezing” or R06.2 “Wheezing”. The second group (n=7) defined RW as a subset of asthma or synonymous with asthma. This group used either a combination of wheezing (786.07 or R06.2) and asthma codes (493.X “Asthma”, J45 “Asthma”) or asthma codes in isolation. The third group of studies (n=7) used a broader set of ICD codes related to a variety of respiratory outcomes in addition to wheezing and asthma codes. None of the identified studies validated pediatric RW through a medical record review. Among studies we excluded, RW was commonly identified through unstructured data including patient notes and parent reports

Conclusions: Diagnosis codes can streamline outcome identification in real world data. This literature review sheds light on some of the difficulties of defining and identifying RW in children. There was no clear consensus on how authors defined RW as evident from the various reported algorithms. Authors’ choices of codes are aligned with their definitions of RW. Some authors treated RW as an extension of pediatric asthma, and either identified RW exclusively with asthma codes or a combination of asthma and wheezing codes. It may be valuable for future studies to explore natural language processing as an alternative or complementary option to diagnosis codes for identifying pediatric RW.