(P-629) How Under-coded Is Pediatric Hypertension?: Comparison of Claims and Electronic Health Record (EHR) Data

Background: Under capture of vital signs measurements in claims data may require supplementary EHR data to properly classify study variables.

Objectives: To determine concurrence between claims and clinical identifiers of pediatric hypertension (HTN) in Data Partners with both claims and EHR data in the Food and Drug Administration‘s Sentinel System.

Methods: Using claims and EHR data from 7 Data Partners (1/2006-12/2018), we identified and classified patients aged 3-17 years according to the 2017 American Academy of Pediatrics (AAP) guidelines for HTN. We required same day diastolic/systolic blood pressure (BP) records from the Sentinel Common Data Model Vital Measures Table, and a valid height within 6 months of BP reading. Separate cohorts looked at patients with 1) any valid BP measure; 2) any high BP measure; and 3) a high BP measure after which subsequent BP measures never return to normal. High BP was defined as either elevated BP (EBP) or hypertensive BP. We assessed concurrence between high BP EHR measures and claims records in the year after the EHR measure.

Results: We identified 1,658,773 patients with any valid BP measure in their EHR record and a total of 10,021,452 valid BP measures across all patients. Of these total measures, 14.0%, 14.4% and 71.6% indicated EBP (not qualifying as hypertensive), hypertensive, and normotensive measures, respectively. Among patients with any BP measures, 1,024,299 (61.8%) had at least one EBP or hypertensive measure, and 364,093 (21.9%) had consistent EBP or hypertensive measures. AAP guidelines require three hypertensive measures to meet the clinical definition of HTN. Among patients with an initial hypertensive BP measure (N=510,407), 90.5% of patients had at least two additional measures in their subsequent enrollment history. Of those, 48,739 patients (9.5%) had at least two subsequent EBP or hypertensive measures. Correlation between claims and EHR definitions of HTN was poor; only 1.7% of hypertensive patients and 0.5% of EBP patients had any high BP or HTN-specific ICD code in the following year. Among those with consistently high BP, 3.6% of hypertensive patients and 0.9% of EBP patients had a corresponding claims code.

Conclusions: More than 94% of patients with HTN BP in EHR do not have a claim for HTN, suggesting that studies relying exclusively on claims may under capture patients’ HTN status. Incorporation of EHR data could improve specificity. The percentage of hypertension measures among all patients (14.4%) is higher than national survey estimates but may be a product of using one rather than three BP measures. Future EHR analyses should explore rates of pediatric HTN when strictly applying AAP guidelines.