VALIDATION OF THE HEPARIN-INDUCED THROMBOCYTOPENIA DIAGNOSIS CODE USING A LARGE-SCALE ELECTRONIC HEALTH RECORD DATABASE
Author(s)
Amanda M. Moore, PharmD, PhD;
TriNetX, Cambridge, MA, USA
TriNetX, Cambridge, MA, USA
OBJECTIVES: Heparin-induced thrombocytopenia (HIT) is a life-threatening condition suspected in patients with new onset thrombocytopenia after exposure to heparin; however, laboratory testing to confirm HIT can be time-consuming and patients may carry a suspected diagnosis in their medical record for an extended period. The purpose of this study was to evaluate the performance of the ICD-10-CM diagnosis code for HIT against lab-validated results.
METHODS: This was a retrospective, observational cohort study of adult patients ≥18 years old in the TriNetX Dataworks-USA research network of de-identified electronic health record (EHR) data with an incident HIT diagnosis (ICD-10-CM code D75.82) (index date) from 01/01/2018 to 12/31/2024. Patients were required to have ≥1 year of baseline history and were followed for up to 42 days after index date. Lab-confirmed HIT was determined based on an ELISA optical density ≥2.00 and/or a positive serotonin release assay or heparin-induced platelet aggregation assay. The sensitivity, specificity, and positive predictive value (PPV) of the HIT diagnosis code was estimated.
RESULTS: A total of 8,537 patients had ≥1 diagnosis code for HIT. Performance of the code included an estimated 36.5% sensitivity, 99.99% specificity, and 21.1% PPV. As further restrictions were applied, the PPV increased to 30.9% (≥1 inpatient code) and 34.4% (≥2 codes 2-14 days apart). Across all groups, demographics and comorbidities/risk factors were similar between patients who were true and false positives. Patients who were true positives were more likely to have recent unfractionated heparin exposure (SMD: 0.33-0.50) and be treated with a non-heparin anticoagulant (e.g., argatroban [SMD: 0.20-0.45], bivalirudin [SMD: 0.46-0.60], fondaparinux [SMD: 0.21-0.32]) after diagnosis.
CONCLUSIONS: This study uses a diagnostically complex condition to highlight the importance of using laboratory-confirmed diagnosis codes in pharmacoepidemiology studies, particularly for conditions with low prevalence, as the predictive value of the diagnosis code itself did not reliably predict HIT.
METHODS: This was a retrospective, observational cohort study of adult patients ≥18 years old in the TriNetX Dataworks-USA research network of de-identified electronic health record (EHR) data with an incident HIT diagnosis (ICD-10-CM code D75.82) (index date) from 01/01/2018 to 12/31/2024. Patients were required to have ≥1 year of baseline history and were followed for up to 42 days after index date. Lab-confirmed HIT was determined based on an ELISA optical density ≥2.00 and/or a positive serotonin release assay or heparin-induced platelet aggregation assay. The sensitivity, specificity, and positive predictive value (PPV) of the HIT diagnosis code was estimated.
RESULTS: A total of 8,537 patients had ≥1 diagnosis code for HIT. Performance of the code included an estimated 36.5% sensitivity, 99.99% specificity, and 21.1% PPV. As further restrictions were applied, the PPV increased to 30.9% (≥1 inpatient code) and 34.4% (≥2 codes 2-14 days apart). Across all groups, demographics and comorbidities/risk factors were similar between patients who were true and false positives. Patients who were true positives were more likely to have recent unfractionated heparin exposure (SMD: 0.33-0.50) and be treated with a non-heparin anticoagulant (e.g., argatroban [SMD: 0.20-0.45], bivalirudin [SMD: 0.46-0.60], fondaparinux [SMD: 0.21-0.32]) after diagnosis.
CONCLUSIONS: This study uses a diagnostically complex condition to highlight the importance of using laboratory-confirmed diagnosis codes in pharmacoepidemiology studies, particularly for conditions with low prevalence, as the predictive value of the diagnosis code itself did not reliably predict HIT.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
EPH39
Topic
Epidemiology & Public Health
Topic Subcategory
Disease Classification & Coding
Disease
SDC: Systemic Disorders/Conditions (Anesthesia, Auto-Immune Disorders (n.e.c.), Hematological Disorders (non-oncologic), Pain)