Abstract TP266: Linking Inpatient Coverdell Stroke Program Registry to Commercial Claims Data in the United States

    loading  Checking for direct PDF access through Ovid

Abstract

Background: Linking inpatient stroke registries to commercial claims data can provide a unique data source with detailed in-hospital data and long-term information on patient history, post-stroke care, and health outcomes, to answer clinical and pharmacoepidemiological research questions. Such linked data can offer a broader picture than the solely linkage to Medicare Fee-for-Service data.

Objectives: To assess the feasibility of linking the Paul Coverdell National Acute Stroke Program registry (PCNASP) to commercial claims data and to evaluate the representativeness of the identified linked population.

Methods: All stroke admission records entered in PCNASP between 2008-2015 were considered for linkage to longitudinal patient claims records from a large U.S. commercial health insurer using an algorithm based on age, gender, admission date, discharge date, primary diagnosis, and state. We assessed the linkage quality via the uniqueness level, i.e., the percentage of unique records among the linked subset. The representativeness of the claims-PCNASP-linked population was assessed via absolute standardized differences (SD) for a range of patient and hospital characteristics among linked and unlinked patients with a hospital discharge diagnosis of ischemic stroke.

Results: From a claims-based population of 105,033 stroke hospitalizations, a total of 5,975 records (5.7%) were linked to PCNASP with a 99.1% uniqueness. Baseline patient characteristics and in-hospital information between linked and unlinked patients were overall balanced, as assessed by SD<0.1, despite minor imbalances (0.1

Conclusions: A high-quality linkage between PCNASP and commercial claims data using indirect identifiers is feasible allowing to combine detailed disease markers of the acute stroke care episode with long-term follow-up. Despite minor imbalances, the linked subset appeared to be representative of the overall claims-based population with stroke.

Related Topics

    loading  Loading Related Articles