Abstract 049: Smart Rx

    loading  Checking for direct PDF access through Ovid

Abstract

Introduction: Chronic cardiac conditions, such as heart failure (HF) and atrial fibrillation (AF), have complex pathophysiology that can often pose challenges when treating comorbid conditions, especially with respect to drug-disease interactions. Although much is known scientifically about these interactions, integrating such evidence into clinical practice with the intent of avoiding negative health outcomes and avoiding excess healthcare costs has proved challenging. Here, we present: (1) a novel software platform that enables real-time assessment of incidence of drug-disease interactions among patients with CHF taking medications for other comorbid indications in a large cohort; and (2) development of a natural language processing (NLP)-driven software platform for performing such quality improvement assessments through automated searches of the electronic health record (EHR) of patients with any cardiac disease.

Methods and Results: An NLP-driven search tool (Figure 1A) was developed based upon the Queriable Patient Inference Dossier (QPID) created by the Laboratory of Computer Science at the Massachusetts General Hospital (MGH) that incorporates an automated assessment of chronic cardiac diseases and of medications that have been identified as absolutely or relatively contraindicated in patients with specific cardiac diseases. The tool, which can be customized to assess for medications or other factors important in cardiovascular care, was used to perform an automated cross-sectional study of medications prescribed to 2,500 patients with CHF currently under care at the MGH. All records were manually reviewed by two independent observers, with a Cohen’s κ of greater than 0.8. Nearly 60% of patients were prescribed agents in the past three months that are known to cause major or moderate adverse events, with three drugs (citalopram, doxazosin and ibuprofen), accounting for over 80% of the contraindicated agents (Figure 1B).

Conclusion: Use of prescription drugs known to cause major or moderate adverse events among patients with HF is quite common. An automated software platform for rapid real-time assessment of drug disease interactions can help identify gaps in care and help improve the quality of care among patients with cardiovascular disease and multiple comorbidities.

Related Topics

    loading  Loading Related Articles