"Strategies to Optimize Neuroimaging in the Emergency Department" by Long Hoang Tu

Strategies to Optimize Neuroimaging in the Emergency Department

Date of Award

Spring 2023

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Investigative Medicine

First Advisor

Forman, Howard

Abstract

Diagnostic errors are a major patient safety concern across all clinical settings, including the emergency department (ED). Acute neurologic conditions such as stroke are among the most common scenarios associated with delayed or missed diagnosis, as well as misdiagnosis-related harms. Despite exponential growth in diagnostic imaging use in the ED over the past few decades, there is no clear associated improvement in patient outcomes. Neuroimaging in particular represents a disproportionate share of advanced imaging used in the ED, with scenarios of suboptimal use contributing to false reassurance and missed diagnosis. Low-value imaging may also produce unnecessary cost, radiation exposure, and incidental findings leading to a domino effect of further testing. The ultimate aim of this work is to develop and evaluate strategies to improve neuroimaging utilization and therefore patient outcomes in acute neurological conditions. This thesis is composed of eight chapters organized around three specific aims: 1) to delineate major trends in the current utilization and diagnostic yield of neuroimaging in the ED, 2) to develop predictive tools assisting in the selection of patients for neuroimaging, and 3) to examine a specific use case (i.e., dizziness) where improved neuroimaging utilization can positively impact patient care, outcomes, and healthcare costs. Throughout this work, we will explore how optimizing the use of a diagnostic tests is an accessible and critical means to address some of the greatest challenges in current emergency care. First, in chapters 1 and 2, we examine trends in the ED utilization of non-contrast head CT and head and neck CTA respectively. CT and CTA imaging represent the vast majority of neuroimaging performed in acute care. We examine practice patterns at three emergency departments within a large academic medical center and use critical results reporting as a proxy for acute or actionable pathology. These chapters delineate presentations with relatively high imaging use and low diagnostic yield, suggesting the potential for improved patient evaluation via risk stratification or alternative approaches. Highlighted chief concerns fitting this profile include dizziness, syncope, paresthesias, migraines, and psychiatric complaints, among other non-focal and non-traumatic presentations. Next, we investigate specific presentations associated with low yield neuroimaging: acute psychiatric concerns and dizziness. In chapter 3, we detail criteria in which CT imaging may be avoidable for medical clearance of common psychiatric complaints (including hallucinations, delusions, and suicidal ideation). We derive a simple decision rule in which patients with isolated psychiatric concerns, but no other indications for neuroimaging, may not require CT as the risk of underlying or detectable abnormality is near-zero. In chapter 4, we expand this concept to the evaluation of dizziness by CTA head and neck. We derive a decision rule, composed of both well-recognized and perhaps underappreciated risk factors, which together delineate a sub-group of dizzy patients with near-zero risk of acute vascular pathology on CTA. In the last section of the thesis, we examine the feasibility and impact of alternative diagnostic approaches to patients presenting with dizziness and potential posterior circulation stroke. In chapter 5 and 6, we present technical specifications and logistical considerations for the use of an abbreviated and specialized MRI protocol, which has greater sensitivity for stroke in dizzy patients than commonly used CT and CTA exams. In chapter 7, we present evidence that the use of MRI in dizzy patients improves detection of underlying pathology, increases secondary preventative measures for stroke, and may reduce ED readmissions. Finally, in chapter 8, we use a decision-analytic model to assess the potential long-term costs and outcomes associated with use of differing imaging in the evaluation of dizzy patients. We find that MRI – specialized MRI most of all – improves long-term outcomes with greater cost-effectiveness than CT or CTA-based approaches. In summary, we outline a variety of strategies that can be used to interrogate and optimize the use of diagnostic testing in emergency care; we show that such improvements can positively impact patient care and outcomes. While we have specifically examined neuroimaging and focused on psychiatric and posterior circulation stroke presentations, the approaches demonstrated in this work are generalizable to other scenarios. Future research may explore the impact of using similar strategies to improve diagnosis across a variety of clinical settings.

This document is currently not available here.

Share

COinS