EUVIMEDEUVIMED entdecken
CrossrefInfokarte

Systemic Challenges in EHR Alert Design: A Thematic Analysis of Physician Discourse on X (Formerly Twitter)

Guy Vishnevsky, Jacob N. Ablin

Applied Clinical Informatics · 2026-08-01 · 17 (04) · 781-788

DOI: 10.1055/a-2945-3280 ↗
AbstractBibliografische Angaben

Abstract

Abstract Background Electronic health record (EHR) clinical decision support (CDS) systems generate excessive, non-contextual alerts, causing clinician desensitization with override rates exceeding 90%. This alert fatigue disrupts workflows, contributes to burnout, and creates patient safety risks. Traditional survey-based methods fail to capture real-time end-user feedback at scale. Objectives This study aimed to evaluate real-world CDS usability signals by analyzing physician discourse on X (formerly Twitter) and identify physician-perceived sociotechnical pain points to inform future, hypothesis-driven EHR optimization. Methods X was systematically searched in August 2025 using Grok 4 (2025-08 build) via an internal API (Application Programming Interface). The AI temperature was set to 0.7 for search generation to ensure a broad retrieval of relevant discourse, utilizing keywords including “EHR alert fatigue,” “clinical decision support usability,” “EMR burnout,” “override EHR,” and “alert fatigue doctors,” yielding 117 posts from 106 self-identified physicians after manual curation (spanning from 2016 to 2025). Only the first post per author was included in the final analysis of 106 posts. We applied Braun and Clarke’s reflexive thematic analysis, with artificial intelligence (AI)-assisted coding verification. Sentiment classification was conducted as a secondary assessment. Results Six CDS usability failure domains were identified: Alert Overwhelm (38.7%, 95% confidence interval [CI]: 30.0–48.2%), Burnout and Fatigue (43.4%, 95% CI: 34.4–52.9%), Patient Safety Threats (9.4%, 95% CI: 5.2–16.5%), Regulatory and Billing Pressures (17.0%, 95% CI: 11.0–25.3%), Mitigation Strategies, including AI prioritization (36.8%, 95% CI: 28.2–46.3%), and Lack of Customization (7.5%, 95% CI: 3.9–14.2%). Sentiment was 76.4% (95% CI: 67.5–83.5%) negative. Co-occurrence analysis showed 50.0% of safety posts co-occurred with alert overwhelm, and 19.6% of burnout posts co-occurred with regulatory and billing pressures. Conclusion Physician discourse on X provides valuable, exploratory usability signals. The prevalence of alert overwhelm and burnout themes highlights recurring pain points in current EHR architectures, generating hypotheses for future user-centered CDS redesign such as tiered alerting, role-based routing, and alert deferral mechanisms.

Wissen, das zu Ihrer Arbeit passt.

Vom kostenlosen Einstieg über die persönliche Recherche bis zur vollständigen Research Suite.

Tarife ansehen →

Sind alle Verlagsvolltexte enthalten?

Nein. Bibliografische Nachweise, Abstracts und Volltexte sind unterschiedliche Inhalte. Ob ein Volltext verfügbar ist, hängt von der Quelle, Open Access und Ihren vorhandenen Lizenzen ab. Ein EUVIMED-Tarif ersetzt keine zusätzlichen Verlagsrechte.

Tarife ansehen →