376. MEDICATION-SEEKING BEHAVIORS AND CORRELATED CHARACTERISTICS OF ZOLPIDEM USERS IN TAIWAN - A SHARED PATIENT NETWORK ANALYSIS

Abstract Background Increasing insomnia signals a public health problem, alongside rising zolpidem use. This study investigates the factors behind the disproportionate rise in zolpidem prescriptions in Taiwan. Aims & Objectives It aims to identify the determinants of high-dose zolpidem users in Taiwan’s Yilan County and employ an innovative approach to outline their medication-seeking patterns, using Taiwan’s healthcare database. Method The associations between sociodemographic and clinical factors and low-dose and high-dose users were analyzed using multiple logistic regression. Social network analysis was employed to explore medication-seeking behavior among these user groups across different healthcare institutions. Results Among 5290 participants, 22.82% are high-dose users. Males have a 1.33-fold higher risk and participants with chronic diseases have more than four times the risk of being high-dose zolpidem users. A social network analysis showed a higher density (0.52) for high-dose users, revealing their frequent visits, for zolpidem, to different healthcare institutions. Psychiatrists have a central role in both low-dose and high-dose user networks, with a greater influence on low-dose users (64.4) than high-dose users (32.2). Discussion & Conclusions Males, individuals aged 40–60 years, and those with chronic diseases, catastrophic illness certifications, economic disadvantages, or mood and physical disorders exhibit a higher risk of high-dose zolpidem use. High-dose users demonstrate frequent doctor shopping and medication refills across institutions. Psychiatrists play a key role in zolpidem-seeking networks but with reduced influence among high-dose users. Social network analysis highlights distinct patterns between high- and low-dose users, emphasizing differing medical needs. These findings guide policymakers in addressing zolpidem-related medication-seeking behaviors effectively.