SUD Care Transitions - PROJECT SUMMARY The proposed study focuses on comprehensively characterizing the pathways of care utilized by people who obtain care for substance use disorder (SUD) and comparing the quality and effectiveness of these care pathways. Many effective SUD care pathways are recommended for use in healthcare systems. However, patients with SUD often report poor care experiences particularly regarding timely follow-up, provider and general satisfaction ratings, and care communication. Because SUD care pathways involve transitions across providers and venues of care, adequate treatment of SUD requires coordination across providers and settings to ensure that unmet needs are addressed, and appropriate SUD care is delivered. Surprisingly little is known about the real-world care pathways that people engage in once they are identified by a healthcare provider as having SUD. We propose to leverage a large novel data source from Kaiser Permanente Washington, a large integrated healthcare system in the northwestern United States. The dataset will include electronic healthcare records, insurance claims, and self-reported measures of substance use and SUD symptoms. To identify care pathways, we will apply continuous time, multi-state Markov modeling methods to empirically observe the longitudinal course of SUD care transitions undertaken by patients over time; each “state” or occurrence of care will be characterized by the intervention received (e.g., withdrawal management, psychotherapy, medication treatment) and setting of care (e.g., emergency department, primary care, behavioral health). Additional states and model parameters will be used to characterize SUD care quality, such as wait times between visits for SUD, concordance with SUD quality measures, and premature treatment termination. To compare the effectiveness of identified pathways, we will examine outcomes including overdose, mortality, remission, and substance-related medical conditions. Finally, employing an exploratory quantitative -> qualitative sampling design, we will purposively sample patients who engaged in the empirically derived pathways and interview them to understand their care experiences; we will also observe and interview healthcare providers to elucidate health system factors that facilitate or hinder their ability to link patients to adequate SUD care. SPECIFIC AIMS are: (1) Apply multi-state Markov models to comprehensively characterize care transitions across the full spectrum of SUD care among patients with SUD within a large integrated health system, (2) Generate data-driven insights to improve care delivery for SUD using estimated multi-state models, and (3) Observe and explore patient and provider experiences with care transitions across common care pathways using qualitative research methods. IMPACT: Researchers and health care constituents do not know how to optimize care for SUD in healthcare systems. This study will identify and compare the quality and effectiveness of common pathways that patients take when they obtain treatment for SUD and provide decision makers with information on how to better organize the delivery of SUD care.
