Participatory System Dynamics vs Audit and Feedback: A Cluster Randomized Trial of Mechanisms of Implementation Change to Expand Reach of Evidence-based Addiction and Mental Health Care

我们的广泛目标是使所有医疗保健利益相关者能够为所有患者提供最高质量的护理。

Our broad aim is to empower all healthcare stakeholders to provide the highest quality care to all patients. Our specific aims address the complexities and tradeoffs associated with implementing evidence-based practices (EBPs) in outpatient addiction and mental health systems. There is scientific consensus about the best evidence-based psychotherapies and pharmacotherapies (EBPs) to meet the needs of patients with opioid and alcohol use disorder, PTSD and depression. However, EBP coordination over time, within and across multidisciplinary teams of providers, is complex and constantly changing. Veterans Health Administration (VA) policy mandates, national training programs, and incentivized quality measures, have been insufficient for reaching more than 3 to 28% of patients with the highest quality treatments. In fact, limited EBP reach is common in health systems and the field of implementation science seeks to address it. One routine strategy is data auditing with provider feedback (audit-and-feedback; AF), however, the impact is highly variable. As an alternative, participatory system dynamics (PSD) has been used to explain causes of complex problems in business management for 60 years. We partnered with frontline staff using PSD to determine how EBP reach emerges from local resources and constraints, and is determined by system dynamics, such as delays and feedback. The dynamics of EBP reach were formally specified in differential equation models, and tested against VA data drawn from a national VA SQL database. We used this existing enterprise data to tailor model parameters to each care team. PSD models were made accessible via a `Modeling to Learn' interface and training, during which teams safely evaluated local change scenarios via simulation to find the highest yield options for meeting Veterans' needs. PSD learning simulations produce immediate, real-time feedback to the teams who coordinate care, improving day-to-day decisions and long-term improvement plans. We propose a two-arm, 24-site (12 sites/arm) cluster randomized trial to test the effectiveness of PSD simulation as compared to more standard team AF data review. We hypothesize that PSD will be superior to AF for improving EBP initiation and dose (Aim 1). We will test the PSD theory of change that the effect of PSD on improved EBP reach is explained by improvement in team systems thinking (Aim 2). To confirm the potential for widespread usefulness of PSD, we will also test the generalizability of PSD causal dynamics across PSD and AF arms (Aim 3). This study has the potential to inform a new paradigm, by determining what works to improve health system quality defined as EBP reach, why it works, and under what conditions. If PSD is effective, study activities will address a national priority to improve Veterans' addiction and mental health care to prevent chronic symptoms, relapse, suicide and overdose. Findings from the proposed tests of effectiveness, causality, and generality, could also catalyze future applications to make a significant public health impact across the continuum of healthcare.