Impact of initial buprenorphine prescribing strategies on retention in care for patients with opioid use disorder

项目摘要/摘要 阿片类药物过量继续以惊人的速度夺走生命，#年达到创纪录的69,710人死亡。 2020年。治疗乌德的药物，如丁丙诺啡，可以降低阿片类药物过量死亡的风险。虽然保单 努力扩大了丁丙诺啡的治疗能力，丁丙诺啡由于贫乏而仍未得到充分利用 留守在看护中心。来自真实世界环境的观测数据显示，180天的保留率低至 25%-40%。因此，联邦过量用药预防战略确定了改善保留率的研究是关键 优先考虑。为此，检查丁丙诺啡给药方案的研究可以确定改进策略 留存。来自试验和观察性研究的数据表明，较高的诱导和维持剂量 可以提高保持率。很少有文献描述当前丁丙诺啡的剂量实践，但我们的数据 确实有建议的可变剂量做法，许多患者从未达到推荐的剂量目标。 此外，研究表明丁丙诺啡剂量水平有下降趋势，这意味着 剂量不足的情况可能正在加剧。在这份导师研究职业发展奖的申请中，我建议 检查丁丙诺啡剂量方案与OUD治疗保留的关系。具体来说， 这项工作的目标是1)描述目前丁丙诺啡的剂量实践 轨迹，2)评估两种不同的丁丙诺啡剂量策略对保留率的影响，以及3) 评估改进的丁丙诺啡给药方法对留置的人群水平、临床和成本的影响 以及马萨诸塞州的服药过量。同时，我确定了三个培训目标：1)发展 使用机器学习聚类方法的技能，2)使用目标试验仿真学习因果推理，以及 3)培养仿真建模技能。通过这些培训目标，我将全面发展 数据科学方面的协同能力。这项建议中的研究直接建立在我之前使用 药房声称数据，以检查丁丙诺啡收据的趋势。这项提议充分利用了这一先例 研究丁丙诺啡特定剂量策略对保留的作用的经验。我已经确定了一个 支持性的研究环境和经验丰富的指导团队，具有)内容和方法专业知识 (B)在指导受训人员从事独立研究职业方面有广泛的记录。这些力量的力量 资源将确保我获得成功完成拟议工作范围所需的专业知识 并实现我的长期目标，成为一名独立的内科研究员，使用大量数据进行研究 物质使用障碍治疗的创新。在这个奖项结束时，我将处于有利地位，可以申请 使用国家医疗补助数据检查a)新出现的丁丙诺啡剂量的影响的R级奖项 在芬太尼掺假渗透阿片类药物供应时的保留策略，以及b)在 丁丙诺啡剂量按种族、民族和地区划分。

PROJECT SUMMARY / ABSTRACT Opioid overdose continues to take lives at an astounding rate, reaching a record high of 69,710 deaths in 2020. Medications for OUD such as buprenorphine reduce risk of opioid overdose mortality. While policy efforts have expanded buprenorphine treatment capacity, buprenorphine remains underutilized due to poor retention in care. Observational data from real-world settings demonstrate 180-day retention rates as low as 25-40%. Thus, the federal Overdose Prevention Strategy identifies research to improve retention as a key priority. To this end, research examining buprenorphine dosage regimens could identify strategies to improve retention. Data from trials and observational studies suggest that higher induction and maintenance dosage could improve retention. Scarce literature describes current buprenorphine dosage practices, but the data we do have suggest variable dosage practices, with many patients never achieving recommended dosage targets. Furthermore, research has demonstrated downward trends in buprenorphine dosage levels, meaning that underdosage may be growing. In this Mentored Research Career Development Award application, I propose to examine the relationship between buprenorphine dosage regimens and OUD treatment retention. Specifically, the goals of this work are to 1) describe current buprenorphine dosage practices in terms of dosage trajectories, 2) estimate the impact of two distinct buprenorphine dosage strategies on retention, and 3) estimate the population-level clinical and cost impact of improved buprenorphine dosage practices on retention and overdose for the state of Massachusetts. In parallel, I have identified three training objectives: 1) develop skills using machine learning clustering methods, 2) learn causal inference using a target trial emulation, and 3) develop simulation modeling skills. Through these training objectives, I will develop comprehensive synergistic skills in the data sciences. The research in this proposal builds directly on my prior work using pharmacy claims data to examine trends in buprenorphine receipt. This proposal leverages this prior experience to investigate the role of specific buprenorphine dosage strategies on retention. I have identified a supportive research environment and an experienced mentor team with a) content and methodologic expertise and b) extensive track records in guiding trainees to independent research careers. The strength of these resources will ensure that I gain the expertise necessary to succeed in completing the proposed scope of work and achieve my long-term goal of becoming an independent physician investigator using large data to study innovations for substance use disorder treatment. At the end of this award, I will be well-positioned to apply for an R-level award using national Medicaid data to examine a) the impact of emerging buprenorphine dosage strategies on retention as fentanyl adulteration penetrates the opioid supply, and b) disparities in buprenorphine dosage by race, ethnicity and region.