Jackson John W, Fulchino Lisa, Rogers James, Mogun Helen, Polinski Jennifer, Henderson David C, Schneeweiss Sebastian, Fischer Michael A
Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, MA, USA.
Schizophrenia Research Program, Department of Psychiatry, Massachusetts General Hospital, Boston, MA, USA.
Pharmacoepidemiol Drug Saf. 2018 Jan;27(1):95-104. doi: 10.1002/pds.4354. Epub 2017 Nov 23.
To quantify and explain variation in use of long-acting injectable antipsychotics (LAIs) in the United States, and understand the relationship between patient characteristics, drug reimbursement policies, and LAI prescribing after relapse.
A cohort of recently relapsed patients with schizophrenia ages 18 to 64, were identified immediately after discharge from a related inpatient hospitalization, partial hospitalization, or emergency room visit, drawn from 2004 to 2006 Medicaid claims, and followed for 90 days until LAI initiation. Data on state-level Medicaid prior authorization (PA) policies for LAIs were collected. Sequential longitudinal Poisson regression models were developed to understand the relationship between patient and PA policy variables and LAI prescribing, including prior adherence to oral antipsychotics, demographics, clinical variables, and presence of PA policy for LAI.
Among 36 282 patients, 3.1% received risperidone LAI, and 3.8% received a first-generation (FGA) LAI with wide variation across states. Prior adherence ranged from 29% to 89% but was marginally associated with initiation and did not explain variation for LAI prescribing. FGA initiation was associated with geography and race/ethnicity but not PA policy. For risperidone LAI initiation, demographics and clinical factors explained, respectively, 5.0% and 3.0% of the variation; PA policy had a large negative association with initiation (RR = 0.41; 95%CI 0.20-0.87) and explained 8.4% of the variation.
PA policies may represent a major treatment barrier for risperidone LAI among relapsed patients. Non-adherence plays a little role in predicting which patients receive LAIs. Policy makers and health insurers will need to consider these findings when guiding the use of LAIs. KEY POINTS Among a nationwide cohort of relapsed schizophrenia patients enrolled in US Medicaid, 3.1% received Risperdal Consta, a long-acting injectable antipsychotic (LAI), and 3.8% initiated a first-generation first-generation LAI within 90 days after discharge. During 2004 to 2006, there was marked variation in 90 day post-relapse initiation of Risperdal-Consta-a newly marketed medication during this period-and also marked variation in 90 day post-relapse initiation of any first-generation LAI, which appeared to be associated with race/ethnicity and geography. Prior authorization policies were associated with substantially lower initiation of Risperdal Consta in this cohort of relapsed patients even after accounting for clinical indication (non-adherence), relapse history, demographics, adjunctive medication, and mental health service use.
量化并解释美国长效注射用抗精神病药物（LAIs）使用情况的差异，了解患者特征、药物报销政策与复发后LAIs处方之间的关系。
从2004年至2006年医疗补助计划索赔中，识别出一组年龄在18至64岁、近期复发的精神分裂症患者，这些患者在相关住院治疗、部分住院治疗或急诊就诊出院后立即被纳入研究，并随访90天直至开始使用LAIs。收集了各州医疗补助计划对LAIs的预先授权（PA）政策数据。建立了序贯纵向泊松回归模型，以了解患者和PA政策变量与LAIs处方之间的关系，包括先前口服抗精神病药物的依从性、人口统计学特征、临床变量以及LAIs的PA政策。
在36282名患者中，3.1%接受了利培酮长效注射剂，3.8%接受了第一代长效注射剂，各州之间差异很大。先前的依从性范围为29%至89%，但与开始使用LAIs的相关性很小，且无法解释LAIs处方的差异。第一代长效注射剂的起始使用与地理位置和种族/民族有关，但与PA政策无关。对于利培酮长效注射剂的起始使用，人口统计学和临床因素分别解释了5.0%和3.0%的差异；PA政策与起始使用有很大的负相关性（RR = 0.41；95%CI 0.20 - 0.87），并解释了8.4%的差异。
PA政策可能是复发患者使用利培酮长效注射剂的主要治疗障碍。不依从在预测哪些患者接受LAIs方面作用不大。政策制定者和健康保险公司在指导LAIs的使用时需要考虑这些发现。要点在美国医疗补助计划登记的全国性复发精神分裂症患者队列中，3.1%接受了长效注射用抗精神病药物利培酮康斯坦，3.8%在出院后90天内开始使用第一代长效注射剂。在2004年至2006年期间，复发后90天开始使用利培酮康斯坦（在此期间新上市的药物）存在显著差异，任何第一代长效注射剂在复发后90天开始使用也存在显著差异，这似乎与种族/民族和地理位置有关。即使在考虑临床指征（不依从）、复发史、人口统计学特征、辅助药物和心理健康服务使用情况后，预先授权政策与该复发患者队列中利培酮康斯坦的起始使用显著降低有关。
