Caro J Jaime, Huybrechts Krista F, Xenakis James G, O'Brien Judith A, Rajagopalan Krithika, Lee Karen
Caro Research Institute, Concord, MA 01742, USA.
Curr Med Res Opin. 2006 Nov;22(11):2233-42. doi: 10.1185/030079906X148265.
To present a tool that allows estimation of the budget impact of treatments for acute mania in bipolar I disorder from a US healthcare payer perspective.
Using discrete event simulation, the course of individuals is simulated beginning with hospitalization. Discharge depends on symptom level measured by the Young Mania Rating Scale (YMRS). The treatment effect is determined using time-dependent regression equations derived from trial data, and decision rules obtained from clinical experts. Outcomes include: time to response and symptom resolution; proportion of subjects reaching each outcome; number of adverse events. Costs were obtained from hospital discharge databases, the National Medicare Physician Fee Schedule and RedBook. Different scenarios are examined, each describing the proportion of subjects on the various treatments (lithium, divalproex sodium, olanzapine, risperidone, and quetiapine--monotherapy and in combination with lithium). Analyses are intention-to-treat over 100 days, corresponding to follow-up in mania trials. Despite its flexibility and structural adaptability, the model has some important limitations related to the characteristics of the clinical trials. These include focus on inpatient management of acute mania, use of the YMRS as the model driver, polypharmacy restricted to two-drug regimens, no explicit consideration of titration and dose changes, and relatively short time horizon.
Scenarios with a greater proportion of quetiapine users (5% vs. 40% and 100%) result in a smaller impact on the healthcare budget (6912, 6277, and 5525 dollars per patient, respectively) and improvements in patient outcomes (e.g., 43%, 47%, and 54% responding at day 21; 74%, 77%, and 80% remitting by day 84). Sensitivity analyses showed that the budget impact is influenced by drug prices, discharge criteria and side-effect management.
Results suggest that increased use of quetiapine for bipolar mania in the US is economically justified and improves health outcomes. In addition, this model illustrates that discrete event simulation is a useful and versatile tool for budget impact analyses.
从美国医疗保健支付方的角度，介绍一种能够估算双相I型障碍急性躁狂治疗的预算影响的工具。
使用离散事件模拟，从住院开始模拟个体的病程。出院取决于用杨氏躁狂评定量表（YMRS）测量的症状水平。使用从试验数据推导的时间依赖性回归方程以及从临床专家获得的决策规则来确定治疗效果。结果包括：起效时间和症状缓解时间；达到每种结果的受试者比例；不良事件数量。成本数据来自医院出院数据库、国家医疗保险医师费用表和《红皮书》。研究了不同的情景，每种情景描述了接受各种治疗（锂盐、丙戊酸钠、奥氮平、利培酮和喹硫平——单药治疗以及与锂盐联合治疗）的受试者比例。分析采用意向性分析，为期100天，对应躁狂试验中的随访期。尽管该模型具有灵活性和结构适应性，但与临床试验的特点相关存在一些重要局限性。这些局限性包括专注于急性躁狂的住院管理、使用YMRS作为模型驱动因素、联合用药限于两药方案、未明确考虑滴定和剂量变化以及时间范围相对较短。
喹硫平使用者比例较高的情景（5% 与40% 和100%相比）对医疗保健预算的影响较小（分别为每位患者6912美元、6277美元和5525美元），并且患者结局有所改善（例如，在第21天有反应的比例分别为43%、47% 和54%；到第84天缓解的比例分别为74%、77% 和80%）。敏感性分析表明，预算影响受药品价格、出院标准和副作用管理的影响。
结果表明，在美国增加喹硫平用于双相躁狂的使用在经济上是合理的，并且改善了健康结局。此外，该模型表明离散事件模拟是用于预算影响分析的一种有用且通用的工具。
