Limwattananon Chulaporn, Limwattananon Supon
J Med Assoc Thai. 2014 May;97 Suppl 5:S108-12.
The paper gives an overview of the four fundamental elements that should be considered when constructing a Markov model of cancers, including outcome measures, health state transition, transitional probabilities, and model calibration. The construction of any model of this kind should begin by establishing transition to the death state. The probability of this transition can be estimated using overall survival data from clinical studies. Possible health states over a cycle are defined according to the natural history of diseases and treatment pathways. Validity of the constructed model is tested against real patient data and the parameters are adjusted until the survival results are consistent.
本文概述了构建癌症马尔可夫模型时应考虑的四个基本要素，包括结局指标、健康状态转换、转移概率和模型校准。构建任何此类模型都应从确定向死亡状态的转换开始。这种转换的概率可以使用临床研究中的总生存数据来估计。一个周期内可能的健康状态根据疾病的自然史和治疗途径来定义。根据真实患者数据对构建模型的有效性进行检验，并调整参数，直到生存结果一致。
