McEwan Phil, Bennett Hayley, Qin Lei, Bergenheim Klas, Gordon Jason, Evans Marc
Health Economics and Outcomes Research Ltd, Cardiff, UK.
Swansea Centre for Health Economics, Swansea University, Swansea, UK.
Diabetes Obes Metab. 2017 May;19(5):628-634. doi: 10.1111/dom.12865. Epub 2017 Feb 22.
Time-dependent HbA1c trajectories in health economic models of type 2 diabetes mellitus (T2DM) are typically informed by the UK Prospective Diabetes Study (UKPDS). However, this approach may not accurately predict HbA1c progression in patients who do not conform to the demographic profile of the original UKPDS cohort. This study aimed to develop an alternative mathematical model (MM) to simulate HbA1c progression in T2DM.
A systematic literature review identified studies, published between 2005 and 2015, that reported HbA1c in adult T2DM patients over a minimum duration of 18 months. Pooled data from eligible studies were used to develop an alternative MM equation for HbA1c progression, which was then contrasted with the UKPDS 68 progression equation in illustrative scenarios.
A total of 68 studies were eligible for data extraction (mean follow-up time 4.1 years). HbA1c progression was highly heterogeneous across studies, varying with baseline HbA1c, treatment group and patient age. The MM equation was fitted with parameters for mean baseline HbA1c (8.3%), initial change in HbA1c (-0.62%) and upper quartile of maximum observed HbA1c (9.3%). Differences in HbA1c trajectories between the MM and UKPDS approaches altered the timing of therapy escalation in illustrative scenarios.
The MM represents an alternative approach to simulate HbA1c trajectories in T2DM models, as UKPDS data may not adequately reflect the heterogeneity of HbA1c profiles observed in clinical studies. However, the choice of approach should ultimately be determined by the characteristics of individual patients under consideration and the clinical face validity of the modelled trajectories.
2型糖尿病（T2DM）健康经济模型中随时间变化的糖化血红蛋白（HbA1c）轨迹通常以英国前瞻性糖尿病研究（UKPDS）为依据。然而，这种方法可能无法准确预测不符合UKPDS原始队列人口统计学特征患者的HbA1c进展情况。本研究旨在开发一种替代数学模型（MM）来模拟T2DM患者的HbA1c进展。
通过系统文献回顾，确定了2005年至2015年间发表的、报告了成年T2DM患者至少18个月HbA1c情况的研究。符合条件研究的汇总数据用于开发HbA1c进展的替代MM方程，然后在示例场景中将其与UKPDS 68进展方程进行对比。
共有68项研究符合数据提取条件（平均随访时间4.1年）。不同研究间HbA1c进展差异很大，随基线HbA1c、治疗组和患者年龄而变化。MM方程拟合了平均基线HbA1c（8.3%）、HbA1c初始变化（-0.62%）和观察到的最大HbA1c上四分位数（9.3%）的参数。在示例场景中，MM与UKPDS方法之间HbA1c轨迹的差异改变了治疗升级的时机。
由于UKPDS数据可能无法充分反映临床研究中观察到的HbA1c谱的异质性，MM代表了一种在T2DM模型中模拟HbA1c轨迹的替代方法。然而，最终应根据所考虑个体患者的特征以及建模轨迹的临床表面效度来决定方法的选择。
