用于辅助完整腹主动脉瘤管理的预测模型的开发与验证
本研究开发并外部验证了一种离散事件模拟模型,用于预测腹主动脉瘤患者的自然病史及治疗后30天和2-5年生存率,模型在修复队列(n=320)和监测队列(n=376)中验证,30天死亡率预测c-statistic为0.87,2-5年生存预测c-statistic为0.68-0.73。
要点
- 模型预测腹主动脉瘤修复后30天死亡率c-statistic为0.87(95%CI 0.75-1.0)。
- 模型预测2-5年生存率的c-statistic在修复队列为0.68-0.71,监测队列为0.69-0.73。
- 作者认为该工具可辅助临床决策,但需在更广泛人群中进一步验证。
核心结论
- 腹主动脉瘤临床决策工具能准确预测腹主动脉瘤患者的5年生存率。
- 该工具可在临床决策中用于更好地告知临床医生和患者长期结局。
- 需要在更广泛的腹主动脉瘤人群中进行进一步验证研究以测试该工具的更广泛临床效用。
研究设计与样本
- 构建离散事件模拟模型以模拟腹主动脉瘤患者的自然病史并预测治疗和监测患者的30天及2-5年生存率。
- 模型输入参数来自已发表文献和新西兰国家生命表。
- 使用接受腹主动脉瘤修复的队列(n=320)和接受小腹主动脉瘤监测的队列(n=376)进行外部验证,所有患者随访至少5年。
主要发现
- 每个数据集运行模型三次,产生SD<1%,表明重现性极好。
- 接受腹主动脉瘤修复患者的观察30天死亡率为9/320(2.8%),模型预测死亡率为3.8%(c-statistic 0.87 [95%CI 0.75-1.0])。
- 修复队列预测2-5年生存率的c-statistic范围为0.68至0.71,小腹主动脉瘤监测患者为0.69至0.73。
临床怎么用
- 作者认为该工具可在临床决策中用于更好地告知临床医生和患者长期结局。
- 原文未据此给出推荐。
局限性
- 需要在更广泛的腹主动脉瘤人群中进行进一步验证研究以测试该工具的广泛临床效用。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
OBJECTIVE/BACKGROUND: Predicting outcomes prior to elective abdominal aortic aneurysm repair (AAA) requires critical decision making, as the treatment offered is a prophylactic procedure to prevent death from a ruptured AAA. The aim of this work was to develop and validate a model that may predict outcomes for patients with an AAA and hence aid in clinical decision making. METHODS: A discrete event simulation model was built to simulate the natural history of a patient with an AAA and to predict the 30 day and 2-5 year survival of patients undergoing treatment and surveillance. The input parameters of AAA behavior and impact of comorbidities on survival were derived from the published literature and the New Zealand national life tables. The model was externally validated using a cohort of patients that underwent AAA repair (n = 320) and a cohort of patients undergoing small AAA surveillance (n = 376). All patients had completed at least 5 years of follow up. RESULTS: The model was run three times for each data set to test. This produced a SD < 1%, indicating excellent reproducibility. The observed 30 day mortality for the patients undergoing AAA repair was 9/320 (2.8%) and the expected (model predicted) mortality was 3.8% (c-statistic 0.87 [95 confidence interval 0.75-1.0]). The c-statistic for the predicted 2-5 year survival ranged from 0.68 to 0.71 for the repaired AAA cohort and 0.69 to 0.73 for patients with a small AAA on surveillance. CONCLUSION: The AAA clinical decision tool has the ability to accurately predict the 5 year survival of patients with an AAA. This tool can be used during clinical decision making to better inform clinicians and patients of long-term outcomes. Further validation studies in a wider AAA population are required to test the broader clinical utility of this AAA clinical decision tool.
引用
- Khashram M, Kvizhinadze G, Khashram Z, et al. Development and Validation of a Predictive Model to Aid in the Management of Intact Abdominal Aortic Aneurysms. Eur J Vasc Endovasc Surg. 2018. PMID: 29643016. doi:10.1016/j.ejvs.2018.03.013
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。