In Reply We agree with Dr van Essen and colleagues that observational studies harbor the risk of unmeasured confounding. Treatment of patients with intracerebral hemorrhage has been studied in only a few trials with considerable variance in management protocols. Hence, large multicenter observational studies must be relied on to provide treatment estimates. To address the limitation of unmeasured confounding, we compared our study results (propensity score–matched analysis) with instrumental variable analysis, which is potentially more resilient to unmeasured confounding,^1,2 as suggested by van Essen and colleagues.

中文翻译：

---

小脑颅内出血-应答治疗观察研究中的不可估量的混淆。

在答复中，我们同意van Essen博士及其同事的观点，即观察性研究存在无法衡量的混杂风险。仅在少数试验中研究了脑出血患者的治疗，但治疗方案差异很大。因此，必须依靠大型的多中心观察研究来提供治疗估计。为了解决未测量混杂的限制，我们比较我们的研究结果（倾向得分匹配分析）与工具变量分析，这是潜在的更有弹性不可测量混杂，^{1 ，2}由van Essen的和他的同事的建议。