Log in


  • Home
  • Using multiple natural experimental designs to triangulate the impact of a policy change

Donation goal

0%
Goal:
$500.00
Collected:
$0.00

Using multiple natural experimental designs to triangulate the impact of a policy change

  • Fri, July 17, 2020
  • 09:00 - 10:00
  • https://www.youtube.com/watch?v=PmixDUdpZQw

Time: 9:00 AM - 10:00 AM PDT (12:00 PM - 1:00 PM EST), Friday, July 17, 2020

Online Video Review:https://www.youtube.com/watch?v=PmixDUdpZQw

Zoom.us link will be sent 1 and 24 hours before the event.

Abstract: Estimating causal effects is best done using randomized controlled trials. Regrettably, for many policy-relevant questions only observational data are available. In this talk we explore how (i) decomposing a policy into multiple parts, and (ii) leveraging multiple study designs can produce stronger, more defensible causal estimates. We explore these statistical concepts through an example from health outcomes research: The feasibility and effectiveness of delaying surgery to transfer patients with acute type A aortic dissection—a catastrophic disease that requires prompt intervention—to higher-volume aortic surgery hospitals is unknown. We investigated the hypothesis that regionalizing care at high-volume hospitals for acute type A aortic dissections will lower mortality. We decomposed this hypothesis into subparts, investigating the isolated effect of transfer and the isolated effect of receiving care at a high-volume versus a low-volume facility. We used a preference-based instrumental variable design to address unmeasured confounding and matching to separate the effect of transfer from volume. We used the carefully deployed instrumental variable design, and contrast it with a decomposed propensity score study design, to gain deeper insights into the elements giving rise to confounding. We also develop a simple, but complete, form of sensitivity analysis to bound the effect of censored observations. 




Biosketch:  Michael Baiocchi, PhD, is an Assistant Professor in the Department of Epidemiology and Population Health at Stanford University. He is an interventional-statistician, creating interventions and the means for analyzing them. He specializes in creating simple, easy to understand statistical methodologies for getting reliable results out of messy data and messy situations. His research is in nonparametric estimation and design-based inference. Much of his applied work is in health outcomes and policy research. He was the inaugural Stein Fellow in the department of Statistics at Stanford University.  He is the principal investigator on a large (enrollment: 5,000+ students, 100+ schools) randomized study of a sexual assault prevention intervention in the informal settlements around Nairobi, Kenya. 



文本转语音功能仅限200个字符
选项 : 历史 : 反馈 : Donate 关闭



@Dahshu 2020

文本转语音功能仅限200个字符
选项 : 历史 : 反馈 : Donate 关闭
G
M
T
Y
文本转语音功能仅限200个字符
Powered by Wild Apricot Membership Software