Bayesian Statistics

6/16/2022

Brian Dates: Pseudo control condition

Michigan - went strictly telehealth services after COVID pandemic started

2.1 million services delivered to more than 40,000 people

Bayesian Structural Time Series analysis:

Create a control group in Bayesian framework

Comparing two years worth of information to the prior two years

Preintervention (10/1/2019 - 3/2020) - postintervention

Three factors: regular time series, effect of implementation of telehealth service, end of implementation (end of executive order: 9/30/2021)? Post intervention

Data of the same type - comparable data

State Space model - prior and posterior: Bootstrap or MCMC process