UIP4RCT

The source codes for "Unit Information Prior for Incorporating the Real-World Evidence into Randomized Controlled Trials". See

Notebooks (.ipynb):

• UIP4RCT_ExperimentSettings: Generate the plot that describe distributions of propensity scores under different scenarios. (Figure 1)
• UIP4RCT_Simulation: Perform the simulations of Scenarios 1-3 for continuous, binary and survivial outcomes. (Tables 1-3)
• UIP4RCT_Simulation_Extreme_Linear&Binary: Perform the simulations of Scenarios 4&5 for linear and binary outcomes. (Tables 1&2)
• RealDataApplication_HCQ: Real data analysis for comparing UIP and NIP. (Table 4)

Scripts (.py):

• balance_methods: Functions for different balancing methods (IPW and Match).
• datgen: Functions to generate simulated dataset.
• samplers: MCMC sampler for UIP and models.

Dataset (.csv):

• HCQ_RCT: Patient-level data reconstructed from Figure 2 in the HCQ study (https://doi.org/10.1371/journal.pone.0257238).