Incertus.jl

The package Incertus.jl implements several restricted randomization procedures for two- and multi-arm clinical trials, targeting equal or unequal allocation.

• Randomization, targeting equal 1:1 allocation
• Randomization, targeting unequal allocation in two-arm trial and equal/unequal allocation in multi-arm trial
• Procedures implemented in the package
• • Completely Randomized Design (CRD)
  • Permuted Block Design (PBD)
  • Random Allocation Rule (Rand)
  • Truncated Binomial Design (TBD)
  • Truncated Multinomial Design (TMD)
  • Efron's Biased Coin Design (EBCD)
  • Adjustable Biased Coin Design (ABCD)
  • Generalized Biased Coin Design (GBCD)
  • Big Stick Design (BSD)
  • Biased Coin Design With Imbalance Tolerance (BCDWIT)
  • Block Urn Design (BUD)
  • Ehrenfest Urn Design (EUD)
  • Bayesian Biased Coin Design (BBCD)
  • Mass Weighted Urn Design (MWUD)
  • Doubly-Adaptive Biased Coin Design (DBCD)
  • Maximum Entropy Constraint Balance Randomization (MaxEnt)
  • Drop-the-Loser Urn Design (DLUD)
  • Funcions implemented to calculate allocation probabilities
• Simulation
• Operational characteristics
• • Measures of imbalance
  • • Final imbalance
    • Expected absolute imbalance
    • Variance of imbalance
    • Expected maximum imbalance over first allocation steps
    • A cumulative average loss over the first allocation steps
  • Measures of randomness
  • • Cumulative average of expected proportions of correct guesses over first allocation steps under two different guessing strategies
    • Cumulative average of expected proportions of deterministic assignments over first allocation steps
    • Forcing index
  • Balance-randomness trade-off
  • Allocation ratio preserving property
• Visualizing operational characteristics
• Simulation example 1
• • Simulation setup
  • Simulation run
  • Simulation results
  • • A violin plot of the final imbalance after all treatment assignmnets complete
    • Expected absolute imbalance vs. allocation step
    • Variance of imbalance vs. allocation step
    • Expected maximum absolute imbalance over first allocation steps
    • Cumulative average loss over first allocation steps
    • Cumulative averages of expected proportions of correct guesses over first allocation steps (under the convergence guessing strategy)
    • Cumulative averages of expected proportions of correct guesses over first allocation steps (under the maximum probability guessing strategy)
    • Cumulative averages of expected proportions of deterministic assignments over first alocation steps
    • Forcing indeces vs. allocation step
    • Balance-randomness tradeoff vs. allocation step
• Simulation example 2
• • Simulation setup
  • Simulation run
  • Simulation results
  • • A violin plot of the final imbalance after all treatment assignmnets complete
    • Expected absolute imbalance vs. allocation step
    • Variance of imbalance vs. allocation step
    • Expected maximum absolute imbalance over first allocation steps
    • Cumulative average loss over first allocation steps
    • Cumulative averages of expected proportions of correct guesses over first allocation steps (under the convergence guessing strategy)
    • Cumulative averages of expected proportions of correct guesses over first allocation steps (under the maximum probability guessing strategy)
    • Cumulative averages of expected proportions of deterministic assignments over first alocation steps
    • Forcing indeces vs. allocation step
    • Balance-randomness tradeoff vs. allocation step
    • Allocation ratio preserving (ARP) property vs. allocation step
• Auxiliary functions