markov_model: Definition of Markov model and trace

In packDAMipd: Decision Analysis Modelling Package with Parameters Estimation Ability from Individual Patient Level Data

Description

Definition of Markov model and trace

Usage

markov_model(
  current_strategy,
  cycles,
  initial_state,
  discount = c(0, 0),
  parameter_values = NULL,
  half_cycle_correction = TRUE,
  state_cost_only_prevalent = FALSE,
  state_util_only_prevalent = FALSE,
  method = "half cycle correction",
  startup_cost = NULL,
  startup_util = NULL
)

Arguments

current_strategy	strategy object
cycles	no of cycles
initial_state	value of states initially
discount	rate of discount for costs and qalys
parameter_values	parameters for assigning health states and probabilities
half_cycle_correction	boolean to indicate half cycle correction
state_cost_only_prevalent	boolean parameter to indicate if the costs for state occupancy is only for those in the state excluding those that transitioned new. This is relevant when the transition cost is provided for eg. in a state with dialysis the cost of previous dialysis is different from the newly dialysis cases. Then the state_cost_only_prevalent should be TRUE
state_util_only_prevalent	boolean parameter to indicate if the utilities for state occupancy is only for those in the state excluding those that transitioned new.
method	what type of half cycle correction needed
startup_cost	cost of states initially
startup_util	utility of states initially if any

Details

Use the strategy, cycles, initial state values creating the markov model and trace. As many probabilities /cost/utility value depend on age/time the evaluation and assignment happens during each cycle. At the heart it does a matrix multiplication using the previous row of the trace matrix and the columns of the transition matrix. Also checks for population loss, calculates cumulative costs and qalys (accounts for discounting and half cycle correction)

Value

Markov trace

Examples

tmat <- rbind(c(1, 2), c(3, 4))
colnames(tmat) <- rownames(tmat) <- c("Healthy", "Dead")
tm <- populate_transition_matrix(2, tmat, c(0.5, 0.5, 0, 1))
a <- health_state("Healthy", 1, 1, 0, FALSE)
b <- health_state("Dead", 1, 0, 0, TRUE)
health_states <- combine_state(a, b)
this.strategy <- strategy(tm, health_states, "intervention")
markov_model(this.strategy, 10, c(1, 0))

packDAMipd documentation built on March 3, 2021, 5:07 p.m.