Obtain adverse event probabilities for treatment sequences given existing probabilities by treatment from a network meta-analysis.

ae_probs(n, struct, params_ae = iviNSCLC::params_ae_nma)

Arguments

n

The number of random observations of the parameters to draw.

struct

A model_structure object.

params_ae

Parameter estimates of the probabilities of adverse events in the same format as params_ae_nma.

Value

A list of matrices where each matrix corresponds to a type of adverse event. Rows of each matrix are posterior samples of the parameters and columns are treatment strategies. The list is an object of class "ae_probs" with an attribute "tx_abb" denoting the treatment used to estimate adverse events for each treatment sequence.

Examples

txseq1 <- txseq(first = "erlotinib", second = c("osimertinib", "PBDC"), second_plus = c("PBDC + bevacizumab", "PBDC + bevacizumab")) txseq2 <- txseq(first = "gefitinib", second = c("osimertinib", "PBDC"), second_plus = c("PBDC + bevacizumab", "PBDC + bevacizumab")) txseqs <- txseq_list(seq1 = txseq1, seq2 = txseq2) struct <- model_structure(txseqs) ae_probs <- ae_probs(n = 3, struct = struct) print(ae_probs[1:3])
#> $alt #> seq1 seq2 #> [1,] 0.04947135 0.11533762 #> [2,] 0.02617712 0.08485151 #> [3,] 0.02156232 0.09119395 #> #> $ast #> seq1 seq2 #> [1,] 0.10312491 0.10312491 #> [2,] 0.07267753 0.07267753 #> [3,] 0.10509515 0.10509515 #> #> $diarrhea #> seq1 seq2 #> [1,] 0.007514118 0.008037219 #> [2,] 0.013627760 0.018529599 #> [3,] 0.002408500 0.017021846 #>
tidy_ae_probs <- tidy(ae_probs) head(tidy_ae_probs)
#> strategy_id ae_name prob #> 1: 1 Elevated alanine transaminase 0.04947135 #> 2: 1 Elevated alanine transaminase 0.02617712 #> 3: 1 Elevated alanine transaminase 0.02156232 #> 4: 2 Elevated alanine transaminase 0.11533762 #> 5: 2 Elevated alanine transaminase 0.08485151 #> 6: 2 Elevated alanine transaminase 0.09119395