Engine-specific trial simulation functions

These functions are the inner functions called byrun_engine(). In this help file, only the most common engine functionsare listed. To identify the appropriate engine functionfor your FACTS file, callget_facts_engine().

• run_engine_aipf_contin(): Enrichment continuous.

• run_engine_aipf_dichot(): Enrichment dichotomous.

• run_engine_aipf_tte(): Enrichment time to event.

• run_engine_contin(): Core continuous.

• run_engine_crm(): continual reassessment method (CRM).

• run_engine_dichot(): Core dichotomous.

• run_engine_multep(): Multiple endpoint.

• run_engine_tte(): Time to event.

Arguments

param_files

Character vector of file paths or the outputofprep_param_files(). If a character vector, the elementscan be directories containing*.param files or the paths to the*.param files themselves.Such a directory is returned byrun_flfll().

n_sims

Positive integer, number of simulations per param file.

mode

Character scalar:"s" for simulation mode in non-enrichmentdesigns,"" for simulation mode in enrichment designs,"r" for execution mode, and"p" for prediction mode.For the CRM engine,mode needs to be"s" or"".

seed

Positive integer, random number generator seed for theactual trial simulations. Use thisseed argument instead offlfll_seed (run_facts(),run_flfll()) to controlpseudo-randomness in the actual trial simulations.flfll_seedonly controls how the*.param files are generated.

analysis_data

Character, analysis mode patient data file name.

analysis_mode

Logical, whether to activate analysis mode.

current_week

Numeric, current time in weeks.

execdata

Character, name of the execution mode patient file.

final

Logical, whether to do the final analysis.For execution mode only.

interim

Integer, interim number.

mcmc_num

Integer, MCMC file number. For analysis mode only.

verbose

Logical, whether to print progress information to the Rconsole.

version

Character scalar, version of FACTS corresponding to theFACTS file. Get by callingget_facts_version() on your FACTS file.See possible versions withget_facts_versions().Do not supplyversion torun_engine().run_engine()detects the version automatically from the FACTS file andpasses it to the appropriate engine function.

arm_selection

Logical, whether to activate arm selection.

armsdropped

Character, a comma-separated collection of integersindicating dropped arms.

complete_data_analysis

Logical, whether to do a completedata analysis.

fsimdata

Character, prediction mode patient data file name.

fsimexp

Logical. For expert use only.

fsimparam

Character, name of the prediction mode*.param file.

keepfiles

Logical, whether to deactivate cleanup ofextraneous staged design files.

noadapt

Logical, whether to deactivate adaptive actionsin prediction mode.

s2_aux_paramfile

Character, name of the stage 2 executionauxiliary*.param file.

stage

Integer, trial design stage. For staged designs only.

directory

Character, working directory. CRM only.

allocator

Logical, allocator/execution/recommender mode. CRM only.

charting_info

Logical, unused.

estimator

Logical, use estimator. CRM only.

force_cohort

Logical, whether to force small cohort run-in to end.CRM only.

reduced_priority

Logical, whether to run at reduced priority.CRM only.

Value

Nothing.

Details

If you need to repeatedly invoke an engine, as with mosttrial execution mode workflows, these engine functions may be slowon their own. To avoid the most severe sources of slowness,consider runningprep_param_files()and then passing the result to one of the individual enginefunctions (such asrun_engine_contin()).

See also

Examples

facts_file<-get_facts_file_example("contin.facts")# Can only run if system dependencies are configured:if(file.exists(Sys.getenv("RFACTS_PATHS"))){out<-run_flfll(facts_file, verbose=FALSE)# Generate param files.# Identify which engine you need.get_facts_engine(facts_file)# Run the sims with the engine function or `run_engine()`.run_engine_contin(out, n_sims=1, verbose=FALSE, version="6.2.5")read_patients(out)}