Good Statistical Monitoring QTL `{gsm.qtl}` R package

This README provides a high-level overview of gsm.qtl; see the package website for additional details.

The Good Statistical Monitoring or gsm suite of R packages provides a framework for statistical data monitoring. gsm.qtl is an extension package that contains the additional functions, workflows, and document template to generate reports for quality tolerance limits (QTL). See gsm.core for an overview of the ecosystem.

Background

The gsm.qtl package calculates study level metrics that can be represented by some sort M/N value that is used to evaluate a particular QTL of a particular study, such as inclusion/exclusion criteria violations, early study discontinuation, etc.

Mapping

Datasets necessary to calculate a particular QTL can be requested or customized based off of previously existing mappings, see gsm.mapping. Additional mappings that have been added to support gsm.qtl include the IE and EXCLUSION mappings.

Metrics

qtlxxxx metric yaml files have been added to calculate study level QTLs, these files follow similar structure to those found in gsm.kri.

Reporting

gsm.qtl has a report template to produce the data visualizations, widgets and tables that ultimately go into an html QTL report.

Installation

You can install the development version of gsm.qtl like so:

# install.packages("pak")
pak::pak("Gilead-BioStats/gsm.qtl@dev")

Sample Code

This is a basic example showing how to create interactive widget visualizations based on reporting outputs from the gsm.reporting package:

library(gsm.datasim)
library(gsm.mapping)
library(gsm.reporting)
library(gsm.core)
library(gsm.kri)
library(purrr)
library(dplyr)
library(gsm.qtl)
set.seed(1234)

# Single Study
ie_data <- generate_rawdata_for_single_study(
  SnapshotCount = 6,
  SnapshotWidth = "months",
  ParticipantCount = 1000,
  SiteCount = 10,
  StudyID = "ABC",
  workflow_path = "workflow/1_mappings",
  mappings = c("IE", "PD", "STUDCOMP"),
  package = "gsm.mapping",
  desired_specs = NULL
)

mappings_wf <- gsm.core::MakeWorkflowList(
  strNames =c("SUBJ", "ENROLL", "IE", "PD", "STUDY", "SITE", "COUNTRY", "EXCLUSION", "STUDCOMP"),
  strPath = "workflow/1_mappings",
  strPackage = "gsm.mapping"
)
mappings_spec <- gsm.mapping::CombineSpecs(mappings_wf)
metrics_wf <- gsm.core::MakeWorkflowList(strNames = c("qtl0001", "qtl0002"), strPath = "inst/workflow/2_metrics", strPackage = "gsm.qtl")
reporting_wf <- gsm.core::MakeWorkflowList(strNames = c("Results", "Groups"), strPath = "workflow/3_reporting", strPackage = "gsm.reporting")

lRaw <- map_depth(ie_data, 1, gsm.mapping::Ingest, mappings_spec)
mapped <- map_depth(lRaw, 1, ~ gsm.core::RunWorkflows(mappings_wf, .x))
analyzed <- map_depth(mapped, 1, ~gsm.core::RunWorkflows(metrics_wf, .x))
reporting <- map2(mapped, analyzed, ~ gsm.core::RunWorkflows(reporting_wf, c(.x, list(lAnalyzed = .y, lWorkflows = metrics_wf))))

dates <- names(ie_data) %>% as.Date
reporting <- map2(reporting, dates, ~{
  .x$Reporting_Results$SnapshotDate <- .y
  .x
})

# Bind multiple snapshots of data together
all_reportingResults <- do.call(dplyr::bind_rows, lapply(reporting, function(x) x$Reporting_Results)) %>%
  select(-c(upper_funnel, flatline))


# Only need 1 reporting group object
all_reportingGroups <- reporting[[length(reporting)]]$Reporting_Groups

report_listings <- list(qtl0001 = mapped$`2012-06-30`$Mapped_EXCLUSION,
                        qtl0002 = left_join(mapped$`2012-06-30`$Mapped_STUDCOMP,
                                            select(mapped$`2012-06-30`$Mapped_SUBJ, subjid, country),
                                            by = "subjid"))

# Test if new Report_QTL rmd works
gsm.kri::RenderRmd(
  lParams = list(
    dfResults = all_reportingResults,
    dfGroups = all_reportingGroups,
    lListings = report_listings
  ),
  strOutputDir = getwd(),
  strOutputFile = "test.html",
  strInputPath = system.file("report/Report_QTL.Rmd", package = "gsm.qtl")
)

gsm.qtl