Cookbook - Adverse Event KRI Report

Overview

This report demonstrates Key Risk Indicator (KRI) analysis for adverse events using the {gsm.kri} package. The examples show how to:

1. Generate an Adverse Event (AE) metric using the standard {gsm} workflow
2. Create a Serious Adverse Event (SAE) metric with filtering
3. Visualize metric distributions using bar charts
4. Create scatter plots with confidence bounds

Setup

Load required packages and data:

library(gsm.core)
library(gsm.mapping)
library(dplyr)
devtools::load_all()

# Load source data
dm <- gsm.core::lSource$Raw_SUBJ
ae <- gsm.core::lSource$Raw_AE

Example 1.1: Generate Adverse Event Metric

Generate an Adverse Event Metric using the standard {gsm} workflow:

dfInput <- Input_Rate(
  dfSubjects = dm,
  dfNumerator = ae,
  dfDenominator = dm,
  strSubjectCol = "subjid",
  strGroupCol = "invid",
  strNumeratorMethod = "Count",
  strDenominatorMethod = "Sum",
  strDenominatorCol = "timeonstudy"
)

dfTransformed <- Transform_Rate(dfInput)
dfAnalyzed <- Analyze_NormalApprox(dfTransformed, strType = "rate")
dfFlagged <- Flag(dfAnalyzed, vThreshold = c(-3, -2, 2, 3))
dfSummarized <- Summarize(dfFlagged)

Flag Distribution

table(dfSummarized$Flag)

## 
##   0   1   2 
## 136   8   1

Example 1.2: Serious Adverse Event Metric

Create an SAE Metric by adding a filter, using pipes for cleaner code:

SAE_KRI <- Input_Rate(
  dfSubjects = dm,
  dfNumerator = ae %>% filter(aeser == "Y"),
  dfDenominator = dm,
  strSubjectCol = "subjid",
  strGroupCol = "invid",
  strNumeratorMethod = "Count",
  strDenominatorMethod = "Sum",
  strDenominatorCol = "timeonstudy"
) %>%
  Transform_Rate() %>%
  Analyze_NormalApprox(strType = "rate") %>%
  Flag(vThreshold = c(-3, -2, 2, 3)) %>%
  Summarize()

SAE Flag Distribution

table(SAE_KRI$Flag)

## 
##   0   1   2 
## 138   5   2

Example 1.3: Visualize Metric Distribution

Create bar charts to visualize different aspects of the SAE metric:

labels <- list(
  Metric = "Serious Adverse Event Rate",
  Numerator = "Serious Adverse Events",
  Denominator = "Days on Study"
)

# Metric bar chart
gsm.kri::Widget_BarChart(dfResults = SAE_KRI, lMetric = labels, strOutcome = "Metric")

# Score bar chart
gsm.kri::Widget_BarChart(dfResults = SAE_KRI, lMetric = labels, strOutcome = "Score")

# Numerator bar chart
gsm.kri::Widget_BarChart(dfResults = SAE_KRI, lMetric = labels, strOutcome = "Numerator")

Example 1.4: Scatter Plot with Confidence Bounds

Create a scatter plot showing the relationship between metrics with confidence bounds:

dfBounds <- Analyze_NormalApprox_PredictBounds(SAE_KRI, vThreshold = c(-3, -2, 2, 3))
gsm.kri::Widget_ScatterPlot(SAE_KRI, lMetric = labels, dfBounds = dfBounds)

Summary

This report demonstrated the {gsm.kri} workflow for adverse event monitoring:

• Input: Prepared data for rate calculations
• Transform: Standardized metrics across sites
• Analyze: Applied statistical methods (Normal Approximation)
• Flag: Identified sites exceeding thresholds
• Summarize: Generated final KRI summaries
• Visualize: Created interactive charts and plots

The methodology enables systematic monitoring of adverse event rates across study sites to identify potential safety signals.

---

Generated with {gsm.kri} version 1.3.2