ClinicoPathDescriptives

ClinicoPathDescriptives is an open-source R package and jamovi module developed by Serdar Balci designed specifically for descriptive analysis, statistical reporting, data quality validation, and visualization in clinicopathological and biomedical research.

It bridges the divide between command-line statistical computing in R and point-and-click clinical analysis by providing a dual-interface architecture: full programmatic functions for R pipelines alongside interactive GUI modules within the jamovi desktop environment.

• GitHub Repository: sbalci/ClinicoPathDescriptives

• Documentation & Vignettes: serdarbalci.com/ClinicoPathDescriptives

• jamovi Module Hub: serdarbalci.com/ClinicoPathJamoviModule

• Main Project & Issues: sbalci/ClinicoPathJamoviModule

• Zenodo DOI: 10.5281/zenodo.3997188

• License: GPL-2.0+

---

Dual Interface Design

1. jamovi Graphical Interface: Point-and-click menus integrated under the Exploration ribbon, designed for clinicians, residents, and researchers seeking reproducible outputs without writing code.

2. R Programmatic Interface: Native tidy-friendly R functions accepting data frames, pipe operators (|>, %>%), and standard formula syntax, generating publication-ready tables and ggplot2-based graphics.

---

Analytical Modules & jamovi Menu Structure

In jamovi, the package expands the Exploration menu with five specialized clinicopathological sections:

Exploration
├── ClinicoPath Descriptives
│   ├── Table One (tableone)
│   ├── Summary of Continuous Variables (summarydata)
│   └── Summary of Categorical Variables (reportcat)
├── ClinicoPath Comparisons
│   ├── Cross Tables (crosstable)
│   └── Chi-Square Post-Hoc Tests (chisqposttest)
├── ClinicoPath Data Quality
│   ├── Benford's Law Analysis (benford)
│   ├── Single Variable Quality Check (checkdata)
│   ├── Multi-Variable Visual Quality (dataquality)
│   └── Outlier Detection (outlierdetection)
├── ClinicoPath Descriptive Plots
│   ├── Age Pyramid (agepyramid)
│   ├── Venn Diagram (venn)
│   ├── Variable Tree (vartree)
│   └── Alluvial Diagrams (alluvial)
└── ClinicoPath Data Preparation
    └── Categorize Continuous Variables (categorize)

1. Descriptive Reporting Suite (ClinicoPath Descriptives)

• Table One (tableone): Produces publication-grade baseline patient demographic and clinicopathologic characteristics tables. Supports four distinct journal-ready styling presets (NEJM, Lancet, JAMA formats), automatic variable type detection, non-parametric alternatives, and missingness accounting.

• Continuous Summaries (summarydata): Computes parametric (mean, SD) and non-parametric (median, IQR) summary statistics paired with automated natural language clinical interpretations and distributional diagnostics (skewness, kurtosis, Shapiro-Wilk normality testing).

• Categorical Reporting (reportcat): Generates detailed frequency tables, percentages, cumulative proportions, and missing value breakdowns formatted for clinical trial reports.

2. Contingency & Comparative Analysis (ClinicoPath Comparisons)

• Cross Tables (crosstable): Advanced $2 \times 2$ and $R \times C$ contingency tables offering Pearson's $\chi^2$, Fisher's exact test, continuity corrections, relative risk, odds ratios, and false discovery rate corrections ($q$-values) for multiple comparisons.

• Chi-Square Post-Hoc Tests (chisqposttest): Methodologically guardrailed post-hoc pairwise comparisons that trigger only when the global $\chi^2$ test achieves statistical significance ($p < \alpha$), preventing exploratory data dredging. Includes standardized adjusted residuals and Bonferroni / Benjamini-Hochberg adjustment options.

3. Data Quality & Pre-Analytical Auditing (ClinicoPath Data Quality)

• Benford's Law (benford): First-digit distribution analysis evaluating data integrity, fabrication, or digital rounding artifacts across continuous clinical measurements spanning multiple orders of magnitude.

• Single Variable Quality Check (checkdata): Comprehensive pre-flight screening for single variables. Computes completeness, missingness typology, consensus outlier flagging (requiring agreement across $\ge 2$ independent detection algorithms), repeated value inspection, and clinical plausibility checks, returning an aggregate heuristic quality grade.

• Multi-Variable Visual Quality (dataquality): Dataframe profiling including duplicate record detection, missing value heatmaps, and completeness summaries.

• Outlier Detection (outlierdetection): Integrates multivariate and univariate anomaly detection algorithms from performance:

  • Univariate: $Z$-scores, Interquartile Range (IQR) fences, confidence intervals.

  • Multivariate: Minimum Covariance Determinant (MCD), Mahalanobis distance, Local Outlier Factor (LOF), and OPTICS clustering.

4. Specialized Clinical Visualizations (ClinicoPath Descriptive Plots)

• Alluvial Diagrams (alluvial): Patient pathway tracking across stages, treatment lines, and outcome transitions using ggalluvial and easyalluvial.

• Venn & UpSet Plots (venn): Multi-set relationship visualization supporting 2 to 7 sets via ggVennDiagram, with optional transition to UpSetR / ComplexUpset layouts for high-dimensional set intersections.

• Age Pyramids (agepyramid): Demographic age-distribution pyramids stratified by biological sex or treatment arm.

• Variable Trees (vartree): Hierarchical decision and subset trees powered by vtree to illustrate patient flow, inclusion/exclusion criteria, and subgroup distributions.

5. Data Transformation (ClinicoPath Data Preparation)

• Categorize Continuous Variables (categorize): Discretizes continuous biomarkers into clinically meaningful categories via equal intervals, sample quantiles, manual thresholds, mean $\pm$ SD tiers, median split, or Fisher-Jenks natural breaks. Writes the resulting factor back to the active jamovi spreadsheet.

---

Quick Start (R Code Example)

# Install development version
devtools::install_github("sbalci/ClinicoPathDescriptives")

library(ClinicoPathDescriptives)

# Load included histopathology dataset
data("histopathology")

# 1. Generate Baseline Table One
tbl1 <- tableone(
  data = histopathology,
  grouping_variable = "Treatment_Group",
  explanatory_variables = c("Age", "Gender", "Tumor_Size", "Grade")
)

# 2. Cross-tabulation with Statistical Tests
cross_res <- crosstable(
  data = histopathology,
  dependent_variable = "Response",
  explanatory_variables = c("Treatment_Group", "Biomarker_Status"),
  statistical_test = TRUE
)

# 3. Patient Journey Alluvial Plot
alluv_plot <- alluvial(
  data = histopathology,
  variables = c("Stage", "First_Line_Therapy", "Response", "Recurrence")
)

---

Place in the ClinicoPath Ecosystem

ClinicoPath began as a monolithic jamovi module (ClinicoPathJamoviModule) and was later modularized into targeted packages:

• ClinicoPathDescriptives: Descriptive statistics, Table 1, contingency tables, data auditing, and exploratory visualizations.

• OncoPath: Specialized oncology timeline plots, including swimmer plots (individual patient treatment timelines) and waterfall plots (RECIST tumor burden changes).

• Core ClinicoPath: Survival analysis (Kaplan-Meier, Cox regression), diagnostic agreement (Kappa), ROC analysis, decision curves, and decision trees.

---

Installation

In jamovi

1. Open jamovi (version $\ge 2.7.27$).

2. Click the ⊕ (Modules) button in the top-right corner.

3. Select jamovi library, search for ClinicoPathDescriptives, and click Install.

In R

if (!requireNamespace("devtools", quietly = TRUE)) install.packages("devtools")
devtools::install_github("sbalci/ClinicoPathDescriptives")