top of page

Automated Survival Analysis Platform

                                                                                             

Overview

​

A web-based Shiny application for comprehensive survival analysis with enhanced data detection, reference level control, and automated reporting capabilities. Designed to support researchers and clinicians in conducting rigorous survival analyses using generated test datasets.

Key Features

  • Test Data Generation: Multiple realistic datasets (Clinical Trial, Oncology Study, Generic Survival Data) with comprehensive covariates

  • Intelligent Data Detection: Automatically recognizes various column naming patterns including cardiovascular-specific terms (time: survival_time, os_months, cardiac_months, cv_time; status: death, event, cv_event, cardiac_death)

  • Smart Data Conversion: Handles multiple formats (Yes/No, Dead/Alive, 0/1, Event/Censored) with automatic format recognition and conversion

  • Reference Level Control: Set meaningful baseline categories for hazard ratio interpretation with clinical relevance

  • Complete Analysis Suite: Kaplan-Meier plots, median survival times with confidence intervals, log-rank tests, pairwise comparisons, Cox regression, proportional hazards testing

  • Professional Reporting: Export comprehensive Word documents with formatted tables, statistical interpretations, and methodology notes

 

Supported Data Formats

  • Time Columns: time, survival_time, os_months, duration, followup_time, cardiac_months, cv_time, heart_disease_years, mi_time

  • Status Columns: status, event, death, censored, cardiac_death, cv_event, outcome, vital_status, cardiac_event, cardiovascular_death

​

User's Guide

​

The application is ready to use immediately - no installation or setup required.

Step-by-Step Instructions

  1. Generate Test Data: Select from Clinical Trial (300 patients), Oncology Study (250 patients), or Generic Survival Data (150 patients)

  2. Automatic Detection: App identifies time/status columns and converts formats (with user notifications)

  3. Variable Selection: Choose primary variable for Kaplan-Meier plots and select adjustment variables for Cox regression

  4. Reference Levels: Set clinically meaningful reference categories for categorical variables

  5. Analysis Review: Examine survival curves, median survival times, Cox regression results, and proportional hazards tests

  6. Export Reports: Generate comprehensive Word documents with all analyses, tables, and interpretations

Analysis Tabs

  • Kaplan-Meier Plot: Interactive survival curves with risk tables and p-values

  • Survival Summary: Median survival times, log-rank tests, and pairwise comparisons

  • Cox Regression: Multivariable analysis with hazard ratios and confidence intervals

  • Proportional Hazards Test: Assumption testing with Schoenfeld residuals and diagnostic plots

  • Data View: Review processed dataset with applied transformations

Example Use Cases

  • Clinical trial efficacy analysis

  • Cardiovascular outcomes research

  • Cancer survival studies

  • Epidemiological cohort analysis

  • Quality improvement projects

  • Time-to-event analysis in any medical specialty

Technical Specifications

Performance Notes

  • Generated datasets optimized for immediate analysis

  • Cox regression supports up to 50 covariates

  • Word export handles reports up to 20 pages efficiently

  • Real-time data processing with immediate feedback

Data Privacy

  • All datasets are synthetically generated for testing purposes

  • Data processing occurs in-memory only (no persistent storage)

  • No data retention after session ends

  • Session-based analysis with automatic cleanup

Technical Support

  • Detailed error messages provided in app notifications

  • Comprehensive documentation included in the platform

  • Contact: kadirooj@gmail.com for technical issues

Compliance and Limitations

  • Educational and Testing Use: Designed for learning and testing survival analysis methods

  • Statistical Validity: Results include appropriate statistical warnings and interpretation guidelines

  • Model Assumptions: Proportional hazards assumption testing included for Cox regression validity

  • Synthetic Data: All generated datasets are for demonstration purposes only

  • Professional Oversight: Real clinical analyses should be reviewed by qualified statisticians

bottom of page