IPDfromKM: reconstruct individual patient data from published Kaplan-Meier survival curves

Area of Science:

• Biostatistics
• Medical Informatics
• Survival Analysis

Background:

• Secondary analysis of published survival data is crucial for medical research.
• Access to individual patient data (IPD) is the gold standard but often unavailable.
• Existing methods for IPD reconstruction from survival curves can be complex.

Purpose of the Study:

• To propose a straightforward and robust method for reconstructing IPD from published survival curves.
• To develop a user-friendly software platform to facilitate IPD reconstruction.
• To enhance the utility of secondary data analysis in medical research.

Main Methods:

• A two-stage approach was developed: Stage 1 extracts raw data coordinates from Kaplan-Meier (K-M) curves, and Stage 2 reconstructs IPD.
• An R package (IPDfromKM) and a web-based Shiny application were created.
• The software provides an all-in-one solution for data extraction, IPD reconstruction, visualization, accuracy assessment, and secondary analysis.

Main Results:

• The IPDfromKM R package and Shiny application accurately reconstruct IPD from published K-M curves.
• The method demonstrates high accuracy and reliability in estimating key survival metrics like events, patients at risk, survival probabilities, median survival times, and hazard ratios.
• Simulations and real-world data applications validate the proposed approach's effectiveness.

Conclusions:

• The IPDfromKM tool offers flexibility and accuracy for reconstructing IPD from various K-M curve shapes.
• This development is expected to significantly increase the use of quality IPD for secondary data analysis.
• The tool will advance informed decision-making in medical research by leveraging published survival data.