Real world data and data science in medical research: present and future

Area of Science:

• Medical Informatics
• Biostatistics
• Public Health

Background:

• Real-world data (RWD) analytics are increasingly vital in medical research, encompassing treatment effectiveness, safety, epidemiology, and cost analysis.
• RWD research integrates digital transformation, data science, public health, and regulatory science, yet often articles cover only partial aspects, leading to incomplete understanding.

Purpose of the Study:

• To provide a comprehensive overview of real-world data (RWD) analysis in medical fields from a methodological standpoint.
• To highlight the challenges and future directions in leveraging RWD for medical research and regulatory purposes.

Main Methods:

• Understanding the data sources of RWD, particularly digitized medical administrative data and electronic health records.
• Selecting appropriate statistical and epidemiological methods to control for biases inherent in RWD, which differ from randomized clinical trials.
• Exploring strategies for future RWD analysis, including proxy confounders, unstructured text data, and multi-database linkage.

Main Results:

• The digitization of medical data is a key enabler for RWD progress.
• Appropriate methodological choices are critical for RWD analysis due to potential biases.
• Future RWD applications include regulatory use and value assessment in medical research.

Conclusions:

• A holistic methodological approach is essential for robust RWD analysis.
• Addressing data limitations and biases is key to unlocking the full potential of RWD and real-world evidence (RWE).
• The integration of RWD holds significant promise for advancing medical research and informing regulatory decisions.