Data structuring may prevent ambiguity and improve personalized medical prognosis

Area of Science:

• Biomedical informatics
• Personalized medicine
• Data science in healthcare

Background:

• Personalized medicine (PM) aims to tailor treatments to individual patients.
• Ambiguity, arising from biological variations and data interpretation challenges, is a significant barrier to PM.
• Current medical data practices often lack sufficient biomedical validation and interpretation.

Purpose of the Study:

• To review factors influencing personalized medicine.
• To identify causes of ambiguity in medical data and propose mitigation strategies.
• To explore non-reductionist approaches and data structuring for improved PM.

Main Methods:

• Analysis of factors contributing to ambiguity in medical data.
• Review of delayed innovation adoption, inadequate emphasis on interpretation, and cross-disciplinary tool validation.
• Exploration of non-reductionist alternatives and data structuring techniques.
• Application of a data-information-knowledge-decision-making process using COVID-19 data.

Main Results:

• Identified three key causes of ambiguity: delayed innovation, misplaced emphasis, and inadequate validation.
• Highlighted issues with compositional data (e.g., leukocyte data) and reliance on non-biomedical tools.
• Demonstrated the potential of data structuring and non-reductionist approaches using COVID-19 data.
• Showcased how specific data criteria can inform personalized decisions with fewer observations.

Conclusions:

• Ambiguity in medical data can be prevented or reduced through improved data structuring and validation.
• Non-reductionist methods and a focus on biomedical interpretation are crucial for advancing personalized medicine.
• Validated data criteria, including distinct patterns and interpretable intervals, can enhance early and efficient personalized patient care.