The IMPACT framework and implementation for accessible in silico clinical phenotyping in the digital era

Area of Science:

• Digital Health
• Computational Biology
• Bioinformatics

Background:

• Clinical phenotyping is crucial for digital health applications but is traditionally time-consuming and costly.
• Current in-silico phenotyping tools lack cross-task generalizability and are often inaccessible to users.
• There is a need for reusable and user-friendly in-silico phenotyping solutions.

Purpose of the Study:

• To identify barriers to in-silico phenotyping adoption and usage.
• To propose a framework of desiderata for effective in-silico phenotyping.
• To introduce a software application implementing this framework for improved clinical phenotyping algorithm development.

Main Methods:

• Literature review and analysis of existing in-silico phenotyping approaches.
• Development of a framework based on observed implementation challenges.
• Implementation of a software application prioritizing ease of adoption and reusability.

Main Results:

• Identification of key barriers hindering the widespread use of in-silico phenotyping.
• A proposed framework outlining essential features for generalizable and accessible tools.
• Demonstration of a software application that facilitates clinical phenotyping algorithm development.

Conclusions:

• A novel framework and software application can enhance the efficiency and accessibility of in-silico phenotyping.
• The developed solution supports cross-task reusability, addressing a significant gap in current tools.
• This work aims to accelerate the development of digital health applications by simplifying clinical data abstraction.