A Framework for Classification of Electronic Health Data Extraction-Transformation-Loading Challenges in Data Network...

Area of Science:

• Health Informatics
• Data Management
• Clinical Research

Background:

• Contributing health data to networks requires Extract, Transform, Load (ETL) processes to convert local data into a Common Data Model (CDM).
• These ETL processes demand substantial investment in technical resources, specialized skills, and programming expertise.
• A classification of ETL challenges can identify resource needs and lower barriers for less-technical data partners.

Purpose of the Study:

• To survey and classify the Extract, Transform, Load (ETL) challenges faced by data partners in clinical data research networks (CDRNs) and registries.
• To develop a framework for categorizing these challenges to aid in resource allocation and network participation.

Main Methods:

• Key-informant interviews with data partner representatives were conducted to identify Extract, Transform, Load (ETL) challenges.
• A workshop with diverse network stakeholders, including data partners, researchers, and policy experts, was held to vet the identified challenges.
• A five-point Likert scale was used by participants to rate the importance of each challenge.

Main Results:

• Twenty-four technical Extract, Transform, Load (ETL) challenges were identified and rated as important or very important by participants.
• A framework was developed to categorize these ETL challenges based on ETL phases, themes, and levels of data network participation.

Conclusions:

• Addressing Extract, Transform, Load (ETL) technical challenges necessitates significant investment in information technology and human resources.
• Identifying and categorizing these challenges can optimize resource allocation, reduce entry barriers for new data partners, and enhance network inclusivity.
• This framework provides valuable guidance for data partners navigating the complexities of contributing data to networks.