Incompleteness of Electronic Health Records: An Impending Process Problem Within Healthcare.

Area of Science:

• Health Informatics
• Data Science
• Health Systems Research

Background:

• Digitization of health records aimed to enhance data quality and accessibility.
• Incompleteness in electronic health records (EHRs) remains a significant challenge, affecting clinical care, interoperability, and analytics.
• Missing data in EHRs is driven by systemic process gaps, not solely technical limitations.

Purpose of the Study:

• To synthesize evidence on EHR incompleteness drivers and consequences.
• To examine how data missingness propagates bias in artificial intelligence (AI) and machine learning (ML) systems.
• To propose a pragmatic agenda for mitigating EHR incompleteness.

Main Methods:

• Conceptual integration of incompleteness foundations.
• Synthesis of cross-country evidence on EHR data quality.
• Examination of process-level drivers and consequences of missing data.
• Development of a unifying taxonomy for data incompleteness.

Main Results:

• EHR incompleteness stems from multifaceted process gaps involving patients, providers, technology, and policy.
• Missing data in EHRs can introduce and propagate bias in AI/ML models.
• Complementary approaches like Record Strength Score, distributional testing, and workflow studies can address incompleteness.

Conclusions:

• EHR incompleteness can be systematically mitigated, though not entirely eliminated.
• Key mitigation strategies include implementing standards, redesigning workflows, engaging patients, and strengthening governance.
• These steps are crucial for developing safe, equitable, and effective learning health systems.