A new methodological framework for hazard detection models in health information technology systems.

Area of Science:

• Health Informatics
• Patient Safety
• Health Information Technology (HIT)

Background:

• Health information technology (HIT) adoption increases healthcare quality and efficiency but introduces patient safety risks.
• Electronic health records (EHRs) generate vast digital data, necessitating robust hazard detection tools.
• Existing systems lack structured methods for identifying and mitigating HIT-related patient safety issues.

Purpose of the Study:

• To present a novel methodological framework for developing hazard detection models for HIT systems.
• To provide structure for research and communication among interdisciplinary stakeholders.
• To reduce communication barriers and standardize hazard detection processes.

Main Methods:

• A nine-stage framework was developed, encompassing feature extraction, detector development, and optimization.
• The framework includes a support environment for evaluating detector models.
• The US Department of Veterans Affairs' (VA) Corporate Data Warehouse was used for demonstration.

Main Results:

• The framework provides a structured approach to developing hazard detection tools.
• It addresses risks from data errors (collection, transmission, use, processing) and system errors (programming, configuration).
• The methodology is adaptable to specific HIT systems.

Conclusions:

• The proposed framework is foundational for creating effective hazard detection tools in EHRs.
• It enhances the ability to identify and mitigate patient safety risks associated with HIT.
• This structured approach supports interdisciplinary collaboration in patient safety research.