Quality metrics for detailed clinical models

Area of Science:

• Medical Informatics
• Software Engineering
• Health Data Standards

Background:

• Detailed Clinical Models (DCMs) are crucial for structured health data.
• Existing quality criteria for DCMs lacked formal, quantifiable metrics.
• Standardized quality assessment is needed for reliable DCM development and use.

Purpose of the Study:

• To develop and validate formal quality metrics for Detailed Clinical Models (DCMs).
• To establish a reliable framework for assessing DCM quality.
• To support rational decision-making in DCM development and utilization.

Main Methods:

• Applied the ISO/IEC 9126 software quality model to develop initial metrics.
• Conducted a two-round Delphi survey with 9 international experts for face and content validity assessment.
• Assessed reliability using the kappa coefficient (agreement > 0.60) between two evaluators on example DCMs.

Main Results:

• Achieved high inter-rater reliability with a kappa coefficient of 0.73.
• Finalized 8 quality evaluation domains and 29 distinct quality metrics for DCMs.
• Successfully validated the developed quality metrics through expert consensus.

Conclusions:

• Validated quality metrics for DCMs were established by international experts.
• These metrics define essential qualitative and quantitative requirements for DCMs.
• The metrics are expected to facilitate informed decision-making for DCM developers and clinical users.