Which statement correctly describes the four categories of CDSS types and their deployment?

• A. Knowledge-based CDSS rely on machine learning; non-knowledge-based CDSS rely on data-driven models.
• B. Non-knowledge-based CDSS are embedded in the EHR; integrated CDSS are external apps.
• C. Integrated CDSS are embedded in the EHR and triggered during ordering or documentation; standalone CDSS are separate applications used manually; knowledge-based CDSS rely on explicit rules; non-knowledge-based CDSS rely on data-driven models.
• D. All CDSS types are interchangeable and deployment does not matter.

Answer: (C)

The key idea is that CDSS types are defined by two dimensions: how they generate advice (rule-based explicit knowledge versus data-driven models) and where they operate (embedded inside the EHR workflow versus as a separate, standalone application).

Knowledge-based CDSS relies on explicit clinical rules and guidelines that are encoded for decision support. These rules are interpretable because they are based on predefined criteria set by experts. A knowledge-based system can be integrated into the EHR so it automatically applies those rules during workflows, or it can exist as a separate tool used outside the main clinical workflow.

Non-knowledge-based CDSS, on the other hand, uses data-driven models—often machine learning—that learn from large datasets rather than relying on explicit rules. These models can be embedded in the EHR, triggering during ordering or documentation, or they can be accessed as standalone applications used manually.

Integrated CDSS are those built into the EHR and designed to fire within the clinician’s normal workflow, such as during ordering or documenting. Standalone CDSS are separate applications that clinicians interact with outside the EHR workflow.

This combination explains the four categories: knowledge-based integrated, knowledge-based standalone, non-knowledge-based integrated, and non-knowledge-based standalone. The statement in question accurately captures both the nature of how the advice is produced (rules versus data-driven models) and how deployment occurs (embedded in the EHR and workflow versus separate, manually used tools). The other choices misstate either what knowledge-based means, how deployment works, or claim interchangeability that doesn’t reflect these distinctions.