Clinical Decision Support (CDS)

Key facts

Certification criterion

Decision support interventions, 45 CFR 170.315(b)(11)

FDA device carve-out

FD&C Act Section 520(o)(1)(E), added by the 21st Century Cures Act

FDA guidance

Clinical Decision Support Software guidance, first finalized in 2022

EHR integration standard

HL7 CDS Hooks

What is clinical decision support?

CDS puts relevant knowledge in front of a decision maker when it is useful. The Office of the National Coordinator for Health IT, now ASTP/ONC, describes it as providing clinicians, staff, patients or others with knowledge and person-specific information, intelligently filtered or presented at appropriate times, to enhance health and health care.

Most CDS runs inside EHRs, but it also appears in pharmacy systems, imaging systems, patient apps and standalone software. It ranges from simple rules, such as an allergy alert, to machine learning models that predict deterioration or readmission.

Medicare once required clinicians ordering advanced imaging to consult appropriate use criteria through qualified CDS tools, under a 2014 law. CMS paused that program and rescinded its regulations in the 2024 Physician Fee Schedule final rule.

Types of clinical decision support

Common forms of CDS include:

• Alerts and reminders: drug interaction, allergy, duplicate therapy and overdue screening notices.
• Order sets and care protocols: standard groups of orders for a condition or procedure.
• Documentation templates: structured prompts that capture required data.
• Dosing support: calculators that adjust doses for weight or kidney function.
• Reference information: context-sensitive links to guidelines or drug labels.
• Predictive decision support: models that estimate risk, such as sepsis or readmission scores.

Why clinical decision support matters

CDS sits where regulation, product design and prescribing meet:

• Health IT and AI developers: under the HTI-1 final rule, certified health IT must support source attribute transparency for predictive decision support interventions, which affects how AI models are deployed in certified EHRs.
• Regulatory strategy: CDS software that meets the Cures Act criteria is excluded from the FDA device definition, while other CDS, such as tools that analyze medical images or signals, can be regulated as a device. FDA has revisited its interpretation, so check the current guidance.
• Pharma and medical affairs: order sets and alerts shape prescribing, so guideline and formulary rules built into CDS affect product uptake.
• Patient safety: poorly designed alerts cause alert fatigue, so hospitals review CDS performance as part of safety programs.