Empirical derivation of an electronic clinically useful problem statement system

Area of Science:

• Medical Informatics
• Clinical Terminology
• Health Data Management

Background:

• Problem lists are crucial for patient management, linking diagnoses to treatment and psychosocial factors.
• Computerized problem lists offer advantages but require clinically expressive vocabularies.
• Existing controlled vocabularies lack the necessary clinical detail for problem statements.

Purpose of the Study:

• To derive a new, clinically expressive canonical problem statement vocabulary.
• To improve the representation of clinical concepts in electronic health records.
• To create a standardized resource for clinical practice.

Main Methods:

• Semi-automated analysis and distillation of 891,770 provider-entered problem statements from 74,696 patients over 6 years.
• Combined automated and manual methods to create the Canonical Clinical Problem Statement System (CCPSS).
• Characterized problem statement nature, frequency, interrelations, and epidemiology.

Main Results:

• Developed the CCPSS with 15,534 terms, identifying 23,503 problem relations and 22,690 modifiers.
• Compared CCPSS to the Unified Medical Language System (UMLS) Metathesaurus; UMLS captured only 25% of CCPSS terms exactly.
• Clinicians confirmed CCPSS captured clinical intent for 84% of free-text problem statements.

Conclusions:

• The CCPSS represents clinical concepts at a higher level than previous approaches.
• This new vocabulary significantly enhances the clinical expressiveness of problem lists.
• A national adoption of a similar approach could standardize clinical problem representation.