A usability evaluation of a SNOMED CT based compositional interface terminology for intensive care

Area of Science:

• Medical Informatics
• Clinical Data Management
• Health Information Technology

Background:

• Accurate registration of reasons for intensive care admission is vital for patient data management.
• Large compositional terminologies, such as SNOMED CT, offer potential for standardized clinical data but pose usability challenges.
• Evaluating the usability of such terminologies in real-world clinical settings is essential for optimizing their adoption and effectiveness.

Purpose of the Study:

• To evaluate the usability of a SNOMED CT-based compositional interface terminology and its application for registering intensive care admission reasons.
• To identify specific usability problems encountered by clinicians during routine use of the system.
• To assess changes in usability over time after system implementation.

Main Methods:

• An observational study design was employed, with usability evaluations conducted before and three months after system implementation.
• User-based usability was assessed using qualitative methods (Think-Aloud user testing) and quantitative methods (System Usability Scale questionnaire, Time-on-Task analyses).
• Usability was measured across five aspects: effectiveness, efficiency, learnability, overall user satisfaction, and experienced usability problems.

Main Results:

• The interface terminology exhibited low usability, with System Usability Scale scores of 47.2 before and 37.5 after implementation.
• A significant number of usability problems (n=35) were identified, impacting the effectiveness and efficiency of data registration.
• Most problems (86%) stemmed from the terminology application, while 14% were related to SNOMED CT content, with content-related issues being more severe.

Conclusions:

• The usability of a large compositional interface terminology is significantly influenced by its extensiveness, hierarchical complexity, and language usage.
• Clinician interaction with controlled terminologies highlights the need for user-centered design in terminology applications.
• Developing carefully curated, domain-specific subsets and optimizing the terminology application are critical for enhancing the usability of complex SNOMED CT-based terminologies.