Development and evaluation of methods for structured recording of heart murmur findings using SNOMED-CT post-coordination

Abstract

Objective: This study evaluated an existing SNOMED-CT model for structured recording of heart murmur findings and compared it to a concept-dependent attributes model using content from SNOMED-CT.

Methods: The authors developed a model for recording heart murmur findings as an alternative to SNOMED-CT's use of Interprets and Has interpretation. A micro-nomenclature was then created to support each model using subset and extension mechanisms described for SNOMED-CT. Each micro-nomenclature included a partonomy of cardiac cycle timing values. A mechanism for handling ranges of values was also devised. One hundred clinical heart murmurs were recorded using purpose-built recording software based on both models.

Results: Each micro-nomenclature was extended through the addition of the same list of concepts. SNOMED role grouping was required in both models. All 100 clinical murmurs were described using each model. The only major differences between the two models were the number of relationship rows required for storage and the hierarchical assignments of concepts within the micro-nomenclatures.

Conclusion: The authors were able to capture 100 clinical heart murmurs with both models. Requirements for implementing the two models were virtually identical. In fact, data stored using these models could be easily interconverted. There is no apparent penalty for implementing either approach.

Figure 1.

Heart murmur findings in SNOMED-CT (7-31-2003). SNOMED-CT contains a subhierarchy of murmur findings that includes descriptive murmur findings containing references only to physical description of the murmur as well as murmur findings containing references to the etiology of the murmur. These two very different types of murmur finding concepts are intermixed within the subhierarchy. Of the descriptive murmur finding present, none have more than two physical descriptors in their fully specified name.

Figure 2.

Concept-dependent Attributes model representation. The model uses concept-role-value triples to represent each murmur characteristic. Triples (designated within parentheses) are repeated for each murmur characteristic needing to be described. The finding “Heart murmur (88610006)” serves as the concept. The cardiac murmur attributes and cardiac murmur values created for use in the model serve as the role and value, respectively. To represent an instance of a specific characteristic, the appropriate role and value are chosen from the lists of attributes and values created.

Figure 3.

Interprets/Has interpretation model representation. The model uses concept-role-value triplets that are grouped to represent each murmur characteristic. Groups of triplets (designated within parentheses) are repeated for each characteristic described. The finding “Heart murmur (88610006)” serves as the concept for each triplet. The cardiac murmur characteristics (as observable entities) and cardiac murmur values created for use in the model serve as the values for the roles “Interprets (363714003)” and “Has interpretation (363713009),” respectively. To represent an instance of a specific characteristic, the appropriate values are chosen from the lists of characteristics and values created.

Figure 4.

Initial subsets. Subsets of SNOMED-CT derived for each model. For our purposes, subsets include those concepts that exist in SNOMED-CT and that were either necessary to instantiate murmurs using our models or to organize the micro-nomenclatures.

Figure 5.

Murmur characteristics add to each extension. Murmur characteristics (highlighted in yellow) were added to the nomenclatures of each model. For the Concept-dependent Attributes model, they were added as attributes. For the Interprets/Has interpretation model, they were added as observable entities.

Figure 6.

Qualifier values added to both extensions. This hierarchical display shows the qualifier values that were added to both subsets. Murmur radiation values were too numerous to display here.

Figure 7.

Cardiac cycle timing partonomy. Hierarchical (Is-a) relationships are represented by solid arrows. Part-of relationships are represented by a dashed arrow.

Figure 8.

Mechanism for handling ranges. The model for range values takes the base models and adds triplets for each delimiter needed to specify the range or spatial area. All triplets are then grouped (designated by parentheses).

Figure 9.

Entity relationships diagram for record storage. This database schema was used to capture data from both models. Murmur data were held in the case and murmurs tables. The concepts, relationships, and descriptions tables contained the concepts from the extensions created for each model. Conceptids from the concepts table were used to populate attributes and values in the murmurs table.

Figure 10.

Relationship between the models for this study. Concepts containing the same information are connected with dotted lines.