The Problem with Heuristic Domain Assignment

Oct 13, 2025 by talynt.org 97 views

Introduction

Domain classification is a crucial aspect of healthcare data management, and the OMOP Common Data Model (CDM) relies heavily on it. However, the current approach to domain assignment in OMOP is based on a heuristic concept domain assignment model, which can lead to arbitrary decisions and circular reasoning. In this article, we will explore the limitations of the current approach and propose an alternative extensible cascading ruleset model.

The Problem with Heuristic Domain Assignment

The Book of OHDSI suggests that domain assignment should be heuristic, and even refers back to the vocabularies code base for guidance. This approach creates issues on design level, enabling circular reasoning and arbitrary decisions. For instance, should scar tissue formations be classified as Conditions or as Observations? The only source of truth for answering this is the heuristic laid out in the code, and up to implementers to change at will.

The Difficulty of Domain Assignment

Domain assignment is a very difficult task. Most times, a diagnosis is made, and it is unambiguously a Condition. When a substance is injected, it is likely a Drug. However, there always will be ambiguity around symptoms and syndromes vs. diseases, biological substances vs. pharmaceutical drugs, and measurement vs. observations, and many other edge cases. Having at least somewhat formalized Domain definition page at the Vocabulary wiki would lay down important groundwork.

Precedent Systems for Edge Cases

As a reference example, SNOMED has a system of short template pages, defining how new concepts in particular domains should be modelled. We could have the same hierarchy of templates for domain assignment in edge cases. There is nothing wrong with subjective reasoning (we are operating with made-up high-level abstractions over physical processes and ephemeral emerging properties of biological systems), nor with justifying decisions by precedent. However, as the circle of Vocabulary authors will grow, documentation system is required for consistency and predictability.

The Benefits of an Extensible Cascading Ruleset Model

An extensible cascading ruleset model would provide a more formalized and structured approach to domain assignment. This would enable:

Consistency: A clear and well-documented set of rules would ensure consistency in domain assignment across different implementers and use cases.
Predictability: With a well-defined set of rules, implementers would be able to predict the outcome of domain assignment for a given concept.
Flexibility: An extensible ruleset would allow for easy addition of new rules and domains, making it easier to adapt to changing healthcare data management needs.

Implementing an Extensible Cascading Ruleset Model

SNOMED's multiaxial hierarchy provides a clean and structured approach to domain assignment. This hierarchy can be easily implemented in OMOP, with an arbitrary number of rule definitions in the OMOP SNOMED representation. Other vocabularies can also be mapped to SNOMED as Standard, or divide codes semantically into ranges, which can serve as a sorting mechanism for domains.

Conclusion

The current approach to domain assignment in OMOP is based on a heuristic concept domain assignment model, which can lead to arbitrary decisions and circular reasoning. An extensible cascading ruleset model would provide a more formalized and structured approach to domain assignment, enabling consistency, predictability, and flexibility. By adopting this approach, we can move away from the limitations of the current heuristic model and provide a more robust and reliable framework for healthcare data management.

Recommendations

Develop a clear and well-documented set of rules for domain assignment, based on the extensible cascading ruleset model.
Implement a multiaxial hierarchy in OMOP, similar to SNOMED's, to provide a clean and structured approach to domain assignment.
Map other vocabularies to SNOMED as Standard, or divide codes semantically into ranges, to provide a sorting mechanism for domains.
Provide a documentation system for consistency and predictability, as the circle of Vocabulary authors will grow.

Q: What is the current approach to domain assignment in OMOP?

A: The current approach to domain assignment in OMOP is based on a heuristic concept domain assignment model, which can lead to arbitrary decisions and circular reasoning.

Q: What are the limitations of the current approach?

A: The current approach creates issues on design level, enabling circular reasoning and arbitrary decisions. For instance, should scar tissue formations be classified as Conditions or as Observations? The only source of truth for answering this is the heuristic laid out in the code, and up to implementers to change at will.

Q: What is an extensible cascading ruleset model?

A: An extensible cascading ruleset model is a more formalized and structured approach to domain assignment. This would enable consistency, predictability, and flexibility in domain assignment.

Q: What are the benefits of an extensible cascading ruleset model?

A: An extensible cascading ruleset model would provide:

Consistency: A clear and well-documented set of rules would ensure consistency in domain assignment across different implementers and use cases.
Predictability: With a well-defined set of rules, implementers would be able to predict the outcome of domain assignment for a given concept.
Flexibility: An extensible ruleset would allow for easy addition of new rules and domains, making it easier to adapt to changing healthcare data management needs.

Q: How can an extensible cascading ruleset model be implemented?

A: SNOMED's multiaxial hierarchy provides a clean and structured approach to domain assignment. This hierarchy can be easily implemented in OMOP, with an arbitrary number of rule definitions in the OMOP SNOMED representation. Other vocabularies can also be mapped to SNOMED as Standard, or divide codes semantically into ranges, which can serve as a sorting mechanism for domains.

Q: What are the recommendations for implementing an extensible cascading ruleset model?

A: The recommendations for implementing an extensible cascading ruleset model are:

Develop a clear and well-documented set of rules for domain assignment, based on the extensible cascading ruleset model.
Implement a multiaxial hierarchy in OMOP, similar to SNOMED's, to provide a clean and structured approach to domain assignment.
Map other vocabularies to SNOMED as Standard, or divide codes semantically into ranges, to provide a sorting mechanism for domains.
Provide a documentation system for consistency and predictability, as the circle of Vocabulary authors will grow.

Q: What are the next steps for implementing an extensible cascading ruleset model?

A: The next steps for implementing an extensible cascading ruleset model are:

Develop a clear and well-documented set of rules for domain assignment, based on the extensible cascading ruleset model.
Implement a multiaxial hierarchy in OMOP, similar to SNOMED's, to provide a clean and structured approach to domain assignment.
Map other vocabularies to SNOMED as Standard, or divide codes semantically into ranges, to provide a sorting mechanism for domains.
Provide a documentation system for consistency and predictability, as the circle of Vocabulary authors will grow.

Q: What are the benefits of moving away from the current heuristic model?

A: The benefits of moving away from the current heuristic model are:

Improved consistency in domain assignment across different implementers and use cases.
Improved predictability in domain assignment for a given concept.
Improved flexibility in adapting to changing healthcare data management needs.

Q: What are the potential challenges of implementing an extensible cascading ruleset model?

A: The potential challenges of implementing an extensible cascading ruleset model are:

Developing a clear and well-documented set of rules for domain assignment.
Implementing a multiaxial hierarchy in OMOP, similar to SNOMED's.
Mapping other vocabularies to SNOMED as Standard, or dividing codes semantically into ranges.
Providing a documentation system for consistency and predictability.

Q: What are the potential benefits of implementing an extensible cascading ruleset model?

A: The potential benefits of implementing an extensible cascading ruleset model are:

Improved consistency in domain assignment across different implementers and use cases.
Improved predictability in domain assignment for a given concept.
Improved flexibility in adapting to changing healthcare data management needs.