RTiS Concept Mapping

Context

The Roche Terminology and Interoperability System (RTiS) is the authoritative source for standard terminology concepts used in Roche data assets. Mapping local business terms or data attributes to RTiS (or to underlying ontologies like SNOMED, LOINC, or ICD) is a prerequisite for interoperable data contracts and Collibra asset registration. This prompt prepares a structured mapping proposal for RTiS review, and can also explain a specific standard concept in plain language.

The AI does not have direct access to RTiS. These prompts help you prepare proposals and questions for RTiS review — they are not a substitute for validation against the actual system.

The Prompt

Mapping terms to standard concepts:

I have a list of business terms or data attributes that need to be mapped to standard
terminology concepts. Help me prepare a mapping proposal.

For each term, suggest:
- A candidate standard concept name (using standard ontology language where applicable)
- The likely ontology or terminology system it belongs to (SNOMED, LOINC, ICD, internal Roche terminology, etc.)
- Confidence level (High / Medium / Low — based on how specific or ambiguous the term is)
- Notes on ambiguity or alternative mappings to investigate

Terms to map:
[LIST TERMS — one per line, with any available context or definition]

Domain context: [e.g., Clinical / Product / Customer / Reference Data / ...]

Explaining a terminology concept:

Explain the concept "[CONCEPT NAME]" as it would be defined in [RTiS / SNOMED / LOINC /
relevant ontology]. Include:
- Definition
- How it differs from closely related concepts (if any)
- Typical use cases in [DATA DOMAIN]
- Any known ambiguities or common misuses

Usage Instructions

Open build-cli in your terminal.
Choose the appropriate prompt form above (mapping or explanation).
For mapping: list each term on its own line with as much context as you have (definition, usage examples, source system field name).
Provide the domain context — this significantly affects which ontology system is most relevant.
Use the AI output as input for your RTiS contact or data steward — not as the final mapping.

Example Output

For a mapping request, the AI will produce a table such as:

Term	Candidate concept	Ontology	Confidence	Notes
`patient_sex`	Administrative Sex	SNOMED CT (385442003)	High	Standard administrative sex concept; distinct from biological sex
`diagnosis_code`	Clinical Finding	SNOMED CT / ICD-10	Medium	Ambiguous — could be ICD-10 or SNOMED depending on source system; confirm with RTiS contact
`product_formulation`	Pharmaceutical Dose Form	EDQM / SNOMED	Low	Multiple ontologies cover this space; need domain context to narrow down

Notes:

For High confidence mappings: validate directly in RTiS before using in a data contract or Collibra asset.
For Medium/Low confidence: use as input for a conversation with your RTiS contact or data steward.