OrthoPath

Validation

Last updated: 2026-02-24

Assistive-Only Disclaimer

OrthoPath is an assistive clinical decision support tool only. Validation results reported here cover technical correctness of rule logic only — not clinical efficacy or patient outcomes.

Do not enter patient identifiers or any Protected Health Information (PHI) into this tool.

Validation Approach

OrthoPath uses a multi-layer technical validation strategy:

Rule Logic Validation

• Golden tests: Each protocol pack includes a set of golden test cases (structured input → expected output). These run automatically at CI time.
• Completeness tests: For each protocol, test suites verify that all critical input combinations produce a defined (non-empty, non-silent) output.
• Determinism tests: Tests confirm that identical inputs always produce identical outputs across repeated runs.
• Condition parser tests: The grammar-based rule condition parser is unit-tested for correct AST construction and edge cases.

Safety Language Validation

• Automated gate (`protocol_verify`): Before any protocol is published, an automated check scans all rule output text for banned terms (diagnose, prescribe, order, etc.) using pattern matching with negative-context exclusion.

API Contract Validation

• OpenAPI 3.1 assertion: CI asserts that the API generates a valid OpenAPI 3.1.0 specification.
• Schema strictness: All data models use `extra="forbid"` — any unexpected field causes an immediate validation error.

What Has Not Been Validated

• Clinical efficacy: No prospective clinical trial or observational study of OrthoPath's impact on patient outcomes has been conducted.
• Human factors / usability: No formal usability testing with target end-users has been performed.
• Regulatory clearance: This tool is not cleared as a medical device by any regulatory body.
• Edge-case clinical scenarios: Validation covers protocol-defined inputs only; rare or atypical presentations may not be addressed.

Limitations

• Technical validation does not imply clinical validation.
• Rule logic reflects the evidence cited at time of authoring; authors may have introduced errors in rule encoding.
• All validation is the responsibility of the clinician deploying and using this tool.