A maturity model for AI-augmented clinical biochemistry: the BASM framework for diagnostic stewardship, algorithmic governance, and health-service outcomes

Clinical biochemistry generates approximately 70% of the objective data used in medical decision-making, yet the integration of artificial intelligence (AI), machine learning, and large language models (LLMs) into laboratory operations remains fragmented. Existing digital-health maturity constructs (HIMSS EMRAM, HITI, CAMM) do not jointly address analytical metrology, algorithmic lifecycle governance, and clinical outcome accountability, leaving laboratories without a structured roadmap for AI-augmented diagnostic stewardship. To develop and face-validate the biochemistry–AI Stewardship Maturity (BASM) framework, a system-level maturity model that links laboratory analytics, AI governance, and health-service outcomes across five progressive tiers. A PRISMA-ScR-guided scoping review of six databases (MEDLINE, Embase, Scopus, Web of Science, IEEE Xplore, Cochrane) covering January 2022–July 2026 identified 1,482 records; 29 sources were retained after dual-independent screening (Cohen’s κ = 0.86). Charted evidence was triangulated with IFCC, ADLM, CAP, and CLSI position statements, and consolidated into a seven-dimension rubric. Face validity was assessed with five laboratory directors, and structural comparisons were performed against four precedent maturity models. BASM operationalizes maturity across seven dimensions—analytical foundations, data infrastructure, algorithmic lifecycle governance, diagnostic stewardship, workforce competency, equity and safety, and outcome accountability—resolved into five ordinal tiers (BASM-1 to BASM-5) with 12–24-month transition milestones. A chain of custody aggregation rule prevents high analytical scores from masking governance or equity deficits. Alignment with TRIPOD + AI, DECIDE-AI, and TRIPOD-LLM reporting standards is embedded at tiers 3–5. BASM provides laboratories, health systems, and regulators with an actionable, standards-aligned instrument to plan, benchmark, and audit AI-augmented diagnostic stewardship. A multi-institutional modified-Delphi validation is the priority next step.