Artificial intelligence for nursing data visibility in health technology assessment: policy architecture and implementation considerations

Abstract

<i>Background</i>: Health technology assessment (HTA) increasingly informs reimbursement, adoption, scale-up, and disinvestment decisions, yet many evidentiary traditions remain best suited to discrete, attributable interventions. Continuous, multidisciplinary care processes can therefore become systematically undermeasured, producing structural blind spots that undermine decision validity. Nursing is a paradigmatic test case: much nursing work is preventive, temporally diffuse, and inconsistently represented in routine datasets, so successful prevention can appear as "nothing happened."<i>Objectives/Aims</i>: To propose a pragmatic policy architecture for nursing-visible HTA that strengthens validity, equity, and implementability without increasing documentation burden, positioning artificial intelligence (AI) as an enabling method rather than a substitute for measurement design or causal reasoning.<i>Methods Design</i>: discursive discussion paper. Data Sources: targeted searches of PubMed, CINAHL, and Scopus, supplemented with key HTA, governance, and AI policy/standards documents and citation chaining. Search dates: databases were searched in January 2026. Literature used: English-language sources prioritised from 2015-2025 were screened for conceptual and policy relevance, then organised into three analytic themes aligned with the architecture: scoping, governance, and evidence generation.<i>Findings</i>: The architecture links three actions. Action 1 (Scoping) requires a small, feasible set of nursing-sensitive outcomes and implementation consequences, preferring measures already available from routine data to avoid new documentation. Action 2 (Governance) prevents surveillance harms and inequity through nursing decision authority, limits on punitive workforce surveillance, and equity checks with ongoing monitoring of unintended effects. Action 3 (Evidence) translates routine data into policy-grade evidence via transparent reporting and cautious causal claims; analytics (including ML/NLP when appropriate) are used as enablers rather than shortcuts.<i>Conclusion</i>: Integrating scoping, governance, and routine-data evidence can improve HTA validity, equity, and implementability for team-based care models where outcomes are co-produced by multidisciplinary work.

Ähnliche Arbeiten