<![CDATA[Introduction: Aluminium smelter feasibility studies must address not only technical and economic viability but also the early design of a safe socio-technical work system. This study evaluated how occupational safety, process safety, human factors, contractor interfaces, and operational readiness were integrated into one aluminium smelter feasibility study Methods: An evaluative qualitative case study combined a structured document review, literature- and standards-based benchmarks, and semi-structured interviews with four purposively selected experts. Each analytical indicator was classified as present, partially operationalized, or absent. The judgment was derived from explicit evidence requirements, document content, interview convergence or contradiction, and relevance to feasibility-stage decisions. NVivo supported data organization and auditability but coding frequency was not used as evidence of thematic importance. Results: The document contained conventional feasibility components and generic safety provisions. However, most macroergonomic dimensions were only partially operationalized. Competence planning, occupational hygiene, organizational arrangements, and process safety were mentioned but lacked defined requirements, deliverables, responsibilities, or verification criteria. Human-machine interface evaluation, contractor interface governance, major-accident analysis, and operational-readiness gates were absent or insufficiently specified. Expert evidence converged on six priority gaps: competency assurance, HMI and control-room risk, HAZOP/PHA and major-accident scenarios, contractor interfaces, occupational hygiene baselines, and operational readiness. Conclusion: This single-case study does not establish a general model for all aluminium smelter projects. Its contribution is a case-based evaluative framework that makes macroergonomic judgments traceable through explicit indicators and evidence thresholds. The framework can support identification of feasibility-stage gaps that may otherwise be transferred to engineering, commissioning, and operations, but its predictive value requires testing in additional projects and longitudinal settings.]]>
