Bridging Algorithms and Appointments: A Human-Centered Data Science Early Report on Orthopedic Surgical Scheduling (Preprint)

BACKGROUND Orthopedic surgical scheduling in multi-site health systems is a high-volume, cognitively demanding coordination task shaped by payer rules, local capabilities, and evolving clinical criteria. Decisions about whether a case is appropriate for an ambulatory surgery center (ASC) or requires a hospital operating room (OR), and at which specific site it should occur, are often made using fragmented documentation, tacit institutional knowledge, and ad hoc communication. Prior human-centered data science (HCDS) work has shown that generic scheduling tools and commercial optimization platforms often fail to account for hyperlocal operational realities in complex care environments [1-4]. OBJECTIVE This Early Report describes the development of a location optimization system, a rule-based, human-in-the-loop decision-support tool designed using human-centered design (HCD) and HCDS methodologies. We characterize the participatory design process, socio-technical governance structures, and infrastructural constraints shaping the system, and outline a roadmap for future AI-augmented extensions. METHODS Guided by the HCDS framework [1-3], we conducted iterative problem scoping, stakeholder mapping, contextual inquiry, workflow shadowing, rule elicitation and rapid prototyping with schedulers, surgeons, payer-contract staff, and service-line leaders. Curated electronic health record (EHR) and scheduling views were engineered to support near real-time case processing, with attention to data-quality issues highlighted in prior digital-health and EHR-based improvement work [5-11]. Design decisions emphasized transparency, interpretability, auditability, and preservation of stakeholder agency, consistent with best practices in participatory and community-engaged AI for health [1-4,18-21]. RESULTS The optimization system consolidates key case data, applies encoded safety and payer rules, and produces ranked ASC and OR site-of-care recommendations accompanied by natural-language explanations. Hard stops and data gaps are surfaced explicitly. Structured override mechanisms treat human judgment as a governance feature rather than an exception. Role-based dashboards summarize recommendation patterns, ASC utilization, and override activity for service-line leaders. A phased rollout plan, beginning with high-volume orthopedic subspecialties, supports incremental refinement informed by log data and user feedback, aligned with digital-operations and pathway-redesign studies in orthopedics and trauma care [12-17]. CONCLUSIONS The system demonstrates how HCD and HCDS can guide the creation of a transparent, hyperlocal site-of-care recommendation tool in a complex, multi-institutional orthopedic service line. By foregrounding explainability, structured overrides, and governance, the project establishes socio-technical foundations for subsequent AI-augmented decision support. The methods, stakeholder-engagement structures, data architecture (Figure 1), and stakeholder mapping (Table 1) may generalize to other health-system operational domains even as the specific rules remain locally tailored [1-4,18-21]. CLINICALTRIAL N/A