This study adopted a scenario-based simulation design to evaluate the economic, organizational, and research impact of introducing a robotic rehabilitation unit in a medium-sized hospital. Two hypothetical but structurally equivalent hospitals were modeled and compared over a multiyear time horizon: clinic A, implementing a novel robotic rehabilitation unit integrated within standard rehabilitation services; and clinic B, providing conventional rehabilitation care only. Both scenarios were assumed to operate within the same regional health care system and under identical regulatory and reimbursement frameworks. The model was developed from the hospital perspective, focusing on direct capital and operating costs, organizational processes, and patient throughput. The time horizon of the analysis was set at 5 years to capture both initial investment costs and medium-term operational dynamics, including device depreciation and stabilization of clinical workflows. The model structure combined elements of budget impact analysis and organizational simulation, with baseline assumptions and sensitivity analyses to explore parameter uncertainty. No individual patient-level simulation was performed; instead, average patient profiles and treatment pathways were modeled based on typical caseloads in neurological and orthopedic rehabilitation units.

Table 1 summarizes the quantitative hypotheses and assumptions made for the model. Model parameters were derived from the following main sources: published literature on robotic rehabilitation effectiveness, costs, and organizational use; publicly available data on patient volumes in comparable European rehabilitation hospitals; price ranges reported by manufacturers, brochures, and clinical engineering reports; and feedback from clinics operating in the field.

When direct data were unavailable, conservative assumptions were adopted and tested through sensitivity analysis. All assumptions were explicitly reported in the model description.

Robotic-assisted rehabilitation was assumed to increase the achievable intensity of therapy by enabling longer or more frequent sessions without a proportional increase in therapist physical workload. Functional effectiveness values (Fugl-Meyer Assessment [FMA] motor score improvement) were derived from published randomized trials and meta-analyses comparing robotic-assisted and conventional rehabilitation. A comparative therapy quality framework was derived from the literature to characterize differences between robotic-assisted and standard rehabilitation. Evidence indicates that robotic rehabilitation may provide superior intrinsic therapeutic mechanisms through ad hoc assist-as-needed mechanisms but allows the delivery of a higher and more standardized therapy dose under comparable staffing conditions [5,8,10,11]. These dimensions were therefore used to qualitatively compare the expected quality of therapy between scenarios. Quantitative indicators of rehabilitation effectiveness were derived from published randomized trials and meta-analyses comparing robotic-assisted and conventional therapy. Reported ranges for repetitions per session, therapy frequency, and functional outcome improvements (FMA and activities of daily life [ADL] scales) were extracted and used to estimate cost-normalized effectiveness metrics. These values were combined with model-based cost estimates to compute cost-normalized indicators. This approach aimed to link therapy effectiveness and economic outcomes without assuming superior intrinsic efficacy of robotic rehabilitation but rather by reflecting higher delivered therapy dose and access under comparable staff conditions.

Other hypotheses for the model include the following assumptions. Klamroth-Marganska et al [23] reported at least 45 minutes per session per week. Based on these data and considering the capabilities of low-functioning patients, we opted for a conservative duration of 30 minutes per session.

Previous studies have reported modest but clinically relevant improvements in upper-limb motor function following robot-assisted or electromechanical rehabilitation, with between-group differences in the FMA of approximately 2 to 3 points reported in controlled trials [10], while systematic reviews have confirmed overall beneficial effects of robot-assisted and electromechanical training on upper-limb function [25]. Moreover, real-world observational data have reported larger pre-post FMA improvements following robot-assisted rehabilitation, in the range of approximately 7 to 11 points depending on the robotic technology used [10,25,26], supporting the plausibility of a moderate improvement in motor outcomes under intensive robotic rehabilitation scenarios. Based on these data, we chose a realistic, cautious scenario (+5/+7 FMA points with robot, +3/+5 points with standard therapy).

Maximum robot capacity was set to 100%. Since this value might be unrealistic or at least difficult to achieve, sensitivity analysis will also consider lower robot use rates, as well as 10% of sessions being lost.

Auxiliary personnel expenditures—such as initial technology-specific training programs and variations in treatment efficiency linked to individual therapist expertise—were modeled as structural overheads rather than dynamic per-session variables. While onboarding new robotic devices requires initial staff training, these setup costs are typically amortized across institutional operational budgets and do not alter the macrolevel comparative cost-effectiveness between robotic and manual pathways.

Economic parameters related to hardware acquisition, installation, and maintenance were established using aggregated market benchmarks and standardized vendor quotations reflecting commercial upper-limb robotic technologies, ensuring generalizability beyond specific proprietary brand pricing.

Clinical outcomes are reported as aggregate average FMA score improvements across standardized virtual patient cohorts rather than stratified baseline severities, reflecting macrolevel therapeutic efficacy rather than individual patient-level recovery curves. Calculating cumulative annual clinical gains via weighted averages assumes population-level homogeneity, thereby omitting individual response variance and nonlinear recovery profiles typical of heterogeneous clinical cohorts.

Consumable costs per session were modeled using a standardized parity assumption across both pathways. While robotic systems may involve specific component replacements or specialized hygienic accessories, these marginal expenses account for a minimal share of total operational costs compared with labor and capital investments, preserving the structural validity of the cost comparison.

The 5-year analytical horizon focuses strictly on the initial capital amortization cycle. In real-world hospital operations, robotic rehabilitation devices frequently remain functional and heavily used well beyond this initial period, meaning that long-term operational phases often benefit from postamortization efficiency, where capital acquisition costs are fully absorbed. For this reason, no equipment disposal costs were considered. No discount rate was applied to costs or health outcomes over the 5-year time horizon.

We assumed institutional alignment and compliance with workflow reengineering. In clinical practice, successful technology integration depends on overcoming initial staff resistance through adequate training and change management. However, because robotic devices function as clinical support tools designed to reduce physical fatigue and augment supervision capacity rather than displace personnel, organizational integration is structured as a collaborative enhancement of the existing health care team.

Session-specific consumables (such as hygienic interfaces and accessories) are modeled at a flat rate of €5 per session for both scenarios. Administrative overhead and scheduling logistics are assumed to scale proportionally with patient throughput.

C_Consumables = total annual sessions ∙ €5

A uniform reimbursement rate of €33.75 per rehabilitation session was assumed for both clinics, irrespective of inpatient or outpatient status. Annual revenue was calculated as the number of patients treated per month × 12 months × 13 sessions per patient × €33.75 per session. This uniform rate was used as a simplifying assumption for the economic comparison and was not intended to represent setting-specific reimbursement tariffs.

Total annual direct costs (DCs) for each clinic are calculated as the sum of personnel, technology/depreciation, and consumables (maintenance + consumables). Cost per patient and cost per FMA point gained are evaluated as core value-based metrics to determine economic efficiency.

DC = C_Staff + C_Tech + C_Consumables

Total annual revenue is derived from standard institutional reimbursement rates for outpatient and inpatient rehabilitation cycles. The gross operating margin (earnings before interest, taxes, depreciation, and amortization [EBITDA]) is calculated as:

EBITDA = total annual revenue – DC

Finally, cost per patient and cost per FMA point gained are evaluated as core value-based metrics to determine economic efficiency.

The outcome model was developed to capture three complementary dimensions: clinical, organizational, and scientific/innovation-related outcomes.

Organizational outcomes focused on the operational performance of the rehabilitation service and included annual patient throughput, defined as the total number of patients who successfully complete their treatment program within a year. This metric reflects the facility’s total capacity and is a driver for amortizing fixed technological costs. Total annual sessions refers to the aggregate number of individual therapy sessions delivered by the department annually. It is determined by the patient volume multiplied by the number of sessions prescribed per clinical protocol. Annual therapy hours represent the total volume of effective treatment time provided to the patient population. Given a standard session duration (eg, 30 minutes), this metric represents the cumulative “dose” of therapy delivered by the unit. Workforce needed (FTE) refers to the number of FTE staff members required to sustain the projected clinical workload. This measure highlights how robotic integration can decouple patient volume from human labor requirements. Productivity (sessions/FTE) is an efficiency ratio representing the average number of therapy sessions managed by 1 full-time therapist per year. It quantifies the “workforce multiplier” effect of technology, showing the increase in output per staff member enabled by robotic-assisted supervision models.

Regarding therapists, hours of therapy (annual therapy hours) represent the total cumulative volume of effective treatment time delivered to the entire patient population over 1 year. This metric is calculated based on the total number of sessions multiplied by the duration of each session. It serves as a measure of the total “therapeutic dose” the facility is capable of providing. FTE needed (workforce needed) indicates the number of FTE therapists required to cover the total annual therapy workload. In the model, it accounts for the different supervision ratios, showing how robotic integration (1:2 ratio) allows a facility to manage higher patient volumes with fewer staff members compared with the manual 1:1 model. Efficiency (therapy hours per FTE) measures the amount of therapy time successfully delivered by each full-time therapist unit per year. It quantifies the operational gain of the robotic unit, demonstrating how technology enables therapists to “produce” more treatment time by supervising multiple automated sessions simultaneously.

Economic evaluation was performed using a combination of total annual revenue, which represents the total revenue generated from rehabilitation services in a year; DCs (staff + tech), which are the sum of personnel costs (physical therapists required to cover therapy hours) and technology costs (depreciation of machines, maintenance, and consumables); and operating margin (EBITDA), which indicates the gross operating margin (revenue – DCs).

Clinical outcomes were modeled at an aggregate level and included weighted average FMA improvement, which represents the average change in the FMA score—the gold standard for evaluating motor recovery—achieved by patients after a full treatment cycle. It is “weighted” to reflect the specific clinical evidence found in the literature for both robotic-assisted (scenario A) and manual (scenario B) therapies. Annual clinical gain (total FMA points) is a systemic impact metric that quantifies the “total amount of recovery” produced by the department in 1 year. It is calculated by multiplying annual patient throughput by the weighted average FMA improvement. Therapy delivery effectiveness measures the clinical gain achieved relative to the resources invested (eg, FMA points per hour of therapy or per therapist FTE). It evaluates the “quality” of the organizational model, demonstrating how effectively the clinical protocol translates staff effort and technological support into measurable patient recovery.

To provide a comprehensive evaluation of the transition from manual to robotic-assisted therapy, a multidimensional framework based on health technology assessment (HTA [31]) principles was adopted. Beyond clinical efficacy, specific metrics to quantify organizational scalability (eg, staff engagement ratio) and economic sustainability (eg, cost per FMA point gained) were introduced. These parameters allow for a value-based comparison, aligning clinical recovery with the financial viability of the health care provider.

DC per patient calculates the average direct expenditure incurred by the hospital to provide a complete treatment cycle to a single patient. It includes the cost of specialized staff (physiotherapists) and the proportional cost of technology (robot depreciation and maintenance). It is an indicator of operational efficiency, showing how the robotic model can lower the “unit price of care” by scaling the number of patients treated.

Cost per FMA point gained is a cost-effectiveness indicator. It is calculated by dividing the DC per patient by the weighted average FMA improvement. It represents the financial investment required to achieve a single point of motor recovery on the FMA scale.

Staff engagement ratio (or supervision ratio) defines the relationship between the number of therapists and the number of patients treated simultaneously. In clinic B, this is a 1:1 ratio (1 therapist per patient), whereas in clinic A, it shifts toward a 1:2 or hybrid model. It quantifies workforce leverage, allowing therapists to oversee more treatments without increasing their physical workload.

Marginal clinical gain per patient represents the additional clinical improvement achieved by a patient in the robotic scenario (clinic A) compared with standard manual care (clinic B). It is calculated as the difference between the average FMA score improvement in the 2 scenarios. It quantifies the “added value” of technology in terms of individual recovery.

Clinical reach potential measures the department’s ability to extend specialized rehabilitation to a broader portion of the target population. It is linked to increased patient throughput and reduced waiting lists. It quantifies the public health impact by measuring how many more patients can access high-quality treatment within the same time frame and infrastructure.

Scientific and innovation outcomes were modeled as proxy indicators of research capacity rather than as direct outputs. These included the potential number of observational or pilot studies enabled by structured data collection, the volume of performance datasets generated annually, the estimated number of scientific publications and conference contributions, and the potential to initiate funded research projects and industrial collaborations. These outcomes were estimated based on the volume of treated patients and the availability of quantitative motor data produced by robotic systems.

To assess the robustness of the model to assumptions, one-way and multiparameter sensitivity analyses were conducted on key variables, considering the economic and clinical outcomes. Two variables were chosen: patient volume and the robot use rate, which indicates the extent to which the robot unit is used and encompasses, in aggregate, the effects of the number of patients treated, the time the robot is not in use, and the number of patients handled by the therapists; and the cost per FMA point gained, as a measure of the effectiveness of the robot unit in improving the clinical quality of treatments. Threshold analyses were performed to identify minimum use rates and patient volumes required for economic sustainability.

The present study was based entirely on a hypothetical simulation model and did not involve direct patient recruitment, interventions, access to individual clinical records, or the collection or analysis of identifiable or pseudonymized patient-level data. The model incorporated assumptions informed by current clinical practice and relevant regulatory frameworks for medical devices, including the European Medical Device Regulation (MDR 2017/745) [32] and internationally recognized technical standards issued by the International Organization for Standardization (ISO) and the International Electrotechnical Commission (IEC), including ISO 13485 and the IEC 60601 series. Potential future clinical and kinematic data collection and secondary use were considered only as hypothetical scenarios, assuming appropriate anonymization and compliance with applicable data protection requirements, including the General Data Protection Regulation (GDPR) [33]. Any future observational studies conducted using the robotic unit would be subject to the applicable ethical review and regulatory requirements, including review by the relevant ethics committee where required, in accordance with the Declaration of Helsinki and applicable national regulations.

The model suggests that clinic A achieved a higher annual patient throughput than clinic B (Table 2), with 960 vs 720 patients per year (+33%). This increase was associated with a higher total number of therapy sessions (12,480 vs 9360) and greater annual therapy hours (6240 vs 4680). The modeled transition from 720 to 960 annual patients reflects potential technical capacity rather than guaranteed market expansion. Actual patient volume intake is bounded by external constraints such as regional health care demand, referral networks, and institutional budgetary caps. When local demand limits volume growth, operational performance converges toward lower use bands, underscoring the importance of regional market alignment before capital adoption. Despite the higher activity volume, the required workforce was slightly lower in clinic A (2.34 vs 2.44 FTE), resulting in a substantial gain in productivity, expressed as sessions per FTE (+39%). Overall, the robotic-hybrid organizational model allowed higher service capacity with comparable or reduced staffing levels.

Table 2 also shows that clinic A could manage a higher monthly patient volume than clinic B (80 vs 60 patients, +33%), corresponding to a higher number of therapy hours (520 vs 390 hours per month). Despite this increased workload, the estimated staff requirement was slightly lower in clinic A (2.34 vs 2.44 FTE). Consequently, staff efficiency, expressed as therapy hours per FTE, was markedly higher in the robotic-hybrid model (223 vs 160 hours/FTE and +63 hours/FTE), indicating a more efficient allocation of human resources.

The introduction of robotic units allows for a decoupling of patient throughput and staff requirements. While clinic B requires 2.44 FTEs to treat 60 patients, clinic A handles a 33% higher volume (80 patients) with a lower staff demand (2.34 FTEs), effectively increasing the therapy-to-staff productivity ratio by 39%.

Under the assumptions, clinic A generated higher total annual revenue than clinic B (€421,200 vs €315,900), despite moderately higher DCs (€264,100 vs €236,100). This resulted in a substantially higher operating margin for clinic A (€157,100 vs €79,800). Furthermore, the robotic-hybrid model was associated with a lower cost per patient (€275.10 vs €327.92), corresponding to a reduction of €52.82 per treated patient (Figure 1).

Sensitivity analysis was also conducted on the cost per FMA point gained, the primary indicator of value-based efficiency. The robustness of this metric was assessed by varying the clinical efficacy differential, defined as the additional FMA points gained through robotic therapy compared with standard care. In the baseline scenario, the robotic-hybrid model (clinic A) achieved an average improvement of 5.13 FMA points, compared with 4.00 points in clinic B, resulting in a 34.6% reduction in cost per point (€53.62 vs €81.98). Two alternative clinical outcomes were simulated. In the conservative scenario (clinical parity), both clinics achieved exactly 4.00 points, indicating that robotic therapy provided no additional clinical recovery over manual therapy. In the optimistic scenario (high intensity), the robotic group reached the upper bound of the expected improvement (7.0 points), increasing the weighted average improvement for clinic A to 5.70 points.

Assuming a fixed cost per patient of €275.10 (scenario A2 of the sensitivity analysis), even in the conservative scenario, where the robotic unit provides zero additional clinical benefit over manual therapy (clinical parity), clinic A remains 16.1% more cost-effective than clinic B (€68.77 vs €81.98). This suggests that the sustainability of robotic rehabilitation is “double-protected.” The financial viability is guaranteed by the organizational efficiency (FTE optimization and higher throughput), even if the clinical superiority were to be marginal. Conversely, when clinical superiority is realized (baseline and optimistic scenarios), the cost-effectiveness scales exponentially, offering a massive advantage for health care providers focused on maximizing patient recovery within fixed budgets. Results are summarized in Figure 3.

The results of this scenario-based simulation suggest that the introduction of a robotic rehabilitation unit in a medium-sized hospital (clinic A) may contribute to improving the “production function” of rehabilitative care. Unlike the traditional manual model (clinic B), which relies on a rigid 1:1 therapist-to-patient ratio, the hybrid model in clinic A may enable a decoupling of patient throughput from human workload. By optimizing the robotic units to a saturation point of 45 patients per month (representing approximately 56% of the total monthly caseload of 80 patients), clinic A may achieve a 33% increase in total volume compared with clinic B. Crucially, this expansion could be achieved without additional staffing. In fact, our FTE-based analysis reveals that clinic A requires only 2.34 FTEs to manage 80 patients, whereas clinic B requires 2.44 FTEs for just 60 patients. This efficiency gain is not attributed to an inherent clinical superiority of the robot in terms of recovery mechanisms but rather to its role as a workforce multiplier. In clinic A, the productivity ratio reaches 2666 therapy hours per FTE annually, a 39% improvement over the 1918 hours per FTE in clinic B. This allows the facility to absorb previously unmet rehabilitation demand while reducing the net cost per patient from €328 to €275.

The integrated analysis of clinic A’s hybrid model reveals the possibility of a dual advantage over traditional care. First, the introduction of robotic units may enhance the quality of clinical outcomes. By enabling higher dosage, precision, and repeatability, the robotic treatment could lead to a weighted average FMA improvement of 5.13 points, compared with 4.00 points in the manual-only setting of clinic B (+28.2%). When scaled to the annual hospital volume, the impact could be substantial: clinic A produces a total annual clinical gain of 4924 FMA points, which is 71% higher than the 2880 points generated by clinic B. This metric suggests that a robotic-equipped facility acts as a far more powerful “health generator” for the community, potentially providing not just more care, but more effective recovery per patient.

One of the most compelling findings of this study is the hypothetical possibility of a reduction in the cost of clinical recovery. In clinic B, the lack of technological support results in a high human-labor cost per unit of recovery, with each FMA point gained costing the facility €81.98. Conversely, in clinic A, the combination of a lower net cost per patient (€275.10) and clinical outcomes leads to a cost of €53.62 per FMA point. This 34.6% reduction in the cost of recovery represents a fundamental shift in hospital management. It demonstrates that the initial capital investment in robotics is not merely a “tech-upgrade” but a strategic move toward value-based health care. By lowering the cost of “producing” recovery, the hospital can reinvest the saved resources (approximately €28 per FMA point) into further expanding services or research.

The hybrid configuration of clinic A—where 45 patients use robotic units while 35 remain in standard care—is potentially a compelling organizational model. This setup allows the facility to handle a 33% higher patient volume with a theoretical workforce requirement of 2.34 FTEs, compared with the 2.44 FTEs needed for a lower volume in clinic B. The resulting staff engagement ratio of 0.72 in clinic A (ie, only 0.72 therapist-hours are required to produce 1 hour of therapy) highlights how technology mitigates the physical and temporal constraints of manual therapy. This “liberated time” is what enables the therapist to focus on high-level clinical decision-making and supervision, effectively increasing the therapy delivery effectiveness by 39% (2666 vs 1918 hours per FTE).

From an organizational perspective, the transition to an FTE-based resource management model highlights that technological investment is amortized through increased service capacity rather than staff reduction. The ability of a single therapist to supervise two robotic sessions simultaneously (1:2 ratio) effectively frees nearly 1 full therapist equivalent (0.91 FTE) compared with purely manual 1:1 delivery for the same volume. Furthermore, the robotic system introduces a level of operational standardization that is unattainable in manual therapy. Every session in clinic A follows precise, programmable protocols, reducing intertherapist variability and ensuring that patients with severe impairments receive a consistent, high-intensity dose of therapy (hundreds to thousands of repetitions per session) without increasing therapist physical fatigue.

Beyond economic and clinical metrics, the adoption of robotic technology offers qualitative advantages regarding data collection and research infrastructure. While our model does not formally quantify research output, the generation of continuous digital datasets ensures comprehensive traceability of training sessions. This technological environment serves as a qualitative enabler for observational research, supporting internal data consistency and fostering potential academic and industrial partnerships. Rather than positioning the clinical unit as an immediate quantitative “innovation hub,” these capabilities should be viewed as long-term institutional opportunities that depend heavily on internal research governance and strategic academic engagement. The projected benefits in terms of clinical gains (FMA improvements) and cost efficiencies are heavily contingent upon institutional adoption rates, consistent device use, and local reimbursement structures. Future empirical implementations are necessary to validate these theoretical projections against real-world clinical variabilities.

Despite the promising insights provided by this comparative analysis, some limitations should be acknowledged. First, this study relies on a scenario-based simulation model rather than longitudinal empirical data. While the model parameters—such as patient throughput, device costs, and staffing ratios—were derived from established clinical literature and realistic hospital benchmarks, they may not be applicable to every institutional setting. In fact, while these parameters were derived from robust clinical benchmarks and manufacturer specifications, the absence of real-world calibration means that the projected outcomes reflect idealized operational assumptions. Consequently, the quantitative outputs—such as exact cost reductions and throughput margins—should not be interpreted as absolute predictions but rather as comparative estimates illustrating structural organizational dynamics. Future research will prioritize multicenter prospective validation studies to calibrate these simulation parameters against longitudinal, real-world clinical and financial workflows.

We acknowledge that factors such as specific regional reimbursement policies, variations in labor costs, and the architectural constraints of different hospitals could influence the analysis. The clinical outcomes modeled in this study, specifically the weighted average FMA improvement, were estimated based on aggregated data from meta-analyses. While these values provide a robust baseline, they do not account for the high variability in individual patient recovery trajectories, which are influenced by the timing of intervention, lesion severity, and comorbidities. Third, the organizational “multiplier effect” assumes a consistent 1:2 therapist-to-patient ratio during robotic sessions. In practice, the initial learning curve for clinical staff and potential technical downtime may temporarily reduce this efficiency during the first year of implementation. Finally, the scientific and innovation outcomes are presented as proxy indicators of research potential. While the capacity for data generation is a clear technical advantage of robotic units, the transformation of these datasets into high-impact publications depends on the hospital’s internal research infrastructure and academic partnerships, which were not modeled. Other factors, such as real-world operational complexities—including patient transfer delays, device donning and doffing procedures, hygiene and disinfection protocols, unexpected technical downtime, spatial limitations, and long-term mechanical wear and tear—were not modeled at a microanalytical level. While annualized maintenance costs partially offset hardware degradation, high-frequency operational frictions can temporarily reduce the effective daily use rate. Consequently, actual hospital implementations must account for these microenvironmental overheads when targeting the operational efficiency thresholds identified in our sensitivity analysis. Moreover, the baseline assumption of a 100% robot use rate represents a theoretical operational ceiling rather than a realistic daily average. In actual clinical practice, auxiliary activities such as patient positioning, device donning and doffing, and routine software calibration consume a portion of the operational window. Nevertheless, our sensitivity analysis explicitly models underuse thresholds (eg, lower operational saturation), demonstrating that while efficiency margins compress, the model retains economic viability provided that use remains above the critical break-even limit. The analysis was conducted strictly from the institutional hospital perspective, omitting a patient-level microsimulation or a broader societal perspective. While institutional decision-making is primarily governed by operational budgets, throughput capacities, and direct labor costs, future investigations should incorporate patient-level longitudinal tracking to capture individual variability in recovery trajectories and out-of-pocket expenses. Finally, while conventional therapy is subject to performance variability driven by physical therapist fatigue, robotic pathways rely on electromechanical consistency. Nonetheless, operational continuity in robotic settings can occasionally be modulated by technical factors such as routine software recalibrations or minor maintenance interventions, elements that are mitigated by standardized service agreements and safety protocols. Future research should focus on multicenter prospective studies to validate these simulation-based findings with real-world longitudinal evidence. Results may not generalize to hospitals with different reimbursement schemes or staffing costs.

Although the current sensitivity analysis focused on primary throughput and cost boundaries, we acknowledge that secondary operational factors—such as unexpected robotic downtime, fluctuations in specialized physiotherapist wages, regional variations in reimbursement tariffs, and escalating maintenance agreements—represent real-world variables that could influence economic margins. In practice, a severe increase in technical downtime (reducing device use below critical operational thresholds) or a downward revision of regional Diagnosis-Related Group tariffs would compress the gross operating margin (EBITDA) calculated in clinic A. Nevertheless, the structural robustness demonstrated across our high- and low-use sensitivity ranges suggests that the model maintains economic viability provided that patient volume remains stable. Future implementations should incorporate real-world stochastic tracking of these secondary parameters as longitudinal data become available.