Introduction
Hospitals purchase major medical equipment to expand clinical capability, shorten time to diagnosis or treatment, maintain emergency coverage, and support specialist services. At approval, clinical departments usually present clinical need and projected activity, while management reviews purchase or lease cost, maintenance, staffing, and reimbursement. After adoption, however, the original assumptions about cases, revenue, contribution margin (CM), quality, and recovery period may no longer be available in the same unit in which the investment was justified. The resulting gap is a hospital-internal health technology assessment (HTA) problem: how should an adoption hypothesis be preserved, tested, and returned to the next capital-allocation decision?
Department-level accounting is often too coarse for this task. A department is an administrative and personnel unit, whereas a patient pathway may cross imaging, operating rooms, wards, rehabilitation, and post-discharge services. Clinical service line (CSL) management offers a unit closer to a patient group and care pathway than a formal department (Reference Parker, Charns and Young1). Value-based health care likewise emphasizes outcomes that matter to patients in relation to the resources required to achieve them (Reference Porter2). Financial recovery is therefore relevant to organizational sustainability, but it is not sufficient evidence that an investment generated value for patients, healthcare professionals, the organization, or the health system.
The internationally developed definition of HTA describes a multidisciplinary process that determines the value of a health technology at different points in its lifecycle to inform decision making (Reference O’Rourke, Oortwijn and Schuller3). Lifecycle HTA (LC-HTA) extends this logic by connecting sequential assessment activities when evidence, the technology, or its context may change over time (Reference Kirwin, Round, Bond and McCabe4–Reference Pichler, Boysen and Mittmann6). Medical equipment is a natural application because utilization, staff learning, maintenance, reimbursement, and clinical pathways can change after purchase. A hospital-level method is needed to convert the assumptions approved at adoption into a planned post-adoption evidence record.
Japan provides a useful implementation setting because universal health insurance, regulated fees, emergency-care obligations, regional planning, and public statistics shape hospital investment decisions (7). The procedural problem is nevertheless international. Public claims or equipment datasets may be available in some countries and absent in others, whereas most hospitals retain at least contracts, asset records, activity data, and accounting records. This study therefore had two objectives: to specify a CSL Investment Recovery Framework (CSL-IRF) as a hospital-level post-adoption component of LC-HTA, and to examine initial data and calculation feasibility through a retrospective application using routine aggregate hospital records.
Conceptual position within HTA
National or payer-level HTA does not directly answer hospital-specific questions about local capacity, shared equipment, staffing constraints, referral geography, and competition for capital across clinical programs. Hospital-based health technology assessment (HB-HTA) and mini-HTA support local adoption decisions by combining general evidence with hospital context and implementation requirements (Reference Sampietro-Colom, Lach and Pasternack8–11). The CSL-IRF is not a substitute for these approaches. It adds a post-adoption layer that retains the approved CSL hypothesis and returns AIB, CM, allocation, value indicators, and evidence status to reassessment and renewal decisions. This responds to the continuing difficulty of measuring the effects and repercussions of local HTA when follow-up and comparable baselines are inconsistently documented (Reference Gagnon, Desmartis, Poder and Witteman12).
The framework is complementary to LC-HTA rather than equivalent to it. LC-HTA connects assessment activities across time and requires the decision problem, activity sequence, and optimization criteria to be specified prospectively (Reference Kirwin, Round, Bond and McCabe4–Reference Pichler, Boysen and Mittmann6). Within that broader process, the CSL-IRF operationalizes one hospital-level sequence: approval, evidence-field preservation, post-adoption measurement, reassessment, and feedback to the next investment. Coverage with evidence development contributes the logic of conditional use with planned evidence generation (Reference Hutton, Trueman and Henshall13;Reference Trueman, Grainger and Downs14), while health technology reassessment and disinvestment support renewal, replacement, scale-down, or withdrawal decisions for technologies already in use (Reference MacKean, Noseworthy and Elshaug15;Reference Calabrò, La Torre and de Waure16). Value-based procurement adds lifecycle cost, maintenance, training, utilization, and outcomes to the purchasing decision (Reference Rahmani, Karimi, Rezayatmand and Raeisi17;Reference Diaconu, Chen, Cummins, Jimenez Moyao, Manaseki-Holland and Lilford18).
Implementation also depends on leadership, data infrastructure, multidisciplinary participation, and integration with existing hospital decision processes, which remain recognized barriers in HB-HTA (Reference Lipska, Di Bidino and Niewada19). The CSL-IRF addresses a related information-asymmetry problem. Executives are accountable for capital allocation and organizational sustainability, whereas clinical applicants hold more detailed knowledge of need, demand, and workflow. Agency theory explains the risk of forecasts that will not later be reviewed in the same responsibility unit (Reference Jensen and Meckling20;Reference Eisenhardt21). Organizational learning requires results to be compared with the assumptions that produced action, so that the rules themselves can be revised (Reference Argyris and Schön22).
Management accounting, capital budgeting, and time-driven activity-based costing provide tools for responsibility units, annual burden, margin, and progressively finer cost attribution (Reference Anthony and Govindarajan23–Reference Keel, Savage, Rafiq and Mazzocato26). The balanced scorecard supports joint financial and nonfinancial monitoring but does not by itself preserve an equipment-specific investment hypothesis, allocation responsibility, and evidence status (Reference Kaplan and Norton27). The CSL-IRF is therefore intended as a learning and accountability system, not a punitive scorecard.
Methods
Framework development and source mapping
Framework development followed four steps: problem specification; targeted mapping of source domains that contributed a design requirement; construction of indicators, allocation rules, evidence-status labels, value guardrails, and governance triggers; and internal logic testing. The source domains were CSL management and value-based health care; HTA, HB-HTA, mini-HTA, and LC-HTA; coverage with evidence development, reassessment, disinvestment, and value-based procurement; management accounting and capital budgeting; time-driven activity-based costing; agency theory; organizational learning; and external health-system statistics. A targeted update of the current HTA definition and LC-HTA literature was completed on 20 August 2026. This was not a systematic or scoping review, and the mapping should be interpreted as a transparent design rationale rather than an exhaustive evidence synthesis. Search concepts, mapping boundaries, and component-level traceability are provided in Supplementary File 1.
Framework specification and internal checks
The artifact was specified around a fixed CSL definition; pre-adoption activity, revenue, variable-cost, value, and allocation hypotheses; post-adoption measurement using the same definitions; and governance action. Internal checks examined double counting between depreciation and lease payments, confusion between total and incremental CM, counterfactual status, shared-asset allocation, quality or access deterioration despite favorable finance, and sensitivity to burden, demand, and margin assumptions. Synthetic calculations were retained as boundary tests and moved to Supplementary File 1; they were not treated as empirical validation.
Retrospective records-based feasibility application
A retrospective application was conducted at one acute-care hospital within Fukujukai Medical Corporation. Three non-retired medical-equipment assets suggested by the study team were purposively screened from the asset list available in February 2026 because they appeared mappable to a CSL. This was an exploratory sample rather than a census of investments. A leased 4 K arthroscopy camera head linked to shoulder arthroscopy was selected because the asset and lease record and corresponding procedure activity could be identified for 12 months before and after its August 2025 introduction.
Available sources were the asset and lease record and aggregate claims-based procedure statistics. The pre-implementation period was 1 August 2024, through 31 July 2025, and the post-implementation period was 1 August 2025, through 31 July 2026. The same reimbursed procedure items were used in both periods; automobile-liability and workers’ compensation cases were excluded. Revenue represented nominal procedure-fee revenue only, not total episode or service-line revenue. A reimbursement revision occurred during the post-implementation period; revenue was therefore treated as a secondary descriptive measure. No patient identifiers, patient-level records, or clinical outcomes were accessed.
The application first assessed whether each minimum framework field could be retrieved: original application and planned values, contract and annual burden, activity and revenue, variable cost, committed and actual-use allocation, and value indicators. Available values were summarized descriptively. Because variable costs and original planned CM were unavailable, actual CM, contribution margin ratio (CMR), investment burden coverage ratio (IBCR), and ex-post performance ratio (EPR) were not estimated. Instead, the minimum required CM per observed additional procedure and an attribution-margin threshold were calculated. No statistical test or causal before-and-after claim was made.
Results
Framework architecture
The CSL-IRF links a pre-adoption decision record to post-adoption measurement and the next investment decision (Figure 1). Intended users include hospital executives, clinical and CSL leaders, finance and fixed-asset staff, management-planning units, clinical engineering, operating-room or imaging managers, and quality teams. Minimum internal inputs are the investment application, CSL definition, asset or lease record, activity data, revenue and variable-cost data, an allocation rule, and prespecified value indicators. External claims, registry, equipment-capacity, or regional data are optional contextual inputs rather than mandatory components or automatic thresholds.
Clinical service line investment recovery framework architecture. The clinical service line investment recovery framework (CSL-IRF) links a fixed pre-adoption clinical service line (CSL) hypothesis to post-adoption measurement, reassessment, governance action, and feedback to the next capital decision. Optional external evidence can refine assumptions but does not replace minimum hospital records. AIB, annualized investment burden; CM, contribution margin.
Figure 1. Long description
The flowchart is organized into three horizontal layers.
Top Layer:
* A dashed box labeled Optional external contextual evidence includes national claims, registries, capacity surveys, or regional benchmarks. Two downward arrows point from this box to the first and second stages of the core process.
Middle Layer (Core Process):
* Stage 1: Pre-adoption commitment. Lists fixed C S L and target population, planned activity, revenue, variable cost, and C M, minimum A I B and allocation, and review date and value guardrails. An arrow points right to Stage 2.
* Stage 2: Post-adoption measurement. Lists same C S L definitions, activity, revenue, variable cost, and C M, A I B and actual-use allocation, and value indicators and evidence status. An arrow points right to Stage 3.
* Stage 3: Reassessment and action. Lists Tier 1, Tier 2, or feasibility-only, financial interpretation plus value guardrails, and actions: continue, remediate, reallocate, scale down, replace, or renew.
Bottom Layer:
* A wide box labeled Minimum internal data architecture includes application and C S L definition, contract or asset ledger, activity, revenue and variable cost, use or allocation, and safety and access indicators. Three upward arrows point from this box to each of the three stages above.
* At the very bottom, a text line reads: Feedback to the next capital-allocation decision.
At application, the hospital records the target CSL and patient group, expected cases, revenue, variable cost, planned CM, AIB, review timing, value indicators, and ex-ante allocation weights. After adoption, the same definitions are used for observed activity, revenue, variable cost, CM, AIB, allocation, and value status. At reassessment, deviations are classified as demand or referral shortfall, staffing or slot constraint, reimbursement or material-cost change, allocation mismatch, or a value-related restriction. The finding is then linked to continuation, remediation, reallocation, scale-down, replacement, or renewal.
Financial and value-governance indicators
For CSL
$ s $
, AIB is defined as the annualized accounting burden allocated from equipment or contract
$ e $
. It includes depreciation
$ {D}_e $
, maintenance
$ {M}_e $
, lease cost
$ {L}_e $
, update or upgrade cost
$ {U}_e $
, and annualized installation or training cost
$ {F}_e $
, with CSL allocation weight
$ {w}_{e,s} $
. Thus, CSL-specific AIB is
$ {\mathrm{AIB}}_s=\sum \limits_e{w}_{e,s}\left({D}_e+{M}_e+{L}_e+{U}_e+{F}_e\right) $
.
AIB is an accounting view of annual burden rather than cash flow. Depreciation and lease cost are mutually exclusive for the same component, and mixed arrangements must be itemized. When new dedicated fixed operating costs are required, they should be recorded separately or added to an expanded recovery denominator.
CSL-specific CM is revenue
$ {R}_s $
minus variable cost
$ {\mathrm{VC}}_s $
, or
$ {\mathrm{CM}}_s={R}_s-{\mathrm{VC}}_s $
. Variable cost includes materials, drugs, outsourced tests, consumables, and other costs that vary with cases; depreciation and fixed staff cost are not included in this definition. For a new expansion investment, the preferred numerator is incremental CM, defined as post-investment CM minus the counterfactual CM that would have occurred without the investment:
$ \varDelta {\mathrm{CM}}_{s,t}={\mathrm{CM}}_{\mathrm{post},s,t}-{\mathrm{CM}}_{\mathrm{counterfactual},s,t} $
.
Counterfactual CM cannot be generated by framework logic alone. A credible strategy may use a prespecified pre-period trend, matched comparator, interrupted time series, or capacity-adjusted baseline. Tier 1 denotes a defensible incremental-CM analysis. Tier 2 denotes planned-versus-observed commitment tracking and must not be interpreted causally. If even the original plan is unavailable, the application is classified as a data-feasibility assessment rather than a completed Tier 2 review.
The CMR is
$ {\mathrm{CM}\mathrm{R}}_s={\mathrm{CM}}_s/{R}_s $
. Required additional revenue to cover the annualized burden is
$ {R}_{\mathrm{req},s}={\mathrm{AIB}}_s/{\mathrm{CMR}}_s $
. If case-level CM is
$ {\mathrm{CM}}_{\mathrm{case},s} $
, the break-even volume is
$ {Q}_{\mathrm{BE},s}={\mathrm{AIB}}_s/{\mathrm{CM}}_{\mathrm{case},s} $
. For expansion investments,
$ {\mathrm{IBCR}}_{s,t}^{\mathrm{inc}}=\varDelta {\mathrm{CM}}_{s,t}/{\mathrm{AIB}}_{s,t} $
. For replacement or continuity investments, total CSL CM may be used as a supplementary numerator,
$ {\mathrm{IBCR}}_{s,t}^{\mathrm{total}}={\mathrm{CM}}_{s,t}/{\mathrm{AIB}}_{s,t} $
. An IBCR below 1.0 means that CM has not covered the annualized accounting burden during that period; it does not automatically mean that cash flow is negative because AIB contains depreciation.
Margin of safety (MOS) compares observed volume with break-even volume as
$ {\mathrm{MOS}}_{s,t}=\left({Q}_{\mathrm{obs},s,t}-{Q}_{\mathrm{BE},s,t}\right)/{Q}_{\mathrm{obs},s,t} $
. The EPR compares observed cumulative CM with the planned CM hypothesis stored at approval:
$ {\mathrm{EPR}}_{s,t}=\frac{\sum \limits_{\tau =1}^t{\mathrm{CM}}_{\mathrm{obs},s,\tau }}{\sum \limits_{\tau =1}^t{\mathrm{CM}}_{\mathrm{plan},s,\tau }} $
. Here,
$ {\mathrm{CM}}_{\mathrm{plan}} $
is the planned contribution-margin hypothesis; it is distinct from committed allocation weight. The cumulative gap is
$ {G}_{s,t}=\sum \limits_{\tau =1}^t\left({\mathrm{CM}}_{\mathrm{obs},s,\tau }-{\mathrm{CM}}_{\mathrm{plan},s,\tau}\right) $
.
Shared assets create a second allocation problem. If only observed use is used to allocate AIB, a CSL that requested the equipment but did not use it can lose all responsibility for the original investment hypothesis. The CSL-IRF therefore stores ex-ante committed allocation separately from ex-post actual-use allocation. Committed AIB is
$ {\mathrm{AIB}}_s^{\mathrm{commit}}=\sum \limits_e{w}_{e,s}^{\mathrm{commit}}\left({D}_e+{M}_e+{L}_e+{U}_e+{F}_e\right) $
. Actual-use AIB is
$ {\mathrm{AIB}}_{s,t}^{\mathrm{actual}}=\sum \limits_e{w}_{e,s,t}^{\mathrm{actual}}\left({D}_e+{M}_e+{L}_e+{U}_e+{F}_e\right) $
. Table 1 summarizes the financial indicators.
Financial indicators for CSL-level investment monitoring
Table 1. Long description
The table consists of four columns: Indicator, Definition, Primary use, and Interpretation boundary.
* A I B: Defined as allocated annual depreciation or lease cost plus maintenance, updates, and annualized installation or training. Primary use is the annual accounting burden of the investment. Interpretation boundary notes it is not initial cash outflow and owned/leased components must not be double counted.
* C M: Defined as revenue minus variable cost. Primary use is contribution before fixed burden. Interpretation boundary requires a consistent service-line variable-cost definition.
* Incremental C M: Defined as observed C M minus counterfactual C M. Primary use is the preferred numerator for expansion investments. Interpretation boundary requires a prespecified and defensible counterfactual Tier 1.
* C M R: Defined as C M divided by revenue. Primary use links required revenue to A I B. Interpretation boundary states it should only be interpreted with the same revenue and variable-cost scope.
* I B C R: Defined as eligible C M divided by A I B. Primary use is whether C M covers annual accounting burden. Interpretation boundary notes that I B C R less than 1 does not by itself establish negative cash flow.
* E P R: Defined as observed cumulative C M divided by planned cumulative C M. Primary use is commitment tracking against the approved plan. Interpretation boundary notes it is not causal and cannot be calculated if planned C M was not preserved.
* Allocation: Defined as committed allocation and actual-use allocation retained separately. Primary use is responsibility and realized use of shared assets. Interpretation boundary notes a proxy based on procedure share is not a direct equipment-use measure.
* Evidence status: Defined as Tier 1, Tier 2, or data-feasibility assessment. Primary use is to prevent overinterpretation. Interpretation boundary states the label is part of the result, not a quality score.
Note: AIB, annualized investment burden; CM, contribution margin; CMR, contribution margin ratio; CSL, clinical service line; EPR, ex-post performance ratio; IBCR, investment burden coverage ratio.
Financial recovery is necessary for sustainability but is not sufficient evidence of value realization. The framework therefore retains separate value domains for patients, healthcare professionals, the organization, and the health system or society (Table 2). Safety and access indicators act as veto gates: favorable financial results cannot support expansion or renewal while a prespecified guardrail is breached. Workforce capability, public service, equity, and environmental sustainability are reported alongside financial recovery rather than collapsed into an unvalidated composite score.
Value domains retained alongside financial recovery
Table 2. Long description
The table consists of four columns and six rows of data.
* Row 1: Patient clinical value. Indicators include safety events and technical success. Minimum evidence is a prespecified clinical or safety source. Governance role: Safety deterioration can veto expansion or renewal.
* Row 2: Access and equity. Indicators include waiting time and geographic access. Minimum evidence is an administrative or access measure. Governance role: Access deterioration can trigger remediation despite favorable finance.
* Row 3: Healthcare professionals. Indicators include workload and training. Minimum evidence is a staffing or competency record. Governance role: Supports workforce planning and identifies hidden operating burden.
* Row 4: Organization. Indicators include capacity and strategic capability. Minimum evidence is an operational or strategic indicator. Governance role: Makes nonfinancial organizational value auditable.
* Row 5: Health system, society, and C S R. Indicators include regional availability and transparent stewardship. Minimum evidence is a locally selected public-value indicator. Governance role: Supports accountability without automatically monetizing public value.
* Row 6: Environmental sustainability. Indicators include energy and waste. Minimum evidence is a prespecified environmental measure. Governance role: Reported as a separate domain with no universal composite score assumed.
Note: CSR, corporate social responsibility; Candidate indicators should be selected before adoption and interpreted within the local decision context.
Decision-making and reassessment process
The review sequence begins with evidence status. If a credible counterfactual was prespecified, incremental CM and incremental IBCR may be interpreted under Tier 1 assumptions. If only planned and observed values are available, Tier 2 supports accountability but not causal inference. If planned fields, costs, allocation, or value indicators are missing, the first output is a data-availability map showing which decisions cannot yet be supported. Governance action follows cause classification rather than a single threshold. Illustrative triggers and synthetic boundary tests are provided in Supplementary File 1 and require local calibration.
Retrospective records-based feasibility application
Of the three purposively screened assets, one could be linked to both an asset and lease record and comparable pre- and post-implementation procedure statistics. The selected 4 K arthroscopy camera head had a 72-month lease, a total contract value of JPY 4,728,240, and a documented annual lease component of JPY 788,040. Separate maintenance, update, installation, and training costs were not identifiable, so JPY 788,040 was treated as the minimum documented AIB rather than a complete AIB estimate.
Shoulder arthroscopy activity increased from 220 procedures in the pre-period to 245 in the post-period, a difference of 25 procedures or 11.4 percent. Nominal procedure-fee revenue increased from JPY 102,782,400 to JPY 114,863,400, a difference of JPY 12,081,000 or 11.8 percent. Mean procedure-fee revenue per procedure changed only from JPY 467,193 to JPY 468,830, an increase of 0.35 percent. Figure 2 shows the descriptive period comparison.
Descriptive retrospective application to the selected equipment investment. Shoulder arthroscopy procedure count and nominal procedure-fee revenue are shown for the 12 months before and after the August 2025 lease start. Revenue represents the selected reimbursed procedure items, not total episode or service-line revenue, and the post-implementation period included a reimbursement revision. Differences are descriptive and are not attributed causally to the equipment.
Figure 2. Long description
A two-panel figure showing performance metrics for a 12 month pre-implementation period and a 12 month post-implementation period.
Left Panel: A bar chart measuring Shoulder arthroscopy procedures (n). The vertical y-axis ranges from 0 to 250 in increments of 50. The horizontal x-axis has two categories. The first bar, Pre-implementation 12 months, has a value of 220. The second bar, Post-implementation 12 months, has a value of 245. A text label above the bars indicates a plus 11.4% increase.
Right Panel: A bar chart measuring Nominal procedure-fee revenue in J P Y million. The vertical y-axis ranges from 0 to 120 in increments of 20. The horizontal x-axis has two categories. The first bar, Pre-implementation 12 months, has a value of 102.8. The second bar, Post-implementation 12 months, has a value of 114.9. A text label above the bars indicates a plus 11.8% increase.
The original investment application, planned activity, planned CM, committed allocation, service-line variable costs, direct equipment-use logs, and prespecified quality or access indicators were not available in the records assembled for this application. A claims-based volume proxy suggested that shoulder procedures comprised 245 of 255 potentially related arthroscopy procedures, but it did not establish actual equipment use and was retained only as a supplementary sensitivity input. Consequently, actual CM, CMR, IBCR, EPR, and value-gate status were not calculated.
Let a denote the proportion of the observed procedure-fee revenue increase attributable to the equipment, m the CMR on that attributable revenue, ΔR the observed revenue increase, and A
min the minimum documented annualized investment burden (AIB). Break-even requires
$ a\;m\hskip0.45em \varDelta R\ge {A}_{\mathrm{min}} $
.
With A
min = JPY 788,040 and ΔR = JPY 12,081,000, the condition becomes
$ a\;m\hskip0.4em \ge \frac{A_{\mathrm{min}}}{\varDelta R}=0.0652 $
under the full documented annual lease burden.
Equivalently, if all twenty-five additional procedures were attributable to the equipment, the required CM was JPY 31,522 per additional procedure. These are recoverability thresholds, not evidence that the camera caused the observed increase or that the investment recovered its cost. Table 3 summarizes available fields, calculations, and unresolved outputs.
Retrospective field availability and feasible outputs for the selected investment
Table 3. Long description
The table consists of four columns and ten data rows.
1. C S L, technology, and observation periods: Available. Asset/lease record and aggregate claims mapping. Single-investment application and comparable 12-month periods.
2. Contract total and lease term: Available. J P Y 4,728,240; 72 months. Contract burden documented.
3. Minimum documented A I B: Available with limitation. J P Y 788,040 per year. Minimum annual lease component; full A I B may be higher.
4. Procedure activity: Available. 220 before; 245 after. Descriptive difference: plus 25 procedures (+11.4 percent).
5. Nominal procedure-fee revenue: Available with limitation. J P Y 102.8 million before; J P Y 114.9 million after. Secondary descriptive difference; post-period included fee revision.
6. Original planned values and committed allocation: Unavailable in records assembled. No planned activity, C M, review date, or allocation recovered. E P R and completed Tier 2 review not feasible.
7. Variable costs and C M: Unavailable. Procedure-fee revenue only. Actual C M, C M R, and I B C R not feasible.
8. Direct equipment-use log: Unavailable. 245/255 procedure share retained only as sensitivity information. Actual-use allocation not established.
9. Prespecified safety and access indicators: Unavailable. No value-gate fields recovered. Value-gate status not feasible.
10. Recoverability threshold: Feasible. J P Y 31,522 per observed additional procedure; attribution times C M R greater than or equal to 6.52 percent. Threshold only; no causal or recovery claim.
Note: AIB, annualized investment burden; CM, contribution margin; CMR, contribution margin ratio; CSL, clinical service line; EPR, ex-post performance ratio; IBCR, investment burden coverage ratio. Values are aggregate and non-person-level.
Discussion
The study combined method specification with a limited real-world application. Routine records were sufficient to identify the lease, a minimum annual burden, and comparable procedure activity, but they were insufficient to reconstruct the approved investment hypothesis, variable costs, actual-use allocation, or value guardrails. The framework could therefore generate a transparent recoverability threshold and a data-gap map, but not an actual CM-based recovery judgment. This result illustrates why post-adoption evaluation must begin at approval rather than be reconstructed only after performance becomes a concern.
The CSL-IRF fits within LC-HTA as a hospital-level post-adoption monitoring and reassessment module. LC-HTA emphasizes change over time and connected, prospectively sequenced assessment activities (Reference Kirwin, Round, Bond and McCabe4–Reference Pichler, Boysen and Mittmann6). The present application shows the practical consequence when that sequence is absent: contract and activity records remain, while the decision hypothesis and optimization criteria needed for reassessment cannot be recovered. The framework’s contribution is therefore not to replace HB-HTA, mini-HTA, or LC-HTA, but to make their post-adoption evidence commitments operational within hospital accounting and governance.
Value generation extends beyond financial recovery. Patients may benefit through safety, outcomes, access, or shorter waits; professionals may gain or lose workload capacity and technical capability; organizations may strengthen emergency, referral, teaching, or regional functions; and health systems may gain equity or public-service value. A structured post-adoption record can also contribute to corporate social responsibility by making these commitments and results auditable rather than leaving them as unsupported narratives. Environmental sustainability is an emerging HTA domain, but international practice remains heterogeneous and data intensive (Reference Bobini and Cicchetti28). The CSL-IRF therefore allows prespecified environmental measures, such as energy use, consumables, waste, or equipment life, to be reported as a separate value domain without claiming a universal method or converting them automatically into a financial score.
International transferability lies in the procedure and minimum data architecture rather than Japanese thresholds. Hospitals without national claims or equipment surveys can still begin with contracts, asset ledgers, activity counts, revenue, variable costs, use logs, and locally selected value indicators. External public data are an optional plausibility layer. The Japanese facility, cost, claims, and inpatient data sources retained in the supplement illustrate this optional layer rather than mandatory inputs (29–32). Reimbursement structure, accounting treatment, data ownership, and value priorities must be adapted locally, but the sequence remains portable: define the CSL, preserve planned values, state allocation and counterfactual assumptions, measure with the same definitions, apply value guardrails, and return the result to the next decision.
The feasibility application also changes the recommended implementation order. A minimal prospective record should first preserve the CSL definition, planned activity, planned revenue and variable cost, planned CM, documented AIB components, committed allocation, review date, and at least one safety and one access indicator. More complex costing or counterfactual methods can be added later. This staged approach reduces data burden while preventing the specific failure observed here, in which actual recovery metrics could not be reconstructed because the numerator, plan, and guardrails were never stored together.
Several limitations remain. The empirical component was a single, purposively selected case from one hospital and was not a census of investments. The before-and-after comparison was descriptive, included a reimbursement revision in the post-period, and did not control for staffing, referral, scheduling, or other concurrent changes. Procedure-fee revenue was not total service-line revenue, variable costs were unavailable, and the full AIB may exceed the documented annual lease component. No direct equipment-use log or patient outcome was used. The study did not evaluate face validity through an expert panel, user acceptability, threshold calibration, decision impact, or clinical and organizational outcomes. Finally, targeted conceptual mapping was not a systematic review. Multicenter retrospective and prospective studies should next test field completion, calculation reproducibility, counterfactual feasibility, burden, fairness, and whether the framework changes renewal or expansion decisions.
Conclusions
The CSL-IRF specifies an auditable hospital process that connects a CSL-level adoption hypothesis to annualized burden, margin, allocation, value guardrails, reassessment, and the next capital decision. Its initial real-world application showed that routine records can support part of the process, while missing planned values, variable costs, use logs, and value indicators prevent a complete recovery assessment. Prospective preservation of these fields is therefore a prerequisite for credible post-adoption and lifecycle HTA. The study does not establish that the selected investment caused higher activity or recovered its cost.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0266462326104358.
Data availability statement
The aggregate service-line summary, field-availability audit, threshold calculations, public-data extract, synthetic scenario inputs, and Python code used to generate the figures are provided in Supplementary File 2. Source contracts and internal administrative records are not publicly shared because they contain commercially and operationally sensitive information. No patient-level data were used.
Acknowledgments
The authors thank the public institutions that maintain the aggregate contextual data sources used in the framework.
Author contribution
T.S. and N.K. contributed to the conception and design of the framework; T.S. contributed to source mapping, aggregate hospital-record extraction and verification, calculations, data curation, visualization, supplementary material preparation, and drafting; N.K. contributed to clinical and hospital-management interpretation, methodology refinement, supervision, and substantive editing. Both authors reviewed and approved the final revised version and agree to be accountable for all aspects of the work.
Funding statement
This research received no specific grant from any funding agency, commercial, or not-for-profit sector.
Competing interests
The authors declare none.
Ethical standards
The retrospective feasibility application used only aggregate, non-person-level service-line activity, procedure-fee, asset, and lease data and involved no human participants, patient identifiers, clinical outcomes, individual-record linkage, or participant contact. Research ethics committee review and informed consent were therefore not applicable.
Use of artificial intelligence tools
OpenAI ChatGPT was used for English translation support and editorial drafting. The authors selected and verified the sources and hospital data, checked the calculations, revised the text, and take responsibility for the interpretation, references, figures, code, and final manuscript. Artificial intelligence tools were not treated as authors and did not generate or alter empirical hospital data.
Open access
