Patient-Based Costing in Belgian Hospitals: Building ABC and TDABC from MZG/RHM and FINHOSTA - To Be

Introduction

Patient-based costing seeks to measure the cost of care at the level of the individual patient rather than by broad departmental or specialty averages. In methodological terms, activity-based costing (ABC) assigns costs to activities and then to cost objects through cost drivers, whereas time-driven activity-based costing (TDABC) simplifies this logic by combining two core parameters: the cost of supplying resource capacity and the time required to perform the activities in the care process. In healthcare, TDABC has become especially relevant because it is explicitly tied to process mapping and the measurement of real care delivery across pathways.

For Belgian hospitals, the policy relevance of patient-level costing is unusually high. Belgium still allocates a major part of hospital financing through an APR-DRG/SOI-based budget allocation whose weights are based on average length of stay (LOS) rather than cost weights. The Belgian Health Care Knowledge Centre (KCE) has stated that length of stay is only a proxy for resource use and that Belgium currently has no compulsory collection of patient-level cost data. That makes the development of hospital-level patient costing not merely a managerial innovation, but an essential precondition for a more cost-based and transparent payment architecture.

The central conclusion of this essay is straightforward. Belgian patient-based costing should start from Minimale Ziekenhuisgegevens (MZG)/Résumé Hospitalier Minimum (RHM) and FINHOSTA, but it cannot end there. MZG/RHM provides the clinical and stay-level spine of the costing system; FINHOSTA provides the standardized financial, cost-centre, allocation-key and personnel-cost structure. Together they are necessary. However, the evidence does not support the claim that they are sufficient on their own for full ABC, and still less for TDABC. Full implementation requires additional local feeder systems and, for TDABC, explicit measurement of time and practical capacity. This conclusion is a grounded implementation inference from the official descriptions of MZG/RHM and FINHOSTA, the KCE reports on Belgian costing, and Belgian TDABC studies that had to collect extra activity, material, equipment and time data locally.

1. The Belgian data foundation: what MZG/RHM and FINHOSTA already provide

The Minimale Ziekenhuisgegevens (MZG)/Résumé Hospitalier Minimum (RHM) is a mandatory federal registration for all non-psychiatric hospitals in Belgium. The 2026 official introduction explains that since 2008 the MZG/RHM integrates the former clinical and nursing minimum datasets (MKG, MVG) and is transmitted twice a year. It is structured into 27 files across six domains: structure, personnel, administrative data, nursing data, medical data and billing data. The files which could be most relevant for costing include patient and stay data (PATHOSPI, STAYHOSP, STAYSPEC, STAYUNIT), nursing items (ITEMDIVG), diagnoses (DIAGNOSE), procedures (PROCEDUR), INAMI prestations (PROCRIZI), test results (TESTRESU), emergency admissions (URGADMIN), and staffing files (EMPLOPER, EMPLODAY). At this moment, not all of these files are being used in a way they could make a contribution.  The MZG/RHM manual states that STAYHOSP is the central file for stay information and that other files may contain one, many or no records per stay. It also notes that the hospital information system can be coupled to local subsystems such as Admission, Discharge, and Transfer (ADT), radiology and laboratory (Laboratory Information Management Systems) systems .

For patient-level costing, this means that MZG/RHM is the natural clinical episode layer. It can identify the patient stay, classify the case, describe diagnoses (ICD-10-CM) and procedures (ICD-10-PCS), and partially describe nursing workload and patient trajectory through specialties and care units (VG-MZG). The nursing circular on VG-MZG is especially relevant because it states that nursing data are used for nursing-care databases, staffing systems, nursing financing, quality processes and the measurement of nursing workload. That makes VG-MZG highly valuable as an activity proxy for nursing-intensive costing models, even though it is not itself a full cost file.

The second indispensable Belgian source is FINHOSTA, the federal collection of hospital financial, personnel and operational data. The FPS Public Health describes FINHOSTA as a mandatory standardized transmission for Belgian hospitals and explicitly states that hospitals must encode accounting data, charges by hospital service, medical and nursing activity, investments and personnel data; submission is a condition for obtaining the Budget of Financial Means (/BFM) or Budget van Financiële Middelen (BFM)/Budget des Moyens Financiers (BMF).

From a costing perspective, the KCE’s 2024 report on indirect hospital costs is especially informative. It identifies FINHOSTA Table 2 (analytical accounting: costs and revenues by cost account and cost centre/Analytische boekhouding (klasse 6 en 7)), Table 3 (allocation keys that distribute temporary cost centres toward other cost centres/Verdeelsleutels en budgettair berekende oppervlaktes van de ziekenhuizen met acuut type), and Table 13 (personnel expenses and personnel data by cost centre/Lasten en personeelsgegevens per kostenplaats) as the most important FINHOSTA tables for cost analysis. The same report explains that allocation keys may use drivers such as square metres, FTE staff numbers and nursing days, and that the dataset for 2019 covered all 104 Belgian hospitals.

In short, MZG/RHM tells the hospital what happened to which patient; FINHOSTA tells the hospital where costs sit in the accounting structure and how common costs are allocated across cost centres. That is precisely why these two sources must be the starting point for Belgian patient-level costing.

Note: Cost centers are departments (e.g., HR, IT) that only incur expenses, focusing on budget compliance, while profit centers are business units (e.g., ER/ED, OR, ward, ...) that generate revenue and costs, held accountable for their net profit. Cost centers are cost-control units; profit centers are performance-tracking units. 

Note: Admission, Discharge, and Transfer (ADT) systems are essential healthcare information technology used to track patient movements and data from arrival to departure, serving as the backbone for hospital workflow. They provide real-time updates on patient status, location, and demographics, improving clinical efficiency, bed management, and care coordination.

2. Why MZG/RHM and FINHOSTA are necessary but not sufficient

The main limitation of relying on these two datasets alone is that they do not yet amount to a true patient-level microcosting system. KCE is explicit that FINHOSTA was never set up to calculate the costs of specific medical activities, for example at the level of a nomenclature code, and that the smallest practical unit for many analyses is the cost centre, not the patient or the activity. KCE also emphasizes that there is currently no compulsory Belgian patient-level cost collection.

MZG/RHM has a different limitation. It is a minimum dataset, transmitted only semi-annually, and although it captures diagnoses, procedures, stay trajectory and some nursing information, it does not by itself provide the granular time stamps, direct material consumption, equipment usage, or department-specific activity frequencies needed for bottom-up ABC or TDABC. This is not conjecture: Belgian TDABC studies did not rely on MZG/RHM alone. The KCE multicentre radiotherapy study had to collect local information on activity lists, time registrations, equipment costs, direct and indirect materials, activity consumption by treatment, patient volumes and fractions. That evidences the gap between national minimum registration and actual microcosting requirements.

A further Belgian complication is physician cost measurement. KCE’s manual for cost-based pricing notes that overhead rates derived from FINHOSTA exclude physician costs and that physician inputs are not recorded consistently across Belgian hospitals. The 2024 KCE report likewise describes variability in how physician fees and deductions are recorded. Therefore, hospitals that want accurate pathway costs must usually supplement FINHOSTA with local medical-board, fee-pool or specialty-level physician cost data.

The implication is clear: MZG/RHM + FINHOSTA is the mandatory backbone, but a hospital that stops there will build, at best, a hybrid case-costing system with many proxies. A robust Belgian patient-based costing architecture must add local transactional data and, for TDABC, time and capacity measurement.

3. Which dataset is required for ABC and TDABC in Belgian hospitals?

The correct answer is that no single Belgian dataset is sufficient. What is required is a linked dataset architecture.

For an initial ABC system, the required data stack has four layers. First is the clinical episode layer, built from MZG/RHM: STAYHOSP as the core stay record, PATHOSPI for patient-level attributes, stay trajectory files such as STAYSPEC and STAYUNIT, diagnosis and procedure files, emergency files, and nursing items. Second is the financial and resource-cost layer, built from FINHOSTA: especially Table 2 for analytical costs by cost centre and account, Table 3 for allocation keys, and Table 13 for personnel expenses and staffing data. Third is a local feeder-system layer that traces resource use more directly to patients or activities: pharmacy dispensing, implant and consumables systems, laboratory and radiology orders, operating-room (OR) and anaesthesia systems, ICU and ward bed-movement systems (ADT), rehabilitation and other ancillary systems, and local billing or claims extracts where relevant. Fourth is a costing master-data layer: payroll, physician-cost files, fixed-asset registers, equipment depreciation data, and space/capacity data used for overhead allocation. The need for these additional local data follows from the fact that official MZG/RHM can be coupled to hospital subsystems, while Belgian TDABC studies had to collect detailed activity, equipment, direct material and activity-consumption data that are not present in FINHOSTA or MZG/RHM themselves.

For TDABC, the dataset must go one step further. In addition to the ABC stack, the hospital needs: a standardized process-map library for each targeted care pathway; time measurements or validated time estimates for each significant activity; practical capacity data for staff and equipment; and the ability to model time equations for pathway variations. The healthcare TDABC literature consistently defines TDABC around capacity cost rates and the time needed to perform activities, and the Belgian radiotherapy study operationalized this by collecting actual time registrations during a study period, extrapolating missing time data, and then translating staff and equipment capacity into hourly or per-minute costs.)

Accordingly, the most accurate formulation is this: to implement ABC in Belgian hospitals, the minimum required dataset is MZG/RHM + FINHOSTA + patient-level feeder systems for direct resource use; to implement TDABC, the required dataset is all of the above plus process maps, timestamps or direct observations, and practical-capacity data enabling capacity cost rates.

Note: Activity-Based Management (ABM) is a discipline that focuses on managing activities to improve operational efficiency, reduce costs, balance workload, and increase customer value. It uses Activity-Based Costing (ABC) data to analyze the workload, profitability of products, services, and customers, allowing managers to identify non-value-added activities, workload imbalance, and optimize processes. 

4. How Belgian hospitals could develop patient-based costing from MZG/RHM and FINHOSTA

A feasible Belgian implementation should proceed in stages.

The first stage is to define the cost object. In the Belgian context, the cleanest initial cost object is the hospital stay anchored on the MZG/RHM STAYHOSP record, because STAYHOSP is the central stay file to which other RHM files can be linked. Hospitals can then decide whether the analytic unit remains the stay, or whether it is extended to a broader episode such as a 30-day surgical pathway or an oncology pathway spanning diagnosis, treatment and follow-up. 

The second stage is to build a hybrid ABC model. Here, MZG/RHM is used to classify and stratify cases (APR-DRG/SOI/ROM), while FINHOSTA supplies the reconciled departmental cost pools. Local feeder systems then trace high-value direct costs to the patient wherever possible: pharmaceuticals, implants, expensive disposables, blood products, imaging, laboratory tests, operating theatre time (OR) and ward days (ADT). Residual cost-centre costs can then be allocated through explicit activity drivers derived from local systems, such as number of administrations, minutes in operating theatre (OR), imaging sessions (RX, CT, MRI), nurse interventions, or bed-days. In Belgian practice, this is the most feasible first step because FINHOSTA already provides the standardized cost-centre and allocation-key environment, while KCE’s costing manual provides Belgian unit-cost guidance for personnel, overhead and equipment principles.

The third stage is to upgrade targeted care pathways from ABC to TDABC. Hospitals should not attempt full TDABC everywhere at once. They should start with a limited number of high-volume or high-cost pathways where clinical teams are engaged and process variation matters - for example, childbirth, oncology, joint replacement or other well-defined pathways (e.g. focus clinics). Belgian evidence supports this staged approach. The radiotherapy KCE study mapped activities through interviews, collected time registrations and local resource data, and then calculated treatment costs bottom-up. Later Belgian work showed that TDABC can also be extended to national policy modelling without complete transactional data at national level, but that still presupposes carefully developed pathway models and Belgian inputs.

The fourth stage is reconciliation and validation. A Belgian hospital-level costing system is not credible unless the sum of patient-level costs can be reconciled back to hospital accounts and, where relevant, to FINHOSTA totals. Validation must therefore include financial reconciliation, clinician review of pathway maps and timings, and stability checks by APR-DRG/SOI and site. Without reconciliation, hospitals risk producing elegant but non-auditable numbers. This concern is especially important in Belgium because KCE has documented substantial inter-hospital variation in how costs are registered within the FINHOSTA framework.

5. How patient-based costing can identify inefficiency, reduce costs and improve care quality

The first analytical use of patient-based costing is variance decomposition (Statistical Process Control (SPC)). Belgian hospitals already classify cases by APR-DRG/SOI, but KCE has emphasized that length of stay (LOS) is only a proxy for resource use. Patient-level costing allows hospitals to compare patients within the same APR-DRG/SOI and ask a better question: how much of the remaining cost variation is explained by disease severity, patient complexity, or pathway design? Belgian research demonstrates why that distinction matters. In early-stage invasive breast cancer, TDABC showed that disease-related characteristics were major drivers of cost variability; in childbirth, TDABC showed meaningful variation both between and within delivery types. Therefore, cost variation should not automatically be labelled inefficiency; part of it is clinically warranted and must be separated from operational waste.

The second use is process-level waste identification. TDABC is especially valuable because it makes cost visible at each process step. In the Belgian KCE radiotherapy study, the distribution of personnel-and-equipment cost across steps such as first patient contact and simulation differed materially across centres, even when total costs were relatively similar. The report explicitly interpreted some of those differences as process differences rather than case-mix differences - for example, explanation time shifting from consultation to simulation. That is exactly the kind of insight hospitals need if they want to reduce duplication, unnecessary hand-offs and staffing mismatches without harming care.

The third use is to identify unused or misallocated capacity. TDABC literature emphasizes practical capacity, capacity cost rates and time equations. In health care, that means hospitals can estimate not only the cost of care delivered, but also whether staff and equipment are configured appropriately for demand. This matters in Belgium because common FINHOSTA allocation rules such as square metres, FTEs and nursing days do not always reflect the real consumption pattern and workload of specific services. A patient-level system can therefore replace crude departmental proxies with pathway-specific evidence about where capacity is underused, overused or poorly matched to patient complexity (workload balancing, burnout prevention, Quintuple Aim).

The fourth use is to support cost reduction without quality blindness. The childbirth study concluded that TDABC can inform process improvement and cost-containment initiatives; more broadly, recent reviews describe TDABC as improving transparency and decision-making by mapping process steps and exposing bottlenecks. Yet the managerial lesson is not to cut the cheapest line item. The correct use is to identify high-cost steps with low clinical value, to standardize where variation is unwarranted, and to preserve or expand resource use where outcomes justify it. In practical hospital governance, this means cost dashboards must be paired with outcome indicators such as mortality, readmission, complications, patient experience and, where available, PROMs. Otherwise, a hospital may reduce accounting cost at the expense of workload imbalance, process deterioration, poorer outcomes or more downstream utilization (Quintuple Aim).

Finally, patient-based costing can strengthen Belgian financing reform. KCE has repeatedly linked the move toward more cost-based hospital payment to the need for better underlying cost data. Once hospitals can produce reconciled patient-level costs, Belgium will be in a stronger position to derive more defensible tariffs, distinguish direct from indirect costs more transparently, and move beyond the current dependence on length of stay (LOS) as a proxy in combination with complex compensation mechanisms for its shortcomings. Hospital-level patient costing is therefore both a management tool and a preparatory infrastructure for payment reform.

Note: Process mining for care pathways uses event logs from hospital IT systems (EHR, Patient Data Management System (PDMS), LIMS, Enterprise Resource Planning (ERP), ...) to automatically reconstruct, visualize, and analyze actual patient journeys, moving beyond subjective, manual mapping to evidence-based insights. This data-driven approach identifies bottlenecks, reduces unwarranted variations, improves compliance with clinical guidelines, and enhances resource allocation.

Note: Statistical Process Control (SPC) is a data-driven, scientific method for monitoring, controlling, and improving processes by reducing variation using statistical tools, primarily control charts. By distinguishing between natural variations and assignable causes (defects), SPC allows for real-time, proactive adjustments to maintain quality and boost efficiency.

Note: The Quintuple Aim is a widely accepted framework in healthcare, evolving from the Triple and Quadruple Aims, designed to optimize health system performance. It aims to improve patient care, population health, and clinician experience while reducing costs and advancing health equity. Patient-based costing (ABC, TDABC) should be embedded in the Quintuple Aim.

Conclusion

Belgian hospitals should develop patient-based costing as a linked, staged architecture rather than as a single database project. MZG/RHM should provide the patient, stay, diagnosis, procedure and nursing-activity spine; FINHOSTA should provide the standardized financial, cost-centre, allocation-key and personnel-cost structure. On top of that backbone, hospitals must add local feeder systems for direct patient resource use. If the goal is ABC, that combined architecture is sufficient to produce a robust hybrid patient-costing model. If the goal is TDABC, hospitals must add care process maps, time data and practical-capacity information so that resource costs can be expressed as capacity cost rates multiplied by actual time consumption. 

Thus, the answer to the core question is precise: the required dataset for Belgian patient-based costing is not MZG/RHM alone, nor FINHOSTA alone, nor even the combination of the two in isolation. The required dataset is a linked patient-level costing platform whose backbone is MZG/RHM + FINHOSTA, and whose operational completeness depends on local transactional, staffing, asset and timing data. Used in that way, patient-level costing can separate warranted clinical complexity from unwarranted process variation, reveal unused capacity, support more disciplined (realistic) resource allocation, and reduce costs in ways that are compatible with a more balanced workload and patient care (Quintuple Aim).

Bibliography

Belgian Health Care Knowledge Centre (KCE). Indirect hospital costs as a mark-up percentage on direct costs in Belgium. KCE Report 380C. Brussels: KCE; 2024.

Swartenbroekx N, Obyn C, Guillaume P, Lona M, Cleemput I. Manual for cost-based pricing of hospital interventions. KCE Report 178C. Brussels: KCE; 2012.

Hulstaert F, Mertens A-S, Obyn C, Van Halewyck D, Van Der Straten B, Lievens Y. Innovative radiotherapy techniques: a multicentre time-driven activity-based costing study. KCE Report 198C. Brussels: KCE; 2013.

FPS Public Health. Minimale Ziekenhuis Gegevens (MZG)/Résumé Hospitalier Minimum (RHM). Brussels: FPS Public Health; 2026.

FPS Public Health. Inleiding MZG/Introduction RHM, version Feb. 2026. Brussels: FPS Public Health; 2026. 

FPS Public Health. Finhosta: onmisbare inzamelingen voor Belgische ziekenhuizen/Finhosta: collectes indispensables pour les hôpitaux belges. Brussels: FPS Public Health; 2026.

Canby, J. B. (1995). Applying activity-based costing to healthcare settings. Healthcare financial management: journal of the Healthcare Financial Management Association, 49(2), 50-2.

da Silva Etges, A. P. B., Ruschel, K. B., Polanczyk, C. A., & Urman, R. D. (2020). Advances in value-based healthcare by the application of time-driven activity-based costing for inpatient management: a systematic review. Value in Health, 23(6), 812-823.

Defourny, N., Hoozée, S., Daisne, J. F., & Lievens, Y. (2023). Developing time-driven activity-based costing at the national level to support policy recommendations for radiation oncology in Belgium. Journal of Accounting and Public Policy, 42(1), 107013.

Dubron, K., Verschaeve, M., & Roodhooft, F. (2021). A time-driven activity-based costing approach for identifying variability in costs of childbirth between and within types of delivery. BMC Pregnancy and Childbirth, 21(1), 705.

Etges, A. P. B. D. S., Polanczyk, C. A., & Urman, R. D. (2020). A standardized framework to evaluate the quality of studies using TDABC in healthcare: the TDABC in Healthcare Consortium Consensus Statement. BMC health services research, 20(1), 1107.

Kaplan, R. S., Witkowski, M., Abbott, M., Guzman, A. B., Higgins, L. D., Meara, J. G., ... & Feeley, T. W. (2014). Using time-driven activity-based costing to identify value improvement opportunities in healthcare. Journal of Healthcare Management, 59(6), 399-412.

Keel, G., Savage, C., Rafiq, M., & Mazzocato, P. (2017). Time-driven activity-based costing in health care: a systematic review of the literature. Health policy, 121(7), 755-763.

Nundy, S., Cooper, L. A., & Mate, K. S. (2022). The quintuple aim for health care improvement: a new imperative to advance health equity. Jama, 327(6), 521-522.

Roman, E., Cardoen, B., Decloedt, J., & Roodhooft, F. (2020). Variability in hospital treatment costs: a time-driven activity-based costing approach for early-stage invasive breast cancer patients. BMJ open, 10(7), e035389.

Ross, T. K. (2004). Analyzing health care operations using ABC. Journal of health care finance, 30(3), 1-20.