Feasibility of micro-costing for hospital-at-home in Danish municipalities: a prospective pilot study Authors: Nasrin Tayyari, Iben Duvald, Malene Galle Madsen, Sigrid Engelbrecht Sjøl, Camilla Palmhøj Nielsen, Bettina Wulff Risør Published: 2025 ABSTRACT Abstract AbstractObjectivesTo test the feasibility of identifying and quantifying resource use for a Hospital-at-Home (HaH) model in Danish municipalities, we used a micro-costing approach. Additionally, we aimed to generate a transparent activity and time dataset. This dataset will support subsequent tariff development with time-driven activity-based costing and feed into the economic evaluation of an ongoing randomised controlled trial (RCT).DesignProspective pilot feasibility study.SettingThree municipalities in the Central Denmark Region in collaboration with emergency department specialists and general practitioners.Participants56 elderly acute patients treated in HaH during the pilot phase.Outcome measuresFeasibility of micro-costing data collection (completeness, consistency and acceptability to staff) and descriptive resource-use quantities by activity and provider group. No price assignment or cost estimates are reported.ResultsPatients received a mean of 3.8 HaH treatment days with 7.8 acute team visits and 3.9 municipal-staff visits per treatment course. The acute team spent a mean of 742 min per patient across treatment activities, communication, documentation and transport, while municipal care staff recorded a mean of 213 min. Intravenous medicine administration and vital sign assessments were the most frequent activities. Data completeness and consistency improved over time through co-design and feedback.ConclusionsDetailed resource-use measurement using provider logs was feasible in a municipal HaH model and produced an activity and time dataset suitable for tariff development. Findings are context-specific and not generalisable due to the small sample. The micro-costing log refined through the pilot will be applied in an RCT, where time and activity data will be used to construct a tariff using time-driven activity-based costing.Keywords: HEALTH ECONOMICS, Patients, Health Care Costs STRENGTHS AND LIMITATIONS OF THIS STUDY Co-designed micro-costing logs enabled detailed provider-reported activity and time capture in routine care. Co-designed micro-costing logs enabled detailed provider-reported activity and time capture in routine care. Pre-specified procedures supported completeness, consistency and resolution of boundary cases. Pre-specified procedures supported completeness, consistency and resolution of boundary cases. Results are limited to resource-use quantification; no prices or cost estimates are presented. Results are limited to resource-use quantification; no prices or cost estimates are presented. Reliance on staff self-report may introduce measurement bias relative to independent observation. Reliance on staff self-report may introduce measurement bias relative to independent observation. Small single-region sample limits generalisability and may require local adaptation for transferability. Small single-region sample limits generalisability and may require local adaptation for transferability. Introduction The changing demographics of Denmark, with an increasing population of individuals aged 65 and above, have significant implications for healthcare demand.1 This demographic shift is expected to increase the prevalence of chronic disease and, consequently, the need for healthcare services, including hospital admissions.1,3 Danish hospitals already face challenges related to capacity constraints and staff shortages; therefore, national strategies aim to reduce hospital admissions by enabling specialised treatment to be delivered in primary care settings, closer to or within the patient’s home.4 5 Achieving this ambition requires new and innovative service models, as the current healthcare structure is not yet optimally designed to support advanced treatment outside the hospital environment. Hospital-at-Home (HaH) models represent one such strategy, providing hospital-level treatment in the home environment through collaboration between municipal acute teams and hospital physicians.6,11 Evidence indicates that HaH can improve patient outcomes and caregiver experiences and reduce healthcare utilisation and readmissions.612,14 International studies also suggest that HaH can deliver equivalent or better outcomes at lower overall cost compared with inpatient care.6 12 13 However, the model’s benefits depend on consistent implementation and sufficient clinical oversight. Inadequate coordination or delayed recognition of patient deterioration can increase the risk of complications and compromise safety.8 Moreover, HaH services may become inefficient if patient volumes are low and staff time is under-utilised, or if extensive travel time offsets gains achieved elsewhere.15,17 Delivering HaH in Denmark is inherently complex because it requires cross-sectoral coordination. The state provides overall regulation, the five regions are responsible for hospitals and general practitioners and the 98 municipalities provide home nursing and long-term care.4 Municipal acute teams, staffed by specialised nurses, are financed at municipal level but collaborate closely with regional hospitals.5 Yet municipalities do not routinely collect activity-level resource data, meaning that conventional costing approaches based on aggregated expenditure are insufficient for accurately capturing resource use in HaH delivery. Micro-costing is recommended for new and complex interventions where resource use and costs are uncertain.18,20 This method relies on direct measurement of activities, time and material use, typically through detailed observation or structured logs. Although frameworks for micro-costing hospital-based services are well established, guidance for municipal-level or cross-sectoral services remains scarce. Without transparent micro-costing data, it is difficult to construct tariffs or perform robust cost-effectiveness analyses of integrated care initiatives that span sectors. To address this gap, we conducted a prospective pilot feasibility study to test whether activity-level resource use can be identified and quantified for elderly acute patients treated in a municipal HaH model aiming to subsequently inform the economic evaluation of an ongoing randomised controlled trial (RCT) comparing HaH with standard hospital admission.21 Publishing the findings of the pilot study ahead of the ongoing study is justified because the pilot directly informed the measurement strategy (eg, simplified logging of service counts plus total time, clarified boundary cases and provider-specific communication categories) and provides the validated activity/time inputs needed to construct a transparent tariff based on Time Driven Activity Costing (TDABC). This tariff will be applied in the economic evaluation of the RCT to relate resource use to outcomes in a cost-effectiveness framework. We emphasise that this study reports resource-use quantities only: no valuation or costing results are presented. Research question Is it feasible to identify and quantify resource use for elderly acute patients treated in HaH in Danish municipalities using a micro-costing approach? Methods Study design This was a prospective pilot feasibility study of a newly developed HaH model. The aim was to evaluate the new patient pathway and to test the feasibility and validity of a micro-costing log for detailed resource-use measurement ahead of a larger RCT running from June 2022 to 31 July 2026.21 Setting The study was conducted in three municipalities in the Central Denmark Region in collaboration with general practitioners (GPs) and the Emergency Department (ED) at the regional hospital. Because the distribution of responsibilities across provider groups is central to how resource use is generated in the HaH model, we describe the roles of the involved staff below. Treatment delivery, acute team nurses Acute teams are municipal units staffed with specialised nurses who deliver acute nursing and selected hospital treatments in patients’ homes. In HaH, they work under the clinical responsibility of an ED physician who supervises treatment and conducts daily ward rounds, and they coordinate cross-sector communication as required. Treatment delivery, municipal home nurses Municipal home nurses provide general clinical treatment under delegation in accordance with agreed competence standards. They support ongoing treatment at home, coordinate visits with the acute team and, when relevant, communicate clinical observations to the acute team or the ED physician. Care delivery, municipal care staff Social and healthcare assistants and helpers provide personal care and practical help to support patients during the treatment course at home. Within their training and local policies, they may perform delegated clinical activities. All activities are documented in municipal records, and concerns are escalated to nursing staff when needed. Participants and recruitment The initial study protocol described a pilot conducted in two municipalities and included 44 patients from October 2020 to June 2022.21 After publication of the protocol article, a third municipality joined the project, applying the same inclusion criteria and study procedures. Pilot recruitment in this municipality took place between August 2022 and January 2023, partly overlapping with the initiation of the RCT in the first two municipalities. Thus, while the study began enrolling patients based on the refined protocol, the third municipality continued pilot-phase data collection to ensure implementation of the micro-costing log. Thus, the discrepancy in patient numbers between the published protocol and the present study reflects this extension of the pilot before full implementation. Pilot enrolment and data collection therefore occurred from October 2020 to January 2023 across the three participating municipalities. 56 patients were enrolled; one patient had incomplete logs and was excluded from analysis, leaving 55 for the final dataset. ‘Treatment course’ was defined as the period from HaH admission to discharge, and logs covered the full course prospectively. The pilot’s lessons were incorporated into the micro-costing log, prior to RCT enrolment. Eligible patients were aged 65 years or older, living at home or in a nursing home in the participating municipalities, with an acute condition considered stable enough for HaH management, and able to provide informed consent. Examples of index conditions were cystitis, erysipelas and pneumonia. Patients were excluded if they were unable to give consent or if the acute team lacked capacity at the time of referral. Patients were identified by the referring GP or ED physician who jointly assessed HaH eligibility before notifying the acute team. Micro-costing approach Resource-use measurement followed an activity-based micro-costing framework comprising three stages: identification of relevant activities and resources, measurement of activity counts and time and valuation.20 In this pilot, we carried out only the first two stages. Price assignment and costing are not reported here. Valuation will be undertaken subsequently using a TDABC approach to construct a tariff, which will then be applied in the economic evaluation.22,24 Data collection A micro-costing log was specifically developed for this study by the research team in close collaboration with acute team nurses and the project group. The log was refined through several iterative cycles: early drafts were tested in practice, data from the first 5 and 10 patients were scrutinised in detail and feedback was obtained from both frontline staff and managers. Each revision addressed ambiguities, boundary cases and user-friendliness, and the final version was approved by the project steering committee. This co-design process ensured both validity and applicability of the instrument in everyday clinical practice. A detailed description of the iterative development steps is provided in online supplemental table S1, and the final micro-costing log is provided in online supplemental file 2. The log captured: Staff role and shift (day, evening or night). Staff role and shift (day, evening or night). Treatment activities performed by the acute team (vital sign assessment, intravenous medicine administration, venous blood sampling, C-reactive protein test, electrocardiography, catheterisation, patient assessment, collection and delivery of medication and samples and related transport). Treatment activities performed by the acute team (vital sign assessment, intravenous medicine administration, venous blood sampling, C-reactive protein test, electrocardiography, catheterisation, patient assessment, collection and delivery of medication and samples and related transport). Activities performed by municipal staff (intravenous and oral medicine administration as delegated, personal care, practical help, vital sign assessment where relevant, documentation and transport). Activities performed by municipal staff (intravenous and oral medicine administration as delegated, personal care, practical help, vital sign assessment where relevant, documentation and transport). Time spent on documentation and on communication with ED physicians, GPs, other nurses, patients and relatives. Time spent on documentation and on communication with ED physicians, GPs, other nurses, patients and relatives. Start and end times of home visits, transportation time and additional time for logistics such as collecting medication or delivering samples. Start and end times of home visits, transportation time and additional time for logistics such as collecting medication or delivering samples. A longer list of activities and operational definitions is presented in online supplemental table S3. Completed logs were collected regularly and transcribed into a structured spreadsheet on a secure server. Staff received training and ongoing support. Data quality To ensure data quality, we provided clear definitions and examples of boundary cases, checked early logs in detail and gave feedback to staff on missing or inconsistent information. This process reduced variation and improved completeness over time. Analysis Resource-use data were summarised descriptively with means, SD and minimum/maximum values reported as the mean of the five lowest/highest patient-level values, respectively, to comply with data security rules. Results are reported both as averages across all patients and, where relevant, conditional averages among those who received the service. Patients who did not receive a given service were assigned a value of zero, reflecting genuine absence rather than missing information. Zero values therefore represent true non-use of the service and were retained in the analysis. The data represent effective working time only; non-patient-related activities, slack time and breaks were not recorded. These elements will be incorporated during the subsequent tariff development.25 Analyses were performed in Stata V.18. Patient and public involvement During the pilot phase, four in-depth semi-structured interviews were conducted to explore patients’ and relatives’ preferences concerning treatment site. Findings from the interviews were discussed during the task force and steering group meetings to adjust the treatment pathway.21 Patients were not involved in developing the micro-costing instrument or in data analysis for this pilot study. We acknowledge that the extent of patient and public involvement was limited. Although no patient or public representatives were involved, municipal acute team nurses were extensively engaged during the pilot through co-design and iterative refinement of the micro-costing log, including testing of early versions in practice and participation in feedback sessions. Results Patient exposure and visits 55 patients were included in the final analyses (one excluded due to incomplete logs). On average, patients received 3.8 treatment days in HaH and 7.5 visits from the acute team per treatment course (table 1). In addition, 27 patients (49%) received visits from a municipal home nurse and 14 patients (25%) from municipal care staff, while 10 patients (18%) received visits from both groups. These visits were in addition to the acute team visits. Patients who received visits from municipal home nurses had a mean of 4.4 treatment days, compared with 3.5 days for those with visits from municipal care staff. Table 1Number of patients, visits by different staff and treatment daysVariableNumber of visitsMean per patientSDMinMaxNumber of visits by staff type and number of treatment daysn=55 Visits by the acute team4127.54.2216 Visits by municipal home nurse (n=27)1072.03.1010 Visits by municipal care staff (n=14)1092.05.2015 Treatment days (n=55)NA3.82.219Subgroups (patients with visits additional to visits by the acute team)Patients with visits by municipal home nursen=27 (49% of the included patients) Visits by municipal home nurse1074.03.4110 Treatment daysNA4.42.338Patients with visits by municipal home caren=14 (25% of the included patients) Visits by municipal care staff1097.88.0215 Treatment daysNA3.51.626Patients with visits by municipal home nurse and home caren=10 (18% of the included patients) Sum of visits by municipal home nurse and care staff13713.711.5523 Treatment daysNA4.01.535Max, maximum; Min, minimum; n, number of patients/observations; NA, not applicable. Number of patients, visits by different staff and treatment days Max, maximum; Min, minimum; n, number of patients/observations; NA, not applicable. Most acute team visits (56%) occurred during day shifts, 30% during evenings and 7% during nights; for 7% the shift was not recorded. For municipal staff, 44% of visits occurred in daytime, 43% in the evening and 12% at night. Services delivered by provider group Table 2 shows the number of services delivered by staff group. Intravenous medication and vital sign assessments were the most frequent acute team services, with averages of 4.6 and 3.1 per patient, respectively. Venous blood sampling, C-Reactive Protein tests and patient evaluations occurred less frequently. Table 2Number of services provided by different staffVariableMean per patientSDMinMaxNumber of treatment services provided by the acute teamn=55 Medicine and intravenous treatment4.64.1014 Vital sign assessment3.11.90.25 CRP Test1.51.304 Venous blood sample1.61.204 Intravenous cannula (venflon)1.21.003 Patient evaluation1.51.405 Other services1.42.006Number of treatment services provided by the municipal home nursen=27 Medicine and intravenous treatment3.540.411 Vital sign assessment0.91.604 Patient assessment and evaluation0.30.601Number of care services provided by the municipal care staffn=14 Personal care and practical help13.514.02.623 Vital sign assessment0.41.101 Patient assessment and evaluation1.32.404CRP, C-reactive protein; Max, maximum; Min, minimum; n, number of patients/observations. Number of services provided by different staff CRP, C-reactive protein; Max, maximum; Min, minimum; n, number of patients/observations. For municipal home nurses and municipal care staff, results are reported only for the subgroups of patients who actually received such visits (n=27 and n=14, respectively). Municipal home nurses most often administered intravenous medicine (mean 3.5 per patient in this subgroup) and performed vital sign assessments. Municipal care staff most frequently provided personal care and practical help (mean 13.5 per patient in this subgroup). Zeros in the dataset represent patients who did not receive these services, reflecting genuine variation in need. Communication Table 3 summarises time spent on communication activities. Acute team nurses spent the largest share of communication time with ED physicians (mean 43 min per patient) and with patients and relatives (14 min). Communication with municipal nurses and care staff was also frequent. Municipal home nurses and municipal care staff primarily communicated with acute team nurses and with patients or relatives. Subgroup analyses are presented in online supplemental table S4. Table 3Time spent on communication (minutes) per treatment courseVariableMean per patient (n=55)SDMinMaxAcute nurse time spent on communication with healthcare providers, patient and relatives (minutes) ED physician43.129.60106 Hospital (nurse and other)2.29.5023 Municipal home nurse10.419.6066 Municipal care staff4.811.2034 General practitioner1.03.1010 Other (eg, IT, pharmacy, secretary)1.13.1010 Patient and relatives14.017.7055 Acute nurse time spent receiving communication Municipal home nurse3.511.3036 Municipal care staff0.20.602Municipal home nurse time spent on communication with acute team, healthcare providers and relatives (minutes) ED physician0.21.902 Hospital (nurse and other)0.63.600 Acute team3.514.9036 Municipal care staff0.21.302 Occupational therapist0.43.805 Patient and relatives4.639.1051 Municipal nurse time spent receiving communication Acute team10.419.6066 Municipal care staff0.87.009Municipal care staff time spent on communication with acute team, healthcare providers and relatives (minutes) ED physician0000 Hospital (nurse and other)0000 Acute team0.22.702 Municipal home nurse0.87.209 Occupational therapist0.45.3020 Patient and relatives1.413.7018 Municipal care staff time spent receiving communication Acute team4.811.2034 Municipal home nurse0.21.302ED, Emergency Department; Max, maximum; Min, minimum; n, number of patients/observations. Time spent on communication (minutes) per treatment course ED, Emergency Department; Max, maximum; Min, minimum; n, number of patients/observations. Total time Figure 1 shows the distribution of effective working time per patient across staff groups and activity categories. The acute team accounted for the largest share of total time, particularly for treatment activities and transport, whereas municipal home nurses and care staff contributed smaller but distinct components related to delegated treatment, care, communication and documentation. Detailed numerical values are provided in online supplemental table S5. Mean effective working time per treatment course per patient, by staff group and overall, broken down by activity category. Values represent mean effective working time per patient per treatment course. Non-patient-related time (breaks, slack, indirect administration) is not included. Across all provider groups, the total effective working time per patient was approximately 940 min (15–16 hours) per treatment course. Reported time consumption excludes non-patient-related tasks, breaks and slack time, which will be accounted for in the time-driven activity-based costing tariff construction. Subgroup-specific totals for municipal home nurses and care staff are reported in online supplemental table S6. Feasibility of data collection At the beginning of the pilot, staff were asked to register time for each individual service. They reported that this was too burdensome in daily practice, and the procedure was revised so that staff instead recorded the number and type of services delivered together with the total time spent on services, supplemented by separate registrations of documentation, communication and transport. This adjustment reduced workload, improved acceptability and led to more complete recordings. Throughout the pilot, the project manager and research team maintained close contact with the participating municipalities to encourage data collection and provide feedback. Staff received reminders about definitions and procedures, and when logs contained incomplete or inconsistent information, the responsible staff member was contacted directly to clarify and improve the data. This process contributed to progressive improvement in completeness and consistency of the logs over time. Discussion Principal findings This pilot study shows that detailed, provider-reported micro-costing of HaH is feasible within a Danish municipal context. Using a co-designed log and targeted data quality procedures, we captured transparent activity counts and time across acute team nurses, municipal home nurses and municipal care staff. Intravenous medication and vital signs were the most frequent acute team services; municipal care most often provided personal care and practical help. Communication time was non-trivial, especially between acute team nurses and ED physicians, and with patients and relatives. We report resource-use quantities only. The micro-costing logs are applied in an ongoing RCT and collected data from this study will be used to construct a tariff adopting TDABC; the tariff will then feed into an economic evaluation. Positioning the study within existing evidence and methodological advances This prospective pilot study is, to our knowledge, the first to collect detailed activity-level data on resource use for HaH within a municipal healthcare setting in Denmark. International research has shown that HaH can achieve clinical outcomes comparable to, or better than, conventional hospital admission for selected patient groups, while maintaining high levels of patient satisfaction.9 26 However, evidence on resource use and costs remains mixed and context dependent. Studies have reported heterogeneity related to case-mix, staffing models and financing arrangements.11 13 27 Most existing evaluations rely on aggregate hospital or claims data, which offer limited transparency regarding the specific activities and staff time that drive total costs. The present study contributes to this literature by providing micro-level transparency in a community-based context, where such data are rarely available. By quantifying time spent on direct treatment, personal care, communication, documentation and transport across acute team nurses, municipal home nurses and municipal care staff, we link the clinical organisation of HaH with the underlying distribution of resources. The data show that the largest variation occurs in direct treatment and care activities and, for the acute team, in transport time. This pattern reflects genuine heterogeneity related to patient complexity, local organisation, geography and coordination pathways. These findings complement hospital-based cost studies by extending analysis to the cross-sectoral components of HaH that are often omitted in economic evaluations. They also underline the importance of collecting granular data before drawing conclusions about efficiency or scalability in decentralised healthcare systems. From a health-economic perspective, the study aligns with a growing methodological literature advocating micro-costing as the preferred approach for evaluating new or complex interventions where standard unit costs are not representative.20 27 28 Micro-costing allows transparent documentation of resource quantities and is particularly suited to heterogeneous services such as HaH. The approach also resonates with the principles of TDABC, which explicitly measures capacity and utilisation to derive accurate cost rates.23,2529 While previous applications of TDABC have largely focused on hospital workflows, our study demonstrates that similar principles can be operationalised in municipal healthcare through pragmatic data-collection tools and close collaboration with practitioners. Methodologically, the study contributes in several ways. First, it establishes a practical workflow for community services that integrates activity identification, co-designed instruments and iterative quality-assurance procedures. Early testing showed that recording time for each single activity was unmanageable; therefore, the procedure was revised to capture the number and types of services together with total time for each category. This adjustment reduced reporting burden while maintaining analytical precision. Second, the study highlights the value of continuous engagement between researchers and practitioners. Regular feedback, clarifications of ambiguous entries and direct dialogue with staff responsible for incomplete logs improved completeness and internal consistency. This experience adds empirical evidence to the emerging literature on the feasibility and acceptability of micro-costing in real-world service environments. Third, our reporting distinguishes between whole-sample averages and conditional averages among users, which avoids bias when zero values reflect genuine absence rather than missing data. Furthermore, the study delineates what can be achieved during measurement from what must be handled analytically at a later stage, such as the inclusion of overheads and capacity adjustments. Together, these contributions advance methodological transparency and provide a transferable framework for measuring and analysing resource use in decentralised, cross-sectoral healthcare systems. Strengths and limitations Strengths of this study include the close co-design process with frontline staff, the iterative refinement of the micro-costing instrument and the use of explicit operational definitions combined with an embedded quality-assurance process. The instrument was implemented across three municipalities that differed in population size, geography and staffing models, which supports both face validity and practical applicability. This multi-site implementation showed that detailed micro-costing can be integrated into routine documentation when supported by training, feedback and managerial engagement. The collaborative process also fostered ownership among staff and ensured that data collection reflected the real conditions of municipal acute care. The results show considerable variation in recorded time across patients and provider groups, as reflected by the high SD and wide ranges between minimum and maximum values in the tables. Variation is most pronounced for direct treatment and care activities, and there is additional substantial dispersion in transport time for the acute team. Communication and documentation also vary, but to a lesser extent. This pattern likely reflects genuine heterogeneity in patient complexity and care intensity, differences in local organisation and workflow and contextual factors such as place of residence (home or nursing home), geography and inter-sectoral coordination. From a methodological perspective, these findings underline the importance of micro-level measurement to make real-world heterogeneity visible rather than masked by averages. In addition, the transparent reporting of activity counts and time distributions allows others to adapt specific components, such as transport time, staffing patterns or delegation practices, to their own organisational context. This adaptability strengthens the practical transferability of the findings across settings with differing geographical or operational conditions. The pilot thereby provides valuable insight into the complexity of HaH delivery and highlights the need for future studies to examine determinants of variation more systematically in larger samples. A further consideration relates to capacity and marginal costs. HaH services are typically delivered by teams with fixed staffing levels, meaning that the average cost per patient declines as long as the team operates below full capacity. Once capacity is reached, however, accommodating additional patients requires additional staffing or restructuring, which sharply increases the marginal cost of treatment. Although this study does not examine capacity effects, they are important when interpreting cost estimates and planning the scale-up of HaH services. The main limitations are typical of feasibility studies. Reliance on staff self-report may lead to measurement error through under-recording during busy periods, rounding of time intervals or occasional misclassification of activities. We adopted a conservative approach to minimise the risk of overestimating time use, but we cannot rule out the possibility that some visits were undertaken without corresponding time registrations, which cannot be verified from the available dataset. The built-in feedback and clarification procedures probably reduced, but did not remove, such errors. The sample size is small and limited to one Danish region, and the results should therefore be regarded as descriptive evidence of feasibility rather than generalisable estimates of efficiency. Time was measured for the municipal and acute team providers who delivered HaH, but not for hospital physicians beyond direct consultation contacts, and communication time is reported only from the perspective of municipal and acute team staff. Only effective working time was measured. When subsequently developing HaH tariffs using TDABC, capacity cost rates will incorporate contractual time, annual productive minutes (subtracting leave, holidays and training), and expected utilisation; indirect activities (non-patient administration, availability, breaks) will be allocated through utilisation assumptions rather than merged into direct activity time.25 We did not measure ED physicians’ time beyond contact counts; this is a limitation and will be addressed analytically through tariff construction. Equipment costs were not measured in the pilot but will be incorporated as part of the valuation step by annualising the costs of the specific equipment and consumables used in HaH activities (including purchase, depreciation, maintenance and consumables) and deriving an equipment-specific capacity cost rate (CCR) based on operational hours. Equipment costs will then be entered into the time equations as the mean number of equipment-related activities multiplied by this CCR. The pilot did not collect information on informal caregiver time or support needs. However, HaH models may shift certain practical or supervisory tasks to family members or other informal caregivers, which can constitute a non-trivial component of total societal costs. These potential indirect effects fall outside the scope of the present pilot but should be examined in future evaluations to provide a fuller understanding of the resource implications of HaH. Finally, the geography of the participating municipalities likely influenced several of the recorded times, especially those related to transport and communication. Differences in urban density, distance between patients and the location of acute team bases may all affect total time use and limit direct transferability of results to other settings. Nevertheless, acknowledging and documenting this variation provides an important foundation for understanding how contextual factors shape the practical delivery of HaH services. Implications, transferability and future perspectives The findings from this pilot study have several implications for research, practice and policy. They demonstrate that it is feasible to collect detailed activity-level data on resource use within municipal healthcare, providing a foundation for transparent and reproducible measurement of HaH services. The activity dictionary and time distributions reported here offer a structured framework that can inform subsequent economic evaluations and policy decisions regarding the organisation of acute care in patients’ homes. For Danish decision-makers, the results confirm that municipal services can document their contributions to HaH delivery in a systematic manner, including clinical treatment, communication, documentation and transport. Such transparency is essential in a healthcare system where responsibilities are shared between regions and municipalities and where economic incentives must be aligned across sectors. The workflow tested in this pilot distinguishes universal elements of micro-costing from those that require local adaptation. Core components such as activity identification, categorisation and time recording are transferable across settings, whereas parameters related to geography, staffing models, patient eligibility and coordination structures will vary. By applying the same measurement framework and reporting both overall and user-specific averages, future studies can make heterogeneity in service delivery visible rather than hidden within aggregated data. This transparency is crucial for meaningful comparison of HaH models internationally and for informing the design of sustainable, proximity-based healthcare systems. The pilot also provides methodological insight for ongoing and future research. The next step is to apply this refined micro-costing approach in the ongoing RCT to evaluate HaH compared with standard hospital admission on clinical outcomes and costs. Further studies should explore how differences in local organisation and travel distances influence resource requirements and cost-effectiveness, and whether similar methods can be used to assess other complex cross-sectoral interventions such as rehabilitation, palliative care or integrated mental health services. Embedding micro-costing procedures into routine information systems may ultimately support continuous monitoring, quality improvement and more equitable resource allocation within decentralised healthcare systems. Conclusion This prospective pilot study demonstrated that detailed micro-costing of HaH delivery is feasible within the Danish municipal healthcare system. Through iterative co-design with care providers, we achieved transparent and consistent measurement of activities and time across acute team nurses, municipal home nurses and care staff. The results confirm that acute treatment can be safely organised in patients’ homes and that it requires substantial provider time, particularly for transport, documentation and communication. Beyond establishing feasibility, the study provides a practical framework for transparent resource-use measurement in cross-sectoral care. Co-design, clear operational definitions and structured feedback procedures proved essential to ensure completeness and accuracy of provider-reported data. These findings support Denmark’s broader aim of shifting specialised care closer to patients’ homes while maintaining accountability for resource use. The refined micro-costing approach will be applied in an ongoing RCT to construct a TDABC tariff and to inform future cost-effectiveness analyses and reimbursement models. Supplementary material Acknowledgements The authors would like to thank the members of the steering group and task force for their ongoing support and valuable input into the design of the Hospital at Home (HaH) treatment model. We are also grateful to the clinical staff involved in the project for their commitment and contributions to data collection. Special thanks are extended to the acute team nurses for their contributions to both data collection and the co-design process. Footnotes Funding: The project was funded by Health Foundation (HelseFonden, 20-B-0256), Novo Nordisk Foundation (NNF21OC0070180), Public Health Fund, Central Denmark Region (A4130) and The Quality and Training Committee, Central Denmark Region (1-30-72-29-19). Prepublication history and additional supplemental material for this paper are available online. To view these files, please visit the journal online (https://doi.org/10.1136/bmjopen-2025-106799). Provenance and peer review: Not commissioned; externally peer reviewed. Patient consent for publication: Not applicable. Ethics approval: This study involves human participants and was approved by the Ethical Committee, Central Denmark Region (No 1-10-72-67-20). Participants gave informed consent to participate in the study before taking part. Data availability free text: Data are not publicly available due to Danish data protection regulations. Access to the de-identified data underlying this study may be granted by the Danish Health Data Authority upon reasonable request and in compliance with national regulations. Requests should be directed to the corresponding author. Patient and public involvement: Patients and/or the public were involved in the design, or conduct, or reporting or dissemination plans of this research. Refer to the Methods section for further details. Data availability statement No data are available. References Supplementary Materials Data Availability Statement No data are available.