Perspectives from healthcare providers and decision-makers on strategies toward a future hospital: qualitative insights from a group modelling process

Abstract

Aim

Planning for a hospital is a complex and dynamic process, traditionally not informed by evidence. A systems thinking approach can be useful in informing hospital planning decisions. We derived insights from hospital executives and clinicians regarding the key drivers within the present hospital system that affect resource planning targets and leverage points for transformation in a new future hospital.

Methodology

We utilized a qualitative group model building (GMB) process conducted in one of the largest hospitals in Singapore to inform the development of a systems model for the planning of the new brownfield hospital. Participants included hospital executives supporting hospital planning, strategic management, health system planning, data management, and healthcare providers from clinical and allied health domains who were members of the hospital planning committee. Data were gathered through 10 GMB workshops from September 2023 to March 2024 and analysed using thematic analysis.

Result

A total of 20 participants attended the GMB workshops. Nine themes were identified across four hospital subsystems (Specialist Outpatient Clinic (SOC), emergency department (ED), inpatient, and surgeries & procedures (S&P)). Among these, five themes were linked to potential interventions to optimise hospital operations and patient care, including right-siting of patients (e.g., primary care, community, telehealth) and improving pre- and post-surgical care to reduce length of stay (LOS). Across this dynamic system, the most critical leverage point identified was the reduction of avoidable demand on acute hospital services through efficient patient right-siting and strengthened care transitions, which directly influences ED utilization and inpatient LOS. Four key drivers shaping hospital demand-supply dynamics were identified: ED utilization, extended inpatient LOS, the integration of technological advancements (particularly Artificial Intelligence (AI)), and the expansion of alternative care models. These insights directly culminated in a systems-level causal loop diagram that provided a foundation for identifying key leverage points to guide strategies aimed at better aligning hospital resources with patient demands while mitigating structural inefficiencies.

Conclusion

We identified four key drivers and five potential strategies across four hospital sub-systems that could be leverage points for hospital redesign and transition to the future mode of operations.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12913-026-14638-4.

Introduction

Hospitals are central components of healthcare systems and account for substantial public investment [1]. In the United States, annual expenditure on healthcare facility construction has reached USD $70 billion [2]. Similarly, Singapore has committed significant resources to hospital infrastructure, with direct development expenditures estimated at SGD 822 million in 2021 [3] and SGD $793 million in 2024 [4]. These investments reflect not only the financial cost of hospital construction, but also the long-term commitment of human, informational, and physical resources required to deliver high-quality care [5]. As such, hospital planning and design have far-reaching implications for healthcare delivery, system efficiency, and population health outcomes.

Over the recent years, hospital design has responded to demographic change, technological advancement, architectural innovation, and financial pressures [6]. Globally, population aging and rising multimorbidity have intensified pressure on acute care systems [7]. These challenges were further amplified by the COVID-19 pandemic, which exposed vulnerabilities in hospital surge capacity and infection control [8], leading to renewed attention to hospital preparedness and adaptability [9]. Singapore faces many of these same pressures, driven by rapid population aging and a growing burden of chronic disease [10]. In response, major hospital redevelopment initiatives have been launched to ensure that acute care infrastructure remains responsive to future healthcare needs [11]. One of the examples is to rehabilitate and upgrade the oldest and largest general hospital in Singapore. This Hospital of the Future initiative is intended to transform and better position the hospital in delivering holistic, multidisciplinary, and integrated care by 2035.

Importantly, planning for a future hospital extends beyond increasing physical capacity. Evidence-based design principles [12], such as exposure to natural light, noise reduction strategies, and single-patient rooms, have proven beneficial in improving safety and efficiency in primary care, ensuring better patient outcomes, and also lowering operational costs [13, 14]. At the same time, advances in digital health, alternative care models, and ambulatory services challenge the traditional hospital-centric model of care [15, 16]. Modern acute hospitals must therefore be designed not only as sites of inpatient care, but as adaptable hubs within a broader, integrated health system that spans community, primary, and specialist services [17]. In parallel, health systems in Singapore are increasingly emphasizing right-siting of care, which refers to the deliberate allocation of healthcare services to the most appropriate setting based on clinical complexity, patient needs, and resource intensity [18]. Effective right-siting requires careful coordination across care settings and has important implications for hospital demand, workforce planning, infrastructure utilization, and long-term system sustainability [19, 20].

A growing body of literature recognizes hospital planning as a complex, dynamic process shaped by interactions among clinical workflows, workforce capacity, infrastructure, technology, policy, and patient demand [21–24]. Systems thinking has increasingly been advocated as a useful approach for understanding such complexity, as it emphasizes interdependencies, feedback mechanisms, and non-linear effects within healthcare systems [25]. Prior studies have applied systems-based approaches to topics such as hospital capacity planning [26], operation efficiency [27], and workforce optimization [28]. However, much of this work relies on administrative data [29, 30]. In addition, while prior studies have incorporated qualitative approaches to examine hospital redesign [31], value-based care [32], and infrastructure planning [33], these studies largely remain descriptive and context-specific. They often do not translate stakeholder insights into formal system representations, limiting their ability to capture feedback loops, time delays, and the unintended consequences of localized interventions across the broader hospital system. As a result, existing qualitative approaches have had limited success in informing long-term, system-wide hospital planning, particularly in anticipating how changes in one part of the system (e.g., emergency department (ED) demand) may create downstream effects elsewhere.

To address these gaps, this study applies a systems thinking approach using Group Model Building (GMB) to elicit and integrate perspectives from hospital executives and clinicians involved in strategic planning and care delivery [34]. Insights generated through GMB were synthesized into causal loop diagrams (CLDs), serving as the conceptual foundation for subsequent system dynamics modelling of hospital resources and service demand over time. This paper reports the qualitative analysis of ten GMB workshops and highlights key system dynamics shaping resource planning and transformation in the Hospital of the Future initiative. The description and analysis of the system behaviour have been reported elsewhere [35].

Methods

Study setting

GMB is a participatory method to involve stakeholders in the process of developing system dynamics models to enhance collective understanding [34]. This GMB process was conducted in a general hospital in Singapore to inform a system dynamics model for hospital planning. The study hospital (SH) is part of a public healthcare cluster in Singapore, which consists of three comprehensive acute hospitals, one specialty hospital, five national specialty centres, three community hospitals, and 10 primary care institutions. The healthcare cluster serves both national and regional healthcare delivery functions.

Participants

To capture diverse system-level perspectives related to hospital planning, we employed a multi-strategy purposive sampling approach [36]. Combining criterion-based and maximum variation purposive sampling, participants were purposively sampled based on organizational role, hospital sub-system, and level of involvement in hospital planning. Hospital executives were recruited using criterion-based purposive sampling. The inclusion criteria were as follows: (1) a minimum of five years of work experience in hospital planning, strategic management, health system planning, or health data analytics, as these positions were directly involved in decision-making related to future hospital design; (2) a willingness to participate in this study. Healthcare providers were recruited using maximum variation purposive sampling to ensure representation across clinical and operational contexts. Specifically, clinicians and allied health professionals who were members of the Hospital Planning Committee were identified, with deliberate inclusion of participants from each of the four hospital sub-systems: Specialist Outpatient Clinic (SOC), ED, inpatient, and surgeries & procedures (S&P). Potential participants were identified in consultation with hospital leadership and planning offices, and eligible individuals were invited via email. Some core members of the hospital planning team participated in multiple workshops to provide continuity across sessions.

Data collection

During the GMB process, qualitative data were gathered through ten GMB workshops from September 2023 to March 2024. A discussion guide consisting of eight questions around the base case was developed for the first workshop. Based on the inputs of the first workshop, the research team developed CLDs along with a summary of the discussions. The draft CLDs were then reviewed and refined in the subsequent nine workshops. The first draft and the final version of the CLDs are presented in Supplementary Files 1 & 2. A more detailed description and analysis of the CLDs is published separately [35].

Attendance per workshop ranged from 8 to 15 participants, with an average of 11 to 12 per session. Participation was intentionally structured rather than fluid, reflecting the evolving focus of the modelling process. Early workshops emphasized system-level drivers of hospital performance and primarily involved hospital executives and clinical leaders representing major hospital sub-systems. Subsequent workshops emphasized issues related to the four hospital subsystems, with increased participation from clinicians and allied health professionals directly involved in service delivery. Several core participants, including members of the hospital planning team, attended multiple workshops to provide continuity across sessions. Supplementary File 3 summarizes the topics and key outputs of each workshop.

To maintain coherence and support consensus-building despite changes in attendance, each workshop began with a structured briefing on the most recent version of the CLD. Facilitators reviewed key assumptions, feedback loops, and prior decisions, and participants were invited to validate, challenge, or refine the model. Given the interpretive nature of GMB within a system dynamics approach, the facilitation team, comprising researchers with expertise in system dynamics and health services research, adopted a reflexive stance regarding their potential influence on the modelling process. They were mindful of how their backgrounds and assumptions could influence the framing of problems and the structure of models. To mitigate this, facilitators actively encouraged diverse perspectives, probed for alternative explanations, and regularly checked that emerging model structures accurately represented participants’ perspectives rather than facilitator interpretations. Revisions were incorporated iteratively, ensuring that the CLDs represented a shared and cumulative understanding of the system rather than the views of a single workshop cohort.

Data analysis

The analysis of the qualitative data followed an iterative process. All workshop sessions were conducted face-to-face and audio recorded. Two study team members took notes during the workshops and transcribed the discussions using the recordings. The transcripts were checked and edited manually to ensure accuracy. After familiarizing themselves with the notes and transcripts, two study team members generated codes independently. The codebook was updated after each session. Similar codes were clustered together to create categories and themes. Data management was performed using NVivo R1 (QSR International, Australia).

Results

A total of 20 participants from the departments of Future Health System, Organization Planning & Performance, Strategic Management & Analytics, Data Science, Health Services Research, Emergency Medicine, Oncology, Endocrinology, Orthopaedics, and Nursing, attended the workshops. Six of them were at the director level, and another three of them were senior consultants. Another three held senior positions, including one principal analyst and two research scientists.

The demand for services at the study hospital mainly arises from the SOC, ED, inpatient, and S&P sub-systems, which are the four main functional units of a general hospital. SOC deals with patients referred from primary care or scheduled for consultations with specialist doctors. ED is responsible for delivering urgent and life-saving care to patients with unscheduled care needs, including walk-in patients and those arriving by ambulance. Patients in the inpatient sub-system are those admitted to hospital wards for stays longer than 24 h. These patients can only flow into the hospital through the SOC or ED sub-systems. The S&P sub-system deals with patients who require surgeries or procedures, including day surgeries.

An increasing healthcare demand at the population level, driven by aging and increasing chronic disease burden, drives demand in all four sub-systems, but there is also a high level of interdependency between these four sub-systems. Similarly, demand within the study hospital is also affected by the capacity of sectors outside the hospital (e.g., primary care, intermediate, and long-term care) for absorbing some of the demand or transitioning patients out of the hospital in a timely manner. Apart from macro-level factors, we have also identified specific interdependencies within each sub-system and refined their relationships over the GMB workshops. A total of nine themes were identified, including five potential interventions in the four hospital sub-systems, two themes on factors affecting ED use and the hospital length of stay (LOS), and two general drivers of resource use (Fig. 1).

Fig. 1.

A scheme showing the nine themes across the four hospital sub-systems

Specialist outpatient clinic (SOC)

Theme 1 right-siting patients to primary care (intervention)

Participants highlighted the need to right-site SOC patients to primary care institutions through the collaboration between SOC and primary care. However, they acknowledged the overwhelming amount of domain knowledge that primary care doctors must have to manage patients across various medical domains. They advocated that the collaboration between specialists and primary care physicians could start with the disciplines of endocrinology, orthopaedics, and gastroenterology because clear care pathways within these disciplines have been established and are easily accessible.

They also highlighted the importance of an easy process to allow patients to transition between tertiary care and primary care, especially for the elderly. One doctor explained, “For elderly patients, their healthcare problems may flare up, recur, or new issues may develop. So care needs to be more fluid, allowing them to come back when needed. That’s where appointment systems need to be very fluid.”

However, challenges to right-site patients to primary care exist. First, the current healthcare system may not have achieved completely seamless communication between primary care and tertiary care, making it difficult to switch care between them. This is true for both the electronic medical record system and transition pathways. Some patients’ common concern is described as: “Once they are discharged from tertiary care, it can be excruciatingly difficult for them to get back into tertiary care again”. Second, patients’ perceptions of care quality further constrain right-siting effects. Tertiary centres are often viewed as providing superior care compared to primary care, while differences in cost remain relatively narrow. This perceived imbalance reduces patients’ willingness to transition to community-based care, even when clinically appropriate.

These barriers are reinforced by underlying dynamics identified through system dynamics analysis (Fig. 2). In particular, a “Success to the Successful” pattern emerges, whereby sustained investment in specialist services enhances their capacity and reputation, attracting increasing patient demand and referrals. Over time, this reinforcing dynamic draws resources and attention away from primary care, constraining its development and perpetuating system dependence on SOCs for conditions that could otherwise be managed in lower-acuity settings.

Fig. 2.

A casual loop diagram of the specialist outpatient clinic (SOC) view. Abbreviation: LAB/RAD, laboratory and radiology

Theme 2 right-siting patients to the community (intervention)

Participants suggested establishing a collaboration system for specialists to connect patients to various community resources, providing patients with recovery support, and improving health within their familiar local healthcare networks. A participant highlighted, “The whole idea is to keep them in a healthier life within the community”. The participant also shared an initiative in which the community nursing team conducted outreach events to introduce available referral options to SOC physicians and general practitioners (GPs), with the aim of increasing referrals to community-based social care services and/or social workers. However, the participant noted that the care navigator (intended to function as a primary entry point into the system) was currently underutilized and lacked sufficient visibility.

Theme 3 right-siting patients to telehealth (intervention)

While telemedicine is often viewed as a separate modality distinct from physical care, one participant suggested blending both and replacing brief follow-up consultations with teleconsultations. This requires them to identify scenarios where in-person visits are unnecessary and offer accessible alternatives for teleconsultation to save patients’ traveling time and waiting time. According to this participant, “Not every single patient needs to be followed up (physically) at all times. At some point, we can move to telehealth, but there needs to be a more flexible way to do it.”

Emergency department (ED)

Theme 4 right-siting patients away from the emergency department (intervention)

ED doctors mentioned the only two aspects they can right-site ED patients were during post-ED visits and pre-ED visits. In terms of pre-ED visits, participants emphasized the need for public education to encourage individuals to seek care from their GPs first. The GPFirst program was launched in a general hospital to serve residents in the East of Singapore in 2014 [37]. It encourages patients with mild to moderate medical issues to consult their GPs first, rather than going to the ED directly. This program aims to avoid inappropriate ED attendances, allowing patients to receive timely medical treatment and enabling ED physicians to focus on more serious cases requiring medical attention. There are also several urgent care centres (UCC) in Singapore [38, 39], which handle urgent but not life-threatening conditions. However, they questioned whether there would be improvements in their uptake, underscoring the importance of GPs as the initial points of contact for healthcare needs.

In addition to GPFirst, they highlighted the importance of right-siting patients away from wards by preventing avoidable admissions. One physician cited that “There are certain situations we can avoid admissions by referring them to early SOC access or admitting them directly to community hospitals”. He mentioned a phenomenon that many referrals they received involve primary healthcare professionals referring patients to the ED for access to the SOC. Owing to limited SOC resources available, patients may choose to pay an ED attendance fee to expedite their care. However, such a process would drain the ED resources and lead to unnecessary admissions if SOC resources were limited. One participant shared an example of the collaboration between primary care institutions and ED: a pilot program between a public hospital and a primary care institution (polyclinic) to use teleconsultations. In this program, teleconsultations with the polyclinic GPs were provided before any patient was sent to the ED, which allows for onward referral or keep-in-view admission. This approach aims to reassure patients that they will be seen in SOC in a timely manner, potentially saving them from unnecessary ED visits and admissions caused by the long lead time for referrals.

Another aspect of admission avoidance is preventing the deterioration of health conditions. The doctors highlighted the five main types of conditions that can be intervened to prevent the deterioration, which were “diabetes, hypertension, asthma, chronic obstructive pulmonary disease, and flu”. He shared an example of providing a multidisciplinary intervention to a patient with diabetes, “For example, we could send a diabetic nurse to support the patient, to reduce their burden and help prevent deterioration of their condition”. He also highlighted the importance of utilizing telehealth to improve patients’ self-care activities. As he said, “One of our consultants does this—if you have diabetes, we follow up through telehealth or telecare. Someone will call and nag you and make sure you’re managing your condition.” The purpose of both approaches is to reduce the chronic disease burden on patients and prevent deterioration of their health conditions.

Theme 5 reasons for frequent use of the emergency department (ED) (factor)

It was reflected in the GMB that frequent attenders of the ED have posed a huge burden to emergency physicians. There are two broad groups of frequent users. The first comprises patients with underlying chronic illnesses, such as advanced heart failure, who experience recurrent exacerbations. As one physician noted, “You might have a patient with end-stage heart failure who gets discharged but keeps coming back frequently because of decompensation1.” The second group is largely driven by social and behavioural factors, including alcohol abuse, personality disorders, and psychosomatic symptoms. Beyond patient-level factors, participants highlighted that the limited access to after-hours primary care was an underlying service gap contributing to repeated ED use.

From a system dynamics perspective, these gaps contribute to reinforcing patterns of ED utilization. Frequent attenders, although small in number, generate a disproportionately high volume of visits, amplifying system strain. As a participant explained, “A small number of patients can generate a huge number of visits. They may even come every day. So while the number of patients is small, the impact is large, and it often points to gaps in the system.”

As shared by the participants, the Ministry of Health has supported a research project to tackle the issue of alcohol-related frequent attenders (ARFA). This initiative ensures each hospital has an ARFA team comprising patient navigators and medical social workers to engage with ARFAs to prevent frequent ED admissions by addressing their underlying social issues [40]. However, the participant doubted that the allocation of such a substantial number of resources to address the needs of only a few people in such programs is justifiable.

Inpatient

Theme 6 long-stayers and reasons for prolonged length of stay (LOS) (factor)

One participant commented that long-stayers consume a significant proportion of the total bed capacity, “accounting for 20% of the total bed days.” Since long-stayers “soaked up” a lot of bed days, the participant called for more interventions to improve the flow and reduce bed days for these patients.

There are various reasons leading to prolonged LOS, including waiting for laboratory or radiological investigations, turnaround time for tests, waiting for interventions, and limited time for caregiver training due to the restriction of visiting hours. Another reason is the patient’s preference to stay longer. A physician shared that, “Some ‘avoidable’ bed days are not really avoidable, because patients may prefer to stay in the hospital and don’t want to go home.” The participant observed that under pressure to discharge patients, physicians tend to advocate primarily for unavoidable bed days (those that are medically warranted).

The emergence of new treatment methods or technologies has shown the potential to shorten the LOS, such as highly sensitive troponin tests [41], point-of-care testing [42], and day surgeries [43]. Changes in policies or departmental protocols could also shorten LOS. An example related to CT scans was shared in the interviews, “Previously, patients might stay in Short Stay Units2for eight hours just because they could use Medisave3to cover the cost of the scan. Now that some scans are Medisave-claimable. This is less of an issue.”

Surgeries

Theme 7 improving pre- and post-surgical care (intervention)

One doctor suggested providing pre-and post-surgical care, including physiotherapy, rehabilitation, and nutrition, to enhance recovery after surgery. She explained that “all those factors play a big role in both before and after recovery, and they can help shorten recovery time.” Although these interventions may not directly affect the number of surgeries performed, they can influence inpatient rates by preparing patients better for surgery and facilitating faster recovery, ultimately leading to shorter hospital stays.

A nursing director commented that “it’s very hard to motivate patients to exercise before surgeries”, but she emphasized the team effort to encourage patients to adopt healthier behaviours, such as quitting smoking, which may lead to smoother recoveries and potentially change the nature of surgeries from emergency to elective procedures, and more controlled surgeries.

General drivers of resource use

Theme 8 technological advancement and artificial intelligence (AI)

One participant suggested using machine learning (ML) to optimize patient care pathways within the hospital, especially for older adults who often have multiple medical appointments on the same day. By predicting inter-clinic travel times and the expected waiting time at each service location, AI can be used to generate an optimal personalised plan to minimize delays and reduce care fragmentation. The potential to improve patient experience whilst improving overall system efficiency and care effectiveness within the hospital was highlighted. One doctor reflected, “As patients get older, they tend to have a lot of appointments. If we can’t have a fully integrated multidisciplinary clinic, we should at least try to synchronize each visit to make things most efficient.”

To make full use of underutilized spaces and time slots, another participant shared an experience in their clinic of how they applied AI to efficiently schedule appointments during the COVID period. They used AI to analyse historical operational data and determine how long a doctor spends with new and follow-up patients. Through using predictive ML models, they were able to optimize appointment slot distribution by eliminating empty time slots or efficiently rescheduling appointments if patients canceled appointments at the last minute. The main purpose of integrating predictive models into their system was to manage dwell time to prevent crowding during COVID-19.

Theme 9 the development of alternative care models

1. Hospital @ Home (HaH)

HaH is an alternative inpatient care delivery model that aims to provide equitable care for clinically suitable patients, allowing them to be hospitalized in their own homes instead of a hospital ward [44]. This is expected to reduce total in-hospital bed days, which has direct implication on bed requirements. A participant pointed out that “HaH is more of a way to offload patients rather than replacing a significant portion of inpatient beds.” HaH does not necessarily reduce the average LOS. However, those who are not suitable for HaH and stay in the physical hospital are tougher cases and might have higher LOS, contributing to a higher average hospital inpatient stay.

• 2.GPFirst Program and Geri-ED

Encouraging patients to consult a GP first, as right-siting away from the ED was advocated by some participants and described in Sect. "Surgeries". A study on inappropriate attendance at ED in a tertiary hospital in Singapore found that 9.6% of the attendances were inappropriate attendances, contributed by patients with mild or moderate medical conditions [45]. Currently, the GPFirst program has expanded to seven tertiary hospitals and two UCCs [37].

Additionally, doctors are considering developing alternative models of care within the ED. For example, by utilizing an e-learning website Geri-EM [46], facilitating e-learning in geriatric emergency medicine, ED staff are more empowered to screen geriatric patients and identify the appropriate care for them. This kind of tool will allow target patients “to be fast-tracked to outpatient, (getting) early access to geriatric care or community hospitals or slower stream”, potentially reducing the need for hospital admissions.

• 3.Teleconsult in SOC and ED

Teleconsultations are currently conducted on an ad hoc basis and from any location convenient for the doctor (e.g., consult rooms). As telehealth is expected to expand across more than just SOC, several participants suggested the idea of a centralized telehealth centre within the study hospital, equipped with the necessary infrastructure and support. This pooling of resources for telehealth consultations, with the assistance of a 24/7 translator service, could enhance communication with patients and improve efficiency. They noted that future advancements, such as AI translation capabilities, may further streamline this process.

Although virtual EDs have also been widely developed worldwide, the doctor commented that “we have our reservations on that” since not all ED physicians support this concept due to concerns of not being able to physically examine patients. There is currently a very low risk appetite for managing such cases virtually, particularly for non-undifferentiated patients across the acuity spectrum.

Discussion

This is the first study to report qualitative insights into the brownfield hospital design and planning of a major tertiary hospital in Asia. Besides uncovering key drivers of key resource requirements in the four hospital sub-systems, we identified potential interventions to improve hospital efficiency and patient flow, which have important implications for the resource and facility planning of the future hospital. Our findings also have significant implications for the development of new care models and healthcare AI.

The term “right-siting” was frequently mentioned in our GMB workshops, showing the urgent priority of right-siting due to the rapidly aging population and the rising patient loads across the four hospital sub-systems. Since 2004, the term “right-siting” has been introduced to describe the principle that patients with chronic diseases should be managed in primary care rather than specialist settings [47]. The frequent use of right-siting in our study highlighted the urgent need for improved strategies to address the underlying challenges in appropriately directing patients from the specialist and emergency care to primary care or community institutions, to alleviate overcrowding in the ED and SOC. The past focus on specialist care has contributed to higher costs associated with specialist treatment for both patients, providers, and the healthcare systems [48]. Emphasizing specialist care can create access barriers for patients; a specialist-focused system may not adequately address the long-term management of chronic diseases. Thus, in recent years, there has been a shift toward strengthening primary care and integrating it with specialist care to create a more balanced and efficient healthcare system [49, 50].

However, insights from the workshops reveal that right-siting is embedded within reinforcing feedback dynamics that may undermine its effectiveness. Participants highlighted a vicious cycle in which patients’ perceptions of primary care as lower quality reduce their willingness to accept right-siting. This leads to continued or increased reliance on SOC and ED services, reinforcing congestion and long waiting times (Fig. 2). In turn, these pressures further strain specialist services, prompting additional resource allocation toward hospital-based care rather than investment in primary care capacity, thereby perpetuating the imbalance. Efforts to expand access to alternative care providers, such as UCCs, may not always reduce system burden. For example, evidence from the United States suggests that such models can increase overall spending for lower-acuity care if they generate additional demand rather than substituting for existing services [51]. Without addressing underlying perceptions of care quality, trust, and care continuity, right-siting strategies risk creating leakage back into the acute system, ultimately undermining their intended impact.

Participants from this study proposed the use of AI and collaboration with primary care institutions to smooth the workflow and decrease patient load. In a public hospital specializing in healthcare for women and children, a 24-hour online chatbot platform called “the Urgent Paediatric Advice Line (U-PAL)” providing advice to parents and caregivers for common paediatric conditions or common injuries was launched in 2019 [52]. This approach enabled ED physicians to interact with patients and caregivers prior to their ED arrival and attempt to right-site patients. This is an example of the potential of right-siting patients to ambulatory care settings by integrating various clinical support services from both hospitals and the community with technology. However, alternative care models often face multiple implementation barriers, such as unclear benefits from the patient’s perspective, provider resistance, resource allocation challenges, technology issues, and financial incentives issues, which require healthcare organizations’ efforts to navigate the complexities of implementing alternative care models.

Our study also found that community care could serve as an important resource for providing intermediate healthcare to patients, and community care capacity affects the discharge workflow and inpatient bed occupancy in the hospital. As suggested by the study participants, the partnership between community facilities and public hospitals allows patients to recover and improve their health within their familiar local healthcare network after discharge. With an increasing number of aging patients with non-communicable diseases, strengthening the capability and capacity of the community sectors has become a focus of the healthcare transformation [53–55]. In Singapore, one good example of community care integration is the Central Health Integrated Care Network, which aims to develop Communities of Care (CoCs) by building partnerships between the healthcare cluster with other agencies and organizations in the community, and building integrated place-based models of care [56]. This has culminated in over 90 community-based health and social activities and programs that access community resources to support residents in achieving their health goals and plans [56].

It is also worth mentioning that frequent attenders contribute substantially to ED crowding, not unlike in other countries [57–60]. Though definitions vary, frequent attenders refer to a diverse group of patients who have three to 12 or more ED visits per year, often without any clear rationale [61]. Published evidence shows that a significant proportion of ED attendance is inappropriate [59], and the causes of frequent visits are multifactorial. Although many patients have chronic conditions, especially in our comprehensive tertiary study hospital, there is a considerable number of people combined with marked psychosocial morbidity [58, 61] or are less medically urgent [62]. A retrospective cohort study of 451,775 patients reported that 16.9% of the frequent attenders were due to mental health or alcohol/substance abuse [63]. Since this group of patients is more likely to continue using ED with higher intensity, they should be prioritized for intensive intervention such as case management [64, 65] or community services [66].

Limitations of the present study included a small number of participants from a single hospital setting. However, the study hospital functions as the country’s national hospital, ranked as the 8th best hospital in the world [67]. Second, owing to clinicians’ busy schedules, we were unable to include practitioners from all disciplines or other hospitals. However, a variety of viewpoints were captured from hospital executives involved in planning the Future Hospital and practitioners with years of experience in the healthcare setting. As Singapore is a well-developed country with an island-wide network of public healthcare institutions, the findings from this study may lack generalizability to other countries with different geographical and organizational characteristics. However, the findings of this study may apply to other developed countries with aging populations. Third, the quantitative simulation of the system dynamics model was not performed as it was outside the scope of this paper. Lastly, the findings of this study are specific to the Singaporean context, a highly integrated, state-funded/subsidized healthcare system, and may differ in fragmented or purely insurance-based systems.

Conclusion

This study presents insights from hospital executives and clinicians on the key drivers affecting hospital planning and leverage points for future hospitals. The key drivers in the current hospital sub-systems and leverage points for hospital redesign strategies identified from this study served to inform policies aimed at developing a vibrant healthcare ecosystem that delivers a seamless continuum of care.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization: XY, CSL, AK, GGKT, THK and SLSW. #Methodology: XY, CSL, AK, SLSW. #Investigation: XY, CSL, AK, GY, SLSW. #Data curation and formal analysis: XY, CSL. #Supervision: GGKT, THK, SLSW. #Writing – original draft: XY. #Writing – review & editing: all authors.

Funding

No funding was received for this study.

Data availability

The data that support the findings of this study are not openly available due to reasons of sensitivity and are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Ethical approval was sought under the SingHealth Centralised Institutional Review Board (Ref. No 2025 − 0331), and Consent to participate was waived. In accordance with the Declaration of Helsinki, this study does not involve any experiments involving human subjects.

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.