Identifying recruitment and enrollment barriers in home-based acute care: findings from a regional pilot program in Northern Taiwan

Abstract

Background

Hospital-at-home (HaH) programs provide acute medical care in home settings as an alternative to traditional inpatient hospitalization. In July 2024, the National Health Insurance Administration of Taiwan launched an HaH pilot program. This program focuses on older adults with frailty and operates through 2 referral models: emergency passed admission (EPA) and direct home admission (DHA).

Methods

This retrospective observational study was conducted in the Xinwu-Tai region of New Taipei City, Taiwan, from September 2024 to June 2025. Patient screening records were reviewed to determine the reasons for enrollment failure and identify the outcomes of those admitted. Key variables included diagnosis, referral model, functional status, clinical stability, treatment duration, and readmission.

Results

Of 1462 emergency department (ED) person-days screened, only 180 (12.3%) were eligible for the HaH program, with 6 (0.4%) enrolled through the EPA model. Among 1282 ineligible ED person-days, 50.6% were excluded for insufficient functional impairment (Barthel’s index > 60) and 49.4% were excluded for unstable clinical conditions. Another 5 ventilator-dependent patients were enrolled through the DHA model, yielding a 71.4% admission rate and a total of 11 HaH cases. A total of 10 patients completed home treatment without complications; only one required ED revisitation within 14 days.

Conclusions

Enrollment in the HaH program in Taiwan is limited by strict eligibility criteria, particularly through the EPA model. However, the favorable outcomes observed among the admitted patients support the model’s safety, feasibility, and cost-effectiveness. Expanding referral models and refining inclusion criteria may boost participation and enhance the program’s impact.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12913-026-14509-y.

Introduction

The concept of telehealth originated as a solution for bridging health care gaps in rural and medically underserved areas [1, 2]. With health care systems gradually shifting from fee-for-service to value-based models, telemedicine has emerged as a promising approach for enhancing care delivery while controlling costs [1, 3]. Telehealth technologies have been widely employed across various domains, including medical education, remote monitoring, video-based consultation, and diagnostic data transmission. These tools have demonstrated efficacy in improving health care accessibility, reducing wait times, and maintaining or even improving quality of care, particularly in regions with provider shortages [3]. Despite these benefits, the expansion of telehealth remains constrained by technological adoption barriers, particularly among older populations and in areas with insufficient broadband access [3].

Social distancing restrictions during the COVID-19 pandemic accelerated the global expansion of telemedicine, particularly in countries such as Taiwan, where its adoption had previously been limited [4, 5]. In response, Taiwan amended its National Health Insurance Act in 2020 to include coverage for telemedicine services [6]. In Taiwan, telemedicine is primarily applied in 2 contexts: physician-to-physician consultations, particularly between different specialties, and direct patient access to health care services without limitations pertaining to geographic distance or quarantine [7]. A systematic review from 2021 highlighted the contributions of Taiwan and the United States to the field of health care technology acceptance, with growing attention to telemedicine-related research [8].

The hospital-at-home (HaH) model was first developed in the United Kingdom in the 1970s and later introduced in the United States in 1996 [9, 10]. This model initially focused on providing hospital-level care at home for older adults (≥ 65 years of age) with acute illnesses, such as community-acquired pneumonia, cellulitis, congestive heart failure, and chronic obstructive pulmonary disease. Over time, countries such as Australia, Israel, and Canada started implementing this model, indicating its effectiveness in lowering health care expenditures, improving care accessibility, and enhancing service quality [9, 10]. With advances in telemedicine and remote monitoring technologies, the HaH model has evolved into an integrated form of home-based acute care. Telehealth platforms now enable remote clinical assessment, continuous physiological monitoring, and virtual consultations, which support safe and timely decision making and facilitate the delivery of hospital-level care in home settings [11].

As a result of population aging and the growing burden of chronic diseases, traditional institution-based acute care models have become increasingly inadequate to meet the evolving and diverse health care needs of the population. Therefore, on July 1, 2024, the Taiwanese Ministry of Health and Welfare officially launched the National Health Insurance Pilot Program for Home Care of Acute Symptoms, introducing a home-based hospitalization model that differs from conventional care delivery [12]. Given its role as a key innovation in advancing home-based acute care, it has demonstrated major potential to reduce unnecessary hospital admissions, mitigate emergency department (ED) overcrowding, and enhance the overall efficiency of health care delivery [13]. Despite its benefits, the HaH pilot program presents multiple implementation challenges. Health care institutions are required to establish new workflows, standard operating procedures, medical equipment guidelines, and staff training systems. In addition, public awareness of this program remains limited, which may hinder early-stage recruitment and participation [14].

Therefore, this study aimed to identify the key recruitment and enrollment barriers encountered during the first 10 months of implementing the HaH program at a single hospital in Taiwan. We further analyzed the characteristics of eligible and ineligible cases and evaluated clinical outcomes among enrolled patients. Understanding these factors may provide both academic and practical insights for optimizing patient selection strategies and improving the implementation of home-based acute care programs.

Methods

Study design

This retrospective study was conducted at a single hospital to examine the data of patients who visited the ED or were successfully enrolled in the HaH program between September 2024 and June 2025. This study period was selected on the basis of the official initiation of the hospital’s HaH team operations in September 2024.

Ethical approval

This study was conducted in accordance with the Good Clinical Practice guidelines and the principles of the Declaration of Helsinki. The study protocol was reviewed and approved by the Institutional Review Board of Fu Jen Catholic University Hospital, New Taipei City, Taiwan (approval no. FJUH114533). Owing to the retrospective nature of the study, the requirement for written informed consent from participants was waived by the Institutional Review Board.

HaH pilot program

The HaH pilot program recruited patients through 2 main models (Fig. 1):

Fig. 1.

Workflow comparison between conventional inpatient care and hospital-at-home (HaH). The conventional pathway involves emergency department (ED) presentation followed by hospital admission. The HaH pathways include the Emergency Passed Admission (EPA) model and the Direct Home Admission (DHA) model, both enabling hospital-level care at home

1. Emergency passed admission (EPA) model: This model targeted patients who presented to the ED with a major functional impairment, defined as a Barthel index (BI) of < 60, or those for whom hospital-based care was considered to be challenging because of the nature of their illness or disability. Once these patients met the eligibility criteria during their ED assessments, they were directly discharged and allowed to return home, where they waited for an initial home visit and formal home admission by the HaH program team.

2. Direct home admission (DHA) model: This model targeted patients who were already receiving home health care services or living in long-term care facilities. Its eligibility criteria included functional disability, defined as a BI of < 60, active engagement in home care services, and enrollment in home-based palliative care programs. These individuals were collectively referred to as home-based patients. After the HaH program team conducted an on-site evaluation and confirmed their eligibility, these patients were formally enrolled and began receiving acute care at home.

Patients were deemed eligible for the HaH program if the attending physician determined that their primary diagnosis was one of the following: pneumonia (International Classification of Diseases, Tenth Revision, Clinical Modification [ICD-10-CM] codes: J12-J18, J20-J22, and J69.0), urinary tract infection (UTI; ICD-10-CM codes: N10, N34, N30.0, N30.3, N30.8, N30.9, and N39.0), or soft tissue infection (ICD-10-CM codes: L03.0, L03.1, L03.2, L03.3, L03.8, and L03.9). Although these patients were clinically eligible for inpatient treatment, they were also deemed eligible for receiving acute care at home. Table S1 presents these patients’ detailed diagnostic inclusion criteria.

The HaH multidisciplinary care team primarily comprised physicians, nurses, respiratory therapists, and pharmacists. This team provided various services during the care delivery period at home. These services included in-person physician visits (on the first and last days of care delivery), teleconsultations, daily nurse visits, respiratory therapist visits when required, pharmacist visits when required, prescription and medication dispensing services, and bedside laboratory and diagnostic testing. Bedside evaluations included point-of-care testing and point-of-care ultrasound, enabling timely clinical assessment and treatment adjustments in the home setting. Additional support included individualized health management and 24-hour telephone consultations, with emergency home visit availability. Whenever any patient was discharged from the HaH program, the care team conducted a comprehensive evaluation to determine whether that patient required referral to long-term care services.

Population and data collection for the EPA model

This study included individuals aged > 18 years who presented to the ED as outpatients. These individuals were included if they had a condition requiring admission into the HaH program, specifically pneumonia (ICD-10-CM codes: J12-J18, J20-J22, and J69.0), UTI (ICD-10-CM codes: N10, N34, N30.0, N30.3, N30.8, N30.9, and N39.0), or cellulitis (ICD-10-CM codes: L03.0, L03.1, L03.2, L03.3, L03.8, and L03.9).

To capture the dynamic clinical course of patients with prolonged ED stays, ED person-days were used as the primary unit of analysis. Each calendar day that a patient remained in the ED was counted as one person-day, and patients with prolonged stays therefore contributed multiple records. This approach was adopted because patients’ clinical status and willingness to participate in the HaH program may change during extended ED stays, which can influence daily eligibility assessments and enrollment decisions. Accordingly, all HaH eligibility evaluations were conducted based on daily clinical reassessment by the HaH team. Demographic and clinical variables, including age, diagnosis, and HaH program eligibility, were recorded and analyzed on a per person-day basis. This approach captured the changing clinical condition of each patient and their readiness to participate in the HaH program. Eligibility was assessed daily by the HaH team through clinical evaluations.

After patients with a BI of > 60 were excluded, those with an unstable clinical condition were also excluded in accordance with criteria established by the HaH team, with reference to the Taiwan Triage and Acuity Scale (Table S2) [15]. Only those who had a BI of < 60 and were stable were considered to be eligible for enrollment in the HaH program. Additional assessments were conducted to evaluate each patient’s willingness to participate and to identify any medical requirements that may preclude enrollment.

Additional assessments were conducted to evaluate each patient’s willingness to participate and to identify any medical requirements that may preclude enrollment. Willingness was assessed by HaH nurses using a structured enrollment discussion with patients and caregivers based on a standardized screening instrument. Data on demographic and clinical characteristics, including diagnosis, sex, age, and ED length of stay, were also collected.

Population and data collection for the DHA model

All patients within the DHA model were referred by partner home care agencies following initial clinical evaluations. Upon referral, HaH nurses conducted in-home preadmission assessments to confirm whether each patient met the preestablished clinical and functional eligibility criteria. Once eligibility was confirmed, a formal admission date was assigned, and an HaH physician conducted an initial home visit to initiate the treatment plan. Given that our institution does not offer HaH services for patients staying in long-term care facilities, we could not recruit or manage cases from this specific patient population.

Data collection for successful enrollment through the EPA and DHA model

Data were collected on patients who were successfully enrolled in the HaH program through the EPA and DHA models. Both EPA and DHA pathways operate under the same national HaH pilot framework with shared eligibility criteria and care protocols; therefore, findings from both models were analyzed together to reflect real-world implementation. These data included primary diagnosis, age, sex, HaH treatment duration, treatment outcome, ventilator dependency, 14-day ED revisits, and outcomes of long-term care referrals.

Statistical analysis

All statistical analyses were conducted using IBM SPSS Statistics version 25.0 (IBM, Armonk, NY, USA), with a significance level set at P <.05. The normality of the variables under examination was evaluated using the Kolmogorov-Smirnov test. Continuous variables are presented as mean ± standard deviation values. Categorical variables were analyzed using the chi-squared or Fisher’s exact test, as appropriate, and are reported as frequencies and percentages. A binary logistic regression analysis was used to identify the factors associated with HaH eligibility through the EPA model for ED person-days involving pneumonia, UTI, or cellulitis. The independent variables included sex, age, and ED length of stay. Univariate logistic regression models were used to examine the association between each predictor and HaH eligibility. Variables identified as statistically significant in the univariate analyses were entered into multivariate logistic regression models to estimate the adjusted odds ratios (aORs) and 95% confidence intervals (CIs). All patients who were successfully enrolled in the HaH program were registered in a consolidated review list, regardless of whether they were admitted through the EPA or DHA model.

Results

Screening outcomes and eligibility for HaH

A total of 1462 ED person-days for conditions such as pneumonia, UTI, or cellulitis were screened by the HaH team between September 2024 and June 2025, representing 639 unique patients. Of these person-days, 828 (56.6%) were attributable to male patients and 634 (43.4%) were attributable to female patients. The overall mean age of the study sample was 68.2 ± 16.8 years, and the average length of ED stay before HaH screening was 2.3 ± 1.6 days. Pneumonia accounted for the majority of diagnoses (56.5%), followed by cellulitis (24.7%) and UTI (18.8%). Among all ED person-days, 1282 (87.7%) were deemed ineligible for HaH care, and only 180 (12.3%) met the eligibility criteria. Nevertheless, the actual enrollment rate remained extremely low, with only 6 person-days (0.4%) resulting in HaH program admission through the EPA model (Fig. 2).

Fig. 2.

Flowchart of Person-Days Screening and Enrollment Outcomes in the HaH Program Through the Emergency Passed Admission Model ED, emergency department; HaH, hospital at home; UTI, urinary tract infection. ^aSee Fig. 2 for detailed exclusion reasons

A multivariate logistic regression analysis revealed diagnosis-specific factors associated with HaH program eligibility. Among those with pneumonia, female patients were significantly less likely than their male counterparts to meet the eligibility criteria for the HaH program (aOR = 0.42, 95% CI = 0.34–0.52). Higher age (aOR = 1.03, 95% CI = 1.02–1.04) was associated with higher odds of eligibility, whereas longer ED length of stay was associated with lower odds of eligibility (aOR = 0.93, 95% CI = 0.87–0.99). Only sex was identified as a significant predictor of UTI. Among those with UTI, female patients were more likely than their male counterparts to meet the eligibility criteria for the HaH program (aOR = 4.75, 95% CI = 3.54–6.36). Among those with cellulitis, older age was associated with lower eligibility (aOR = 0.97, 95% CI = 0.96–0.97), whereas longer ED length of stay was associated with higher eligibility (aOR = 1.18, 95% CI = 1.09–1.27) (Table 1).

Table 1.

Characteristics of ED Person-Days Stratified by HaH Eligibility and Diagnosis

	Overall (n = 1462)	Ineligible for HaH (n = 1282)	Eligible for HaH (n = 180)	Unadjusted OR (95% CI)	Adjusted OR (95% CI)
Pneumonia, n (%)	826 (56.5)	734 (57.3)	92 (51.1)
Sex, female/male	293/533	266/468	27/65	0.48 (0.39–0.59)	0.42 (0.34–0.52)
Age, years	71.2 ± 15.8	70.4 ± 16.1	77.6 ± 10.6	1.03 (1.02–1.03)	1.03 (1.02–1.04)
ED length of stay, days	2.2 ± 1.5	2.2 ± 1.5	2.4 ± 1.6	0.93 (0.88–1.00)	0.93 (0.87–0.99)
UTI, n (%)	275 (18.8)	231 (18.0)	44 (24.4)
Sex, female/male	200/75	164/67	36/8	4.63 (3.46–6.19)	4.75 (3.54–6.36)
Age, years	68.6 ± 18.3	66.1 ± 18.8	81.3 ± 6.9	1.00 (0.99–1.01)	1.00 (0.99–1.01)
ED length of stay, days	2.2 ± 1.5	2.1 ± 1.4	2.6 ± 2.0	0.95 (0.87–1.03)	0.92 (0.84–1.00)
Cellulitis, n (%)	361 (24.7)	317 (24.7)	44 (24.4)
Sex, female/male	141/220	126/191	15/29	0.79 (0.62–1.01)	0.84 (0.65–1.08)
Age, years	60.9 ± 15.8	59.4 ± 15	71.6 ± 17.6	0.97 (0.96–0.97)	0.97 (0.96–0.97)
ED length of stay, days	2.5 ± 1.9	2.5 ± 1.8	2.6 ± 2.2	1.13 (1.06–1.22)	1.18 (1.09–1.27)

Data are presented as mean ± standard deviation or as number (%)

ED, emergency department; HaH, hospital at home; OR, odds ratio; CI, confidence interval; UTI, urinary tract infection

Reasons for ineligibility and recruitment failure

Of the 1282 ineligible ED person-days, 649 (50.6%) were excluded because the patients had insufficient functional impairments (BI > 60) and 633 (49.4%) were excluded because the patients had unstable clinical conditions that did not meet the HaH eligibility criteria. Among those with a BI of > 60, pneumonia (41.6%) and cellulitis (40.2%) were the most common diagnoses, with UTI accounting for 18.2% of the cases. Among those with unstable clinical conditions, pneumonia (73.3%) was the most common diagnosis, with UTI and cellulitis together accounting for only 26.7% of the cases. Notably, the average age was significantly higher in those with unstable clinical conditions than in those who had insufficient functional impairments (P <.01 for all).

Among ED person-days involving UTI, the mean ED length of stay was significantly longer in those with unstable clinical conditions than in those with a BI of > 60 (P <.01). Among ED person-days involving cellulitis, the proportion of female patients was significantly lower in those with a BI of > 60 than in those with unstable clinical conditions (33.0% vs. 71.4%, P <.001; Table 2).

Table 2.

Characteristics of Patients Ineligible for the HaH Program (n = 1282)

	Barthel index > 60 (n = 649)	Unstable clinical condition (n = 633)	P value
Pneumonia), n (%	270 (41.6)	464 (73.3)
Sex, female/male	86/184	180/284	0.059
Age, years	58.4 ± 14.9	77.4 ± 12.2	< 0.001***
ED length of stay, days	2.1 ± 1.4	2.2 ± 1.6	0.122
UTI, n (%)	118 (18.2)	113 (17.9)
Sex, female/male	89/29	75/38	0.130
Age, years	54.4 ± 16.6	78.4 ± 11.8	< 0.001***
ED length of stay, days	1.8 ± 1.3	2.4 ± 1.4	< 0.01**
Cellulitis, n (%)	261 (40.2)	56 (8.8)
Sex, female/male	86/175	40/16	< 0.001***
Age, years	56.5 ± 13.4	72.9 ± 15.1	< 0.001***
ED length of stay, days	2.5 ± 1.7	2.9 ± 2.3	0.157

^aChi-square test

^*P <.05; ^**P <.01; ^***P <.001

Data are presented as mean ± standard deviation or as number (%)

HaH, hospital at home; ED, emergency department; UTI, urinary tract infection

Among 174 eligible but nonenrolled ED person-days, the main reason for recruitment failure was the presence of other medical needs requiring inpatient care, such as blood transfusion, chemotherapy, surgery, or hemodialysis (n = 110, 63.2%). Additional reasons included patient preference for further inpatient evaluation and monitoring (n = 33, 19%) and the lack of caregiving support at home, for example, for those living alone or having no caregivers (n = 29, 16.7%). Two person-days (1.1%) were excluded because of concerns regarding long travel distances from the hospital, which led to problems related to transportation costs (Fig. 3).

Fig. 3.

Distribution of the reasons for enrollment failure among eligible hospital-at-home person-days through the emergency passed admission model

Enrollment outcomes in EPA and DHA models

A total of 7 ventilator-dependent patients were referred through the DHA model to our HaH team. Of these patients, 2 were found to have unstable clinical conditions, including tachycardia, hypotension, and tachypnea, during their in-home preadmission assessments, which were conducted by our HaH nurses. Ultimately, 5 patients were successfully enrolled, resulting in an enrolled rate of 71.4% through the DHA model.

A total of 6 and 5 patients were successfully enrolled through the EPA and DHA models, respectively, resulting in a combined total of 11 admissions. All 5 patients who were admitted through the DHA model were ventilator-dependent. Additionally, 7 out of all 11 patients had documented long-term care needs. Among all patients, 10 completed their HaH treatment with stable follow-up outcomes, and only one patient required an ED revisit within 14 days of discharge (Table 3).

Table 3.

Successfully Enrolled Hospital-at-Home Cases

No.	Enrollment model	Age, years	Sex	Primary diagnosis	Secondary diagnosis	Duration of care, days	Ventilator dependency	RT visit	Outcome	14-day ED revisit	Referred to long-term care
01	EPA	104	M	Cellulitis	—	5	No	No	Completed, OPD follow-up	No	No, not required
02	EPA	84	F	UTI	—	5	No	No	Completed, OPD follow-up	No	No, not required
03	EPA	79	F	UTI	—	3	No	No	Completed, OPD follow-up	No	Yes, services already in place
04	EPA	66	M	Pneumonia	Tongue cancer	8	No	Yes	Completed, OPD follow-up	No	Yes, services already in place
05	DHA	74	M	Cellulitis	COPD	8	Yes, home NIV through face mask	Yes	Completed, OPD follow-up	No	Yes, services already in place
06	DHA	70	M	Pneumonia	ALS	12	Yes, home MV through Tr	Yes	Completed, OPD follow-up	No	Yes, services already in place
07	EPA	91	F	Pneumonia	—	7	No	Yes	Completed, OPD follow-up	No	Yes, services initiated
08	DHA	71	F	Cellulitis	Lung cancer	8	Yes, home NIV through face mask	Yes	Completed, OPD follow-up	No	Yes, services already in place
09	DHA	62	F	UTI	—	8	Yes, home NIV through face mask	Yes	Completed, OPD follow-up	No	No, not eligible
10	EPA	74	M	Cellulitis	—	8	No	No	Completed, OPD follow-up	Yes, increased wound exudate	No, not eligible
11	DHA	79	M	Pneumonia	COPD, type II DM	13	Yes, home NIV through face mask	Yes	Completed, OPD follow-up	No	Yes, services already in place

RT, respiratory therapist; ED, emergency department; EPA, emergency passed admission; OPD, outpatient department; UTI: urinary tract infection; DHA, direct home admission; COPD, chronic obstructive pulmonary disease; NIV, noninvasive ventilation; ALS, amyotrophic lateral sclerosis; MV, mechanical ventilation; Tr, tracheostomy; DM, diabetes mellitus

Discussion

This single-center study conducted during the early implementation phase of the HaH pilot program in Taiwan identified substantial recruitment and enrollment barriers. Only a small proportion of ED person-days met eligibility criteria, and the actual enrollment rate through the EPA model remained extremely low. Insufficient functional impairment and clinical instability were the primary reasons for exclusion. Despite the limited number of admissions, patients enrolled through both the EPA and DHA models demonstrated favorable short-term clinical outcomes, supporting the feasibility and safety of home-based acute care for appropriately selected populations. However, these findings reflect the early experience of a single institution within a specific region and may be influenced by local health care capacity and program maturity; therefore, they should not be directly generalized to the national HaH implementation in Taiwan.

In 2025, the National Health Insurance (NHI) Administration of Taiwan released updated data on its HaH program. As of May 31, 2025, a total of 174 multidisciplinary care teams and 789 health care institutions had been authorized to engage in this nationwide program [16]. Among nearly 3000 enrolled patients, 70% were aged ≥ 75 years and 78% were completely dependent in activities of daily living (BI < 20). These national statistics confirm that the HaH program primarily serves frail older adults with significant functional limitations.

Based on these nationwide NHI data, the most common diagnoses for HaH admission were UTI (49%), pneumonia (37%), and cellulitis (14%). The reported average duration of care also varied by diagnosis, with pneumonia requiring 7.9 days and both UTI and cellulitis requiring 6.2 days. The overall mean duration of care was 6.9 days, which was shorter than the national average inpatient length of stay of 12 days.

According to the NHI report, this reduction corresponded to an estimated decrease of at least 106 inpatient bed-days per day and an average expenditure of approximately 30 000 NHI points per HaH episode, compared with about 69 000 points for conventional hospitalization. Only 5% of patients required an ED revisit within 14 days of discharge. These national findings indicate that HaH care may provide a financially sustainable alternative to traditional hospitalization while maintaining quality of care [16].

To the best of our knowledge, this is the first study to explore the recruitment and enrollment barriers encountered in the pilot HaH program of Taiwan. Generally, our HaH team has dedicated personnel responsible for patient screening on a daily basis in the ED. Despite this team’s effort, the enrollment success rate through the EPA model remains only 0.4% among 1462 ED person-days assessed during weekdays. This low rate suggests that age, sex, and ED length of stay play differential roles across diagnoses. In contrast to this low enrollment rate, we observed that the DHA model achieved an enrollment success rate of 71.4%, even though the number of enrolled patients was similar (6 patients through the EPA model and 5 patients through the DHA model). This discrepancy was primarily attributable to the fact that all patients enrolled through the DHA model were ventilator-dependent individuals who were already receiving home care through home care agencies. These agencies could effectively conduct preliminary evaluations. To mitigate the risk of recruitment failure and avoid unnecessary physician visits, our team introduced an intermediate step in which a nurse conducted an initial home visit. This visit included assessments such as physical evaluation, vital sign monitoring, bedside C-reactive protein level measurement, and respiratory sound recording with an electronic stethoscope, which was later reviewed by a physician. These ventilator-dependent patients were already familiar with home-based medical services and typically demonstrated a sufficient self-care capacity, which eliminated concerns regarding caregiver availability. In addition, their willingness to participate in the HaH program was consistently high. This high willingness was particularly evident in a patient with amyotrophic lateral sclerosis, for whom hospital transfer posed major logistic challenges. For patients such as this one, the DHA model provided a meaningful improvement in care accessibility. Notably, one of the patients enrolled in the HaH program through the EPA model was 104 years old. This patient’s family and caregivers strongly supported the patient’s early discharge from the ED and welcomed the HaH care option to avoid prolonged ED stays (Table 3).

Of the 11 patients who were successfully enrolled in the HaH program, 7 (63.6%) were either receiving long-term care services or had recently applied for them, indicating a major overlap between those receiving HaH care and those receiving long-term care services. This overlap primarily originated from the eligibility criteria established for HaH care, which mandate a moderate to high level of functional impairment (BI < 60). Notably, all patients completed their treatment at home without major clinical complications, and only one patient (9.1%) required an ED revisit within 14 days of discharge. This low rate of postdischarge acute care utilization suggests that the inclusion criteria and screening procedures adopted were effective in identifying clinically stable individuals with a low risk of deterioration. Collectively, these findings indicate the safety, feasibility, and clinical appropriateness of delivering acute care at home for older adults with manageable medical conditions who would otherwise require hospitalization.

Despite our limited sample size, the average duration of care in our cohort was 7.7 days, which is substantially shorter than the national average duration of inpatient hospitalization. When stratified by diagnosis, the average duration of care was found to be 10 days for pneumonia, 5.3 days for UTI, and 7.3 days for cellulitis, indicating the HaH model’s capacity to tailor care depending on clinical presentation. From an economic perspective, the average cost per case was calculated to be approximately 46 000 NHI points, which is considerably lower than the estimated cost of inpatient services. Only one patient required an ED revisit within 14 days after discharge, indicating appropriate patient selection and stable clinical outcomes in home-based care settings. The monetary value of each NHI point was adjusted quarterly, typically ranging between NT$0.8 and NT$1.2, depending on national policy and budgeting [17]. Collectively, these findings confirm the feasibility, safety, and economic benefits of the HaH program, particularly for older adults with acute yet manageable medical conditions who would otherwise require hospitalization.

Among the 1282 ED person-days that were deemed ineligible for enrollment, 649 (50.6%) were excluded because the patients had a BI of > 60. From a policy perspective, adjusting functional criteria may expand the eligible population, thereby increasing overall enrollment rates and enhancing program reach. In this study, the remaining 633 ED person-days (49.4%) were excluded because the patients had unstable clinical conditions that did not meet the eligibility criteria for the HaH program. This finding highlights the indirect association between patient exclusion and the experience of the clinical team. Teams with greater experience and more extensive home care abilities may be more confident in managing patients with moderately unstable conditions, thereby expanding the effective criteria for eligibility. In the present study, most of the patients were at an advanced age, which increased the clinical complexity of their cases. Therefore, motivating HaH care teams to invest in their skills and increase their interdisciplinary capacity may be essential to increase their willingness and readiness to admit more complex cases.

In addition to clinical eligibility, the decision to enroll patients in HaH programs involves social and environmental factors, such as the availability of caregiving support at home, housing conditions, and patient or caregiver willingness [18]. In our study, of 174 ED person-days that met the eligibility criteria for HaH program enrollment, only 1.1% were excluded because of concerns pertaining to long travel distances and associated transportation costs. Additionally, 16.7% were excluded because the patients either lived alone or lacked caregiver support, and 19% were excluded because the patients preferred inpatient admission for further diagnostic evaluations. Moreover, the leading cause of recruitment failure among eligible patients was identified as the presence of other medical needs requiring inpatient care, such as blood transfusion, chemotherapy, surgery, or dialysis. These care demands are often influenced by the local development of medical infrastructure and the availability of health care resources. In recent years, advancements in medical technology have facilitated the expansion of home-based services, such as home-based chemotherapy [19–21], home-based blood transfusion [22, 23], and even telehealth-guided maggot debridement therapy [24].

In 2008, Presbyterian Healthcare Services in the United States adopted the HaH model developed by the Johns Hopkins University Schools of Medicine and Public Health to deliver more tailored acute care services in home settings. Instead of delivering conventional inpatient care to patients with similar conditions, this model delivered home-based care that resulted in more favorable clinical outcomes, lower fall rates, and higher patient satisfaction. It also resulted in a 19% reduction in medical expenditures [25]. In 2014, the Icahn School of Medicine at Mount Sinai received funding from the Center for Medicare and Medicaid Innovation to evaluate the clinical effectiveness of an HaH program integrated with a 30-day post-acute home transition care component. The results indicated that patients who were enrolled in this program had lower readmission rates, fewer ED revisits, fewer skilled nursing facility admissions, and shorter lengths of stay. Overall, these findings validate the feasibility and clinical advantages of HaH programs in the context of modern health care practice and technology [26].

HaH models have emerged as a promising integrated care approach that warrants further validation through comprehensive large-scale clinical investigations. Although preliminary evidence suggests favorable outcomes in terms of safety, feasibility, and cost-effectiveness, systematic evaluations of the long-term clinical performance and patient-centered outcomes of these models remain essential. Documenting treatment efficacy, adverse events, and health care utilization metrics can help guide future health policy decisions and support the sustainable expansion of HaH models. Overall, recent advancements in health care technology and service delivery have improved the operational feasibility of HaH programs. This improvement has contributed to growing acceptance of HaH models among both health care professionals and patients. Consequently, the clinical scope of HaH care continues to expand, offering a feasible alternative to conventional hospitalization for certain patient populations. Further research coupled with supportive policy frameworks is required to ensure the responsible and effective integration of HaH services into national health systems [27].

This study has several limitations. First, this study was conducted at a single region in New Taipei City, Taiwan. Because of the differences in health care ecosystems and residential patterns across Taiwan, our findings may not be generalizable to other regions or countries. Second, data were collected only during weekdays because of the limited staff capacity, and no weekend assessments were conducted. Third, the definition of an unstable clinical condition may vary across HaH teams depending on their clinical resources and operational protocols. Fourth, certain inpatient procedures such as dialysis, transfusion, and chemotherapy are typically provided in hospitals. Their implementation in HaH settings heavily depends on the medical capabilities of each care team and the local health care infrastructure. Fifth, because the HaH program was still in its initial implementation phase during the study period, the limited awareness and familiarity of patients, ED personnel, and attending physicians with this program may have led to hesitancy or a lack of trust in it, which may have influenced the final enrollment outcomes. Sixth, the number of successfully enrolled patients was limited, particularly under the EPA model, which may have reduced the statistical power for outcome evaluation. The primary objective of this study was to examine eligibility patterns and recruitment barriers rather than to identify predictors of enrollment success or clinical outcomes among enrolled patients. Therefore, findings related to enrolled cases should be interpreted cautiously and regarded as exploratory observations. Finally, HaH care represents a single alternative to hospital-based care. Other internationally recognized models, such as early supported discharge, have been found to facilitate timely transitions from inpatient to home-based care by integrating acute and subacute treatment in home settings. These models are designed to shorten hospital length of stay while ensuring continuity of care through the delivery of coordinated follow-up services. However, they are not currently included in the national HaH program of Taiwan. Therefore, we could not evaluate or compare their effectiveness, feasibility, or implementation outcomes. Future studies and health policy planning efforts may benefit from the integration of early supported discharge principles into the current HaH framework to further strengthen the continuum of care. In addition, prospective qualitative studies are warranted to further explore patient and caregiver decision-making and barriers to HaH enrollment.

In this study, we examined the key factors contributing to enrollment failure in HaH programs, particularly through the EPA model. We identified functional status exceeding eligibility thresholds and unstable clinical conditions as the most common exclusion criteria. Our findings underscore the importance of patient selection criteria, care team capacity, and clearly defined clinical stability thresholds in influencing enrollment outcomes. Despite these barriers, all patients who met the stringent eligibility criteria successfully completed their treatment without major complications. Future studies should focus on refining inclusion criteria and enhancing clinical readiness to improve the recruitment outcomes and expand the reach of home-based acute care services.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

HaH

Hospital-at-home

EPA

Emergency passed admission

DHA

Direct home admission

ED

Emergency department

BI

Barthel index

UTI

Urinary tract infection

aOR

adjusted odds ratio

Author contributions

KYC, TYT, TLW, HPF, CYL, SHH conceived and designed the study. KYC, CYC, YZC, YCC, HSS, WRC, conducted the experimental work and collected the data. KYC, TYT, YZC, and CYL contributed to data analysis. KYC, TYT, WLL, JJH drafted and revised the manuscript. All authors have approved the final version for publication.

Funding

This study received funding from Fu Jen Catholic University (A0114001). The funder had no role in the study design; the collection, analysis, or interpretation of the data; the drafting of the manuscript; or the decision to submit the manuscript for publication.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethical approval

This study was conducted in accordance with the Good Clinical Practice guidelines and the principles of the Declaration of Helsinki. The study protocol was reviewed and approved by the Institutional Review Board of Fu Jen Catholic University Hospital, New Taipei City, Taiwan (approval no. FJUH114533). Owing to the retrospective nature of the study, the requirement for written informed consent from participants was waived by the Institutional Review Board.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.