Lin et al.

BMC Medicine (2024) 22:250 BMC Medicine

https://doi.org/10.1186/s12916-024-03463-3

RESEARCH ARTICLE Open Access

Early discharge hospital at home

as alternative to routine hospital care for older
people: a systematic review and meta‑analysis
Lulu Lin1†, Mengyuan Cheng2,3†, Yawei Guo1†, Xiaowen Cao2, Weiming Tang2,4, Xin Xu1, Weibin Cheng2,5,6* and
Zhongzhi Xu1*

Abstract
Background The global population of adults aged 60 and above surpassed 1 billion in 2020, constituting 13.5%
of the global populace. Projections indicate a rise to 2.1 billion by 2050. While Hospital-at-Home (HaH) programs
have emerged as a promising alternative to traditional routine hospital care, showing initial benefits in metrics such
as lower mortality rates, reduced readmission rates, shorter treatment durations, and improved mental and functional
status among older individuals, the robustness and magnitude of these effects relative to conventional hospital set-
tings call for further validation through a comprehensive meta-analysis.
Methods A comprehensive literature search was executed during April–June 2023, across PubMed, MEDLINE,
Embase, Web of Science, and Cumulative Index of Nursing and Allied Health Literature (CINAHL) to include both RCT
and non-RCT HaH studies. Statistical analyses were conducted using Review Manager (version 5.4), with Forest plots
and I2 statistics employed to detect inter-study heterogeneity. For I2 > 50%, indicative of substantial heterogene-
ity among the included studies, we employed the random-effects model to account for the variability. For I2 ≤ 50%,
we used the fixed effects model. Subgroup analyses were conducted in patients with different health conditions,
including cancer, acute medical conditions, chronic medical conditions, orthopedic issues, and medically complex
conditions.
Results Fifteen trials were included in this systematic review, including 7 RCTs and 8 non-RCTs. Outcome measures
include mortality, readmission rates, treatment duration, functional status (measured by the Barthel index), and men-
tal status (measured by MMSE). Results suggest that early discharge HaH is linked to decreased mortality, albeit sup-
ported by low-certainty evidence across 13 studies. It also shortens the length of treatment, corroborated by seven
trials. However, its impact on readmission rates and mental status remains inconclusive, supported by nine and two
trials respectively. Functional status, gauged by the Barthel index, indicated potential decline with early discharge
HaH, according to four trials. Subgroup analyses reveal similar trends.

†
Lulu Lin, Mengyuan Cheng, and Yawei Guo contributed equally to this work.
*Correspondence:
Weibin Cheng
[email protected]
Zhongzhi Xu
[email protected]
Full list of author information is available at the end of the article

© The Author(s) 2024. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the
original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or
other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this
licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecom-
mons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Lin et al. BMC Medicine (2024) 22:250 Page 2 of 15

Conclusions While early discharge HaH shows promise in specific metrics like mortality and treatment duration, its
utility is ambiguous in the contexts of readmission, mental status, and functional status, necessitating cautious inter-
pretation of findings.
Keywords Early discharge hospital at home, Home care services, Older people, Meta-analysis

Background such as mortality rates, readmission frequencies, dura-

Hospital-at-Home (HaH) emerges as an innovative tions of treatment, and mental and functional statuses.
healthcare model, redefining the boundaries of hospital Our exploration is channeled towards discerning the
care by extending clinical management to the patient’s impact of these programs on older adults, aged 60 and
residence. Leveraging a meticulous blend of early dis- above, within the architectural frameworks of tradi-
charge and admission avoidance strategies, HaH is driven tional inpatient and HaH care paradigms. This endeavor
by specialized teams that conduct comprehensive health is inspired by a commitment to enriching the empirical
and rehabilitative evaluations, either remotely or through foundations that guide the optimization of HaH strate-
home visits, ensuring the delivery of nuanced and gies in alignment with the evolving contours of patient
patient-centric care [1]. needs and healthcare excellence.
The inception of HaH is particularly poignant against
the backdrop of a burgeoning aging population, heralding Methods
a paradigm that fosters the efficient allocation of health- This systematic review and meta-analysis adhered to
care resources while accentuating the centrality of patient the Preferred Reporting of Items for Systematic Reviews
welfare [2]. It has garnered notable attention and applica- and Meta-Analyses (PRISMA) guidelines, as detailed in
tion, particularly in the care of older adult patients grap- Additional file 1: Table S1. The methodology employed
pling with a spectrum of conditions such as orthopedic in this review was guided by the Cochrane Handbook for
anomalies [3] and chronic obstructive pulmonary disease Systematic Reviews of Interventions. This study was reg-
(COPD) [4]. istered with PROSPERO (registration number: 321343).
A synthesis of empirical explorations into HaH under- The protocol can be accessed at https://​www.​crd.​york.​ac.​
scores its potential to recalibrate the cost-effectiveness uk/​PROSP​ERO/​displ​ay_​record.​php?​Recor​dID=​321343.
landscape of healthcare delivery. The model, through its
emphasis on early hospital discharge, appears to nur-
Search strategy and eligibility criteria
ture an ecosystem that not only preserves but poten-
From April to June 2023, we conducted a rigorous search
tially enhances the quality of clinical outcomes [3, 5]. For
across multiple electronic databases to locate studies
example, a randomized controlled trial demonstrated
published subsequent to Gonçalves-Bradley DC’s 2017
that a short hospital stay followed by a well-managed
review [9]. This encompassed peer-reviewed articles
home care program is as effective as a traditional 10-day
from databases such as PubMed, MEDLINE, Embase,
hospitalization course. This approach not only reduces
Web of Science, and CINAHL. Additionally, our search
hospitalization costs but also fosters closer relationships
was comprehensive, including grey literature such as
between patients and their relatives [6].
abstracts and conference proceedings, and a manual
However, the trajectory of HaH is not without its inter-
review of references from relevant studies and key trial
sections of uncertainty and contention. Notable voices in
registries like ClinicalTrials.gov.
the academic discourse have raised concerns regarding
We used a combination of Medical Subject Headings
the sufficiency and robustness of evidence elucidating the
(MeSH) and keywords relevant to hospital-sponsored
comparative impacts of care environments on the reha-
healthcare services in the home setting. All studies con-
bilitation outcomes of older individuals [7]. A nuanced
sidered were published in English. The search strategy,
examination of existing literature reveals a confluence of
detailed in Additional file 1: Text S1, resulted in 5796
findings, where HaH programs, despite their transforma-
entries, shaping the pool from which we selected suitable
tive potential, echo with resonances of variability and
literature for our review.
ambiguity, particularly concerning readmission timelines
Upon the elimination of duplicate entries from the ini-
[8].
tial database search, two authors independently reviewed
In navigating these complexities, our study embarks on
titles and abstracts for eligibility. Any discrepancies in the
a systematic review and meta-analysis aimed at unrave-
selection process were first resolved through discussion;
ling the comparative efficacies of early discharge HaH
programs. Anchoring our inquiry are important metrics
Lin et al. BMC Medicine (2024) 22:250 Page 3 of 15

if a consensus cannot be reached, a third author would Disagreements were resolved through consensus or
arbitrate. consultation with a third author if necessary.

Eligibility criteria Data extraction

Our review targeted studies that evaluated the efficacy Two independent reviewers are responsible for extract-
or effectiveness of hospital-sponsored health care ser- ing relevant data from studies that meet the inclusion
vices provided in patients’ homes compared to usual criteria. Information such as first author, publication
in-hospital care. Eligible studies reported or permitted year, geographical setting, study design, sample size,
the extraction of raw data for one or more of our pri- duration of follow-up, classification of caregivers, types
mary outcomes. We focused on studies involving older of health services provided, utilization of telemedicine
adults (aged ≥ 60 years) or those with a subgroup of technologies, control groups, and outcomes meas-
individuals aged ≥ 60 for whom results were separately ured were systematically documented using an elec-
reported. We imposed no disease-specific restrictions tronic data collection form, accessible in the Additional
to assess the home hospital programs’ efficacy across file 1: Table. S2. In cases of missing or ambiguous data,
diverse health care needs. To ensure a thorough anal- inquiries were directed to the corresponding authors
ysis, both randomized controlled trials (RCTs) and for clarification.
non-randomized controlled trials (non-RCTs) were
included.
Quality assessment
In assessing the quality of the included studies, we
Inclusion criteria employed the Version 2 of the Cochrane risk-of-bias
Population: Individuals aged 60 years or older receiv- tool (RoB2 tool) [10] for randomized trials and the Risk
ing health care services at home, who would otherwise Of Bias In Non-randomised Studies—of Interventions
require hospital care. Intervention: Health care services (ROBINS-I tool) [11] for non-randomized studies to
provided by physicians or nurses during acute or non- evaluate potential biases systematically. Our assessment
acute phases of illness at the patient’s home. Compara- covered biases from the randomization process, devia-
tor: Standard inpatient hospital care. tions from intended interventions, missing outcome data,
measurement of outcomes, and the reporting of results.
Exclusion criteria Each domain of bias was judged according to the risk
Excluded were studies focusing on outpatient care, level: low, some concerns, or high. The overall quality of
residential care settings, or primarily involving patient evidence for the outcomes reported in the studies was
self-care at home. Programs offering end-of-life care, appraised using the RoB2 for RCTs and the ROBINS-I
social services (e.g., assistance with daily living), or for non-RCTs and classified into three levels—high, some
transitional “hospital to home” care were also omitted. concerns, and low. Discrepancies in the quality assess-
Review articles, commentaries, and study protocols ment were resolved through discussion or by involving
were excluded due to the absence of outcome data. a third author for consensus. We visualized these assess-
ments using Risk of Bias VISualization (robvis) tools [12]
Outcome measures to aid in the clear presentation of our findings. Although
Primary outcomes included mortality rates during the certain domains within some studies raised “some con-
study period or at specific time points, return hospi- cerns,” the collective evaluation of these studies generally
tal rates (admissions post-HaH or readmissions after indicates a low risk of bias.
inpatient care), functional ability measured by the Bar-
thel Index, and quality of life assessed via standardized
Statistical analyses
questionnaires like SF-36 or EQ-5D. Secondary out-
In conducting a meta-analysis, the conventional practice
comes encompassed cognitive function and depression
typically involves segregating RCTs from non-RCTs due
levels.
to the inherent differences in study design and potential
for bias. This separation is rooted in the aim to ensure
Study selection
the integrity and reliability of the analysis by comparing
The selection process, documented via a PRISMA
like with like. However, when such a distinction results
flow diagram (Fig. 1), began with removing duplicates
in an insufficient number of studies within each category,
from database searches. Two authors independently
the meta-analysis may face challenges related to statisti-
screened titles and abstracts for eligibility, with full-
cal power. There is a trade-off between methodological
text articles reviewed for those preliminarily selected.
purity and the practical necessity of accruing sufficient
Lin et al. BMC Medicine (2024) 22:250 Page 4 of 15

Fig. 1 Flow diagram of study selection process

data to enable a meaningful analysis. To this end, main Manager (Version 5.4), with Forest plots and I2 statis-
analyses were conducted by pooling RCT data and non- tics employed to detect inter-study heterogeneity. For
RCT data separately. For secondary analyses targeting I2 > 50%, indicative of substantial heterogeneity among
specific diseases in this study, we have not conducted the included studies, we employed the random-effects
separate analyses due to an insufficient number of studies model to account for the variability. For I2 ≤ 50%, we
within each category, which could compromise the statis- used the fixed effects model. Substantial heterogene-
tical power of the meta-analysis. ity was further examined, and subgroup analyses were
For datasets exhibiting homogeneity, a pooled meta- conducted for randomized controlled trials and specific
analysis was undertaken. This homogeneity was assessed patient subgroups.
according to the types of outcomes and their respec-
tive measurement time points across studies. Categori- Results
cal outcomes such as mortality were presented as risk Literature search and studies included
ratios with 95% confidence intervals (CIs), while con- A total of 9628 records were identified through our
tinuous outcomes were expressed as mean differences database search. Of these, 8750 were excluded because
(MDs) and 95% CIs. If various measurement techniques they are out of the scope of this analysis. This left 923
were employed, standardized MDs and 95% CIs were records for full-text assessment, from which 892 were
used. Statistical analyses were conducted using Review excluded for various reasons, including incorrect setting,
Lin et al. BMC Medicine (2024) 22:250 Page 5 of 15

inappropriate intervention, and unavailability of full text. and non-RCTs (I2 = 72%) suggests variability in the study
Consequently, 30 studies met the inclusion criteria. Of outcomes, underscoring the need for cautious interpreta-
these, 15 focused solely on “admission avoidance HaH.” tion (Fig. 5).
The remaining 15 studies that highlighted “early dis-
charge HaH” were included in this review. Figure 1 dem- Length of treatment
onstrates the selection flow. The forest plot for length of treatment shows that in
Fifteen trials with a cumulative sample size of 4,190 RCTs, the HaH group exhibits no significant difference
patients were incorporated into this study. The trials compared to routine care, with a mean difference of 0.02
involved patients with various health conditions, includ- (95% CI: − 0.98 to 1.03) and low heterogeneity (I2 = 8%).
ing cancer [13], acute medical conditions [5, 14–17], In contrast, non-RCTs demonstrate a significant reduc-
chronic medical conditions [4, 18, 19], orthopedic issues tion in treatment duration for HaH with a mean dif-
[18, 20–23], and medically complex conditions [18, 24, ference of − 1.66 (95% CI: − 3.18 to − 0.14) but with
25]. Of the 15 studies, 10 were trials related to early dis- substantial heterogeneity (I2 = 90%) (Fig. 6).
charge HaH [4, 14, 15, 19–25]. Meanwhile, five trials
studied HaH interventions that incorporated both early Functional status — Barthel index at the end up
discharge and admission avoidance [15, 17, 18, 21, 24]. Figure 7 indicates a notable reduction in treatment
Seven studies employed a randomized controlled trial length with HaH when compared to standard hospi-
design, one used a prospective quasi-experiment [5], one tal care, as shown by a mean difference of − 1.82 (95%
was a prospective, non-randomized real-world cohort CI: − 2.55, − 1.09) with negligible heterogeneity (I2 = 0%)
comparison [13], one was a retrospective study [22], and in non-RCTs. A meta-analysis for RCTs on this outcome
one was a quasi-experimental longitudinal study [21]. was not performed due to the presence of only a single
The characteristics of the included studies are shown in study.
Additional file 1: Table S2. Meta-analysis related to early
discharge HaH program in the past 20 years is shown in Mental status — MMSE
Additional file 1: Table S3. For the assessment of mental status using MMSE, our
analysis considered one randomized controlled trial
Quality of included studies (RCT) [14] and one retrospective study [22]. Due to the
The risk of bias for RCT and non-RCT studies is shown presence of only one study in each category, we are not
in Figs. 2 and 3. Results demonstrate that, although cer- able to pool the results. The findings from both the ret-
tain domains within some studies raised “some concerns,” rospective study and the RCT indicated no significant
the collective evaluation of these studies generally indi- difference in MMSE scores between the case group and
cates a low risk of bias. the control group. Specifically, the retrospective study
reported a mean difference of − 1.24 (95% CI: − 3.30 to
Main analysis 0.82, N = 235; P = 0.24), and the RCT showed a mean dif-
Mortality ference of − 2.59 (95% CI: − 5.45 to 0.27, N = 70; P = 0.08).
The forest plot analysis for mortality outcomes compar-
ing early discharge HaH programs to routine hospital Subgroup analyses based on disease classification
care reveals a nuanced picture: RCTs show no significant Subgroup analyses were conducted in patients with dif-
difference between HaH and hospital care, with an odds ferent medical conditions. Patients were classified into
ratio of 1.11 (95% CI: 0.75, 1.65) and minimal heteroge- five groups: acute medical conditions [5, 14–17], chronic
neity (I2 = 0%). Non-RCTs, however, indicate a significant medical conditions [4, 18, 19], orthopedic conditions [18,
mortality reduction in the HaH group (OR = 0.43; 95% 20–23], cancer [13], and medically complex conditions
CI: 0.26, 0.70), albeit with slightly more heterogeneity [18, 24, 25].
(I2 = 4%) (Fig. 4).
Mortality
Readmission Figure 8 reported the impact of early discharge HaH pro-
The forest plot comparing early discharge HaH to routine grams on mortality across various medical conditions.
hospital care shows no significant difference in readmis- Cumulatively, the data indicates a significant 25% reduc-
sion rates across both RCTs with an odds ratio of 0.97 tion in mortality risk for HaH participants (RR = 0.75,
(95% CI: 0.51, 1.82) and non-RCTs with an OR of 0.75 P = 1.00). Upon dissection by condition, both acute and
(95% CI: 0.35, 1.56), and this pattern holds in the com- chronic medical cases, as well as orthopedic and complex
bined analysis (OR 0.88; 95% CI: 0.55, 1.40). However, the medical conditions, showed favorable trends towards
high heterogeneity observed within both RCTs (I2 = 71%) HaH programs, even if individual disease categorizations
Lin et al. BMC Medicine (2024) 22:250 Page 6 of 15

Fig. 2 Risk of bias plots for RCT studies

Lin et al. BMC Medicine (2024) 22:250 Page 7 of 15

Fig. 3 Risk of bias plots for non-RCT studies

Lin et al. BMC Medicine (2024) 22:250 Page 8 of 15

Fig. 4 Early discharge hospital at home versus routine hospital care: mortality

Fig. 5 Early discharge hospital at home versus routine hospital care: readmission
Lin et al. BMC Medicine (2024) 22:250 Page 9 of 15

Fig. 6 Early discharge hospital at home versus routine hospital care: length of treatment

Fig. 7 Early discharge hospital at home versus routine hospital care: functional status — Barthel index at the end up

did not always attain statistical significance. Notably, the while reflecting a slight reduction in readmission risk,
relatively low heterogeneity among acute and chronic also lacked statistical significance. Intriguingly, ortho-
conditions (I2 = 20% and I2 = 0%, respectively) suggests pedic cases displayed a notably higher readmission risk
consistent outcomes across these trials. The forest plot (RR = 3.14), though this observation was based on lim-
further reinforces these findings, with the majority of ited data and requires further substantiation. Medical
individual study outcomes leaning towards the early dis- complex conditions and cancer demonstrated neutral
charge HaH advantage and the aggregate result repre- to increased risk, yet again without achieving statisti-
sented by the diamond, solidly positioned on the “early cal significance. An overarching observation is the pro-
discharge HaH” side, underscoring the potential benefits nounced heterogeneity in some disease categories, such
of HaH in reducing mortality. as chronic medical problems (I2 = 68%), underscoring the
importance of interpreting these results with caution.
Readmission The forest plot bolsters these conclusions, with a mix of
Figure 9 illuminated the impact of early discharge HaH study outcomes on both sides of the neutrality line and
programs on readmission rates across diverse medical the collective result, embodied by the diamond, tending
conditions. Overall, the combined data reveals a marginal slightly towards the “routine hospital” side. This suggests
increase in readmission risk for HaH programs, but the that while HaH may offer several advantages, it is crucial
effect is statistically insignificant (RR = 1.04, P = 0.26). to consider patient-specific factors and disease categories
When segregating by disease type, acute medical cases when assessing its implications for readmission.
depict an almost neutral effect, with a pooled risk ratio
near unity (RR = 1.00). Chronic medical problems,
Lin et al. BMC Medicine (2024) 22:250 Page 10 of 15

Fig. 8 Early discharge hospital at home versus routine hospital care: subgroup analysis based on disease classification — mortality

Length of treatment a negligible reduction in treatment duration with HaH

Figure 10 elucidates the impact of early discharge HaH (MDs =  − 0.11), and the encompassing forest plot sub-
programs on the duration of treatment, quantified as the stantiates this with outcomes straddling both sides of
standard mean difference across various medical condi- the neutrality line. The prevailing observation is that
tions. For acute medical problems, the pooled data indi- while HaH may potentially reduce the duration of treat-
cates a modest reduction in treatment duration for HaH, ment, the extent of this reduction varies considerably
albeit with a wide confidence interval that borders on across studies and disease categories. Given the high het-
the null effect (MDs =  − 0.21, P = 0.21). Notably, there erogeneity in results, careful interpretation and further
is high heterogeneity (I2 = 87%), suggesting substantial research are necessary to delineate the specific scenarios
variability among the included studies for this category. where HaH offers significant time-saving benefits.
For orthopedic issues, although a reduction in treatment
duration is observed for HaH, the confidence interval Barthel index at the end up
overlaps with zero, denoting potential insignificance in Figure 11 conducted a subgroup analysis of the Barthel
the effect (MDs =  − 0.09). The heterogeneity in this sub- index scores at the endpoint, comparing early discharge
group is also high (I2 = 93%), emphasizing the variability HaH to traditional routine hospital care, segmented by
of study outcomes. The overall combined effect suggests
Lin et al. BMC Medicine (2024) 22:250 Page 11 of 15

Fig. 9 Early discharge hospital at home versus routine hospital care: subgroup analysis based on disease classification — readmission

acute medical problems and orthopedic conditions. For a reduction in Barthel index scores, the magnitude and
those with acute medical problems, HaH intervention robustness of this impact appear to be condition-specific.
suggests a significant reduction in Barthel index scores Collectively, these findings emphasize the potential ben-
with a standard mean difference of − 0.42 (95% CI: − 0.57 efits of HaH, particularly for patients with acute medical
to − 0.27), showcasing its potential efficacy. This result is problems, but warrant cautious interpretation in the con-
statistically significant with a p-value of less than 0.00001. text of orthopedic issues.
However, in orthopedic conditions, the results are less
definitive. Although there is a trend towards a decrease Discussion
in the Barthel index scores within the HaH group Fifteen trials were incorporated into this systematic
(MDs =  − 0.11), the confidence interval slightly spans the review examining early discharge HaH for older people.
line of neutrality, indicating potential non-significance. Our investigation encompassed a comprehensive explo-
This subgroup also displays moderate heterogeneity ration of the efficacy of early hospital discharge to home
(I2 = 69%), hinting at diverse outcomes across the incor- care across several key domains: mortality, readmis-
porated studies. In totality, while HaH seems to influence sion rates, duration of treatment, functional status (as
Lin et al. BMC Medicine (2024) 22:250 Page 12 of 15

Fig. 10 Early discharge hospital at home versus routine hospital care: subgroup analysis based on disease classification — length of treatment

Fig. 11 Early discharge hospital at home versus routine hospital care: subgroup analysis based on disease classification — Barthel index at the end
up

quantified by the Barthel index), and cognitive health (as forest map, the diamonds tended to be on the “early dis-
assessed by the MMSE). charge HaH” side, emphasizing the role of early discharge
Our analysis suggests that early discharge HaH cor- HaH in mortality, readmission rate, length of treatment
relates with a decrease in mortality rates, a conclusion and Barthel index. Our subgroup analysis based on dis-
drawn from 6 of the non-RCTs. In contrast, the impact ease classifications—encompassing acute medical prob-
of early discharge HaH on readmission rates appears lems, chronic conditions, orthopedic issues, cancer, and
negligible, a conclusion supported by four non-RCTs medically complex conditions—revealed similar trends
and five RCTs. The influence of early discharge HaH when compared with the main analysis.
on cognitive health, based on MMSE scores, is simi- In our analysis, the effectiveness for the same meas-
larly inconsequential, as demonstrated by one RCT and ures—such as mortality, length of treatment, and the Bar-
one non-RCT. Moreover, early discharge HaH might thel index—between RCTs and non-RCTs is not always
be linked to a decline in functional status, as gauged by identical. This variance likely stems from the inherent
the Barthel index, based on evidence from one RCT and differences in the design and execution of these two
three non-RCTs. After analyzing the data of the fifteen types of studies. RCTs, with their higher design rigor and
trials included, regardless of whether there was a statis- controlled environments, minimize selection biases and
tically significant difference between the early discharge provide stronger evidence of causality, while non-RCTs,
HaH group and the routine hospital care group from the reflecting broader and more diverse populations, may
Lin et al. BMC Medicine (2024) 22:250 Page 13 of 15

offer findings that are more generalizable but subject to to increase with population aging, urbanization, and
inherent biases and variability. The observed discrepancy globalization of risk factors [32, 33]. Implementing HaH
of effectiveness for the same measures between RCTs in these settings may alleviate some of these challenges
HaH and non-RCTs HaH highlights the importance of by providing essential healthcare services directly to
considering the real-world scenario in understanding the homes of older adults. This is particularly beneficial
reported intervention effectiveness. in geographical locations where distance and limited
The HaH model is distinguished by their more holistic transportation options make it arduous for individuals
approach compared to traditional hospital care, particu- to access healthcare facilities. HaH leverages telehealth
larly for older adults. Its capacity to prevent the devel- technologies and portable medical equipment, making
opment of geriatric syndromes is achieved through the it a scalable solution that can extend the reach of quality
comprehensive method of care that addresses not only healthcare to remote and underserved communities [34,
the medical but also the physical, emotional, and social 35]. While the effectiveness of HaH on specific clinical
needs of patients, thereby enhancing their overall well- outcomes may remain under debate, these broader socio-
being and reducing mortality risks. medical benefits highlight the importance of its contin-
Furthermore, the decline in functional status observed ued development and implementation. A comprehensive
in some patients underscores the indispensability of assessment of HaH programs should thus go beyond
incorporating a multidisciplinary team into the HaH traditional clinical metrics to include factors like patient
model. Specifically, the integration of physiotherapists safety, accessibility, and equity, especially in the changing
and occupational therapists is crucial. These profession- landscape of global healthcare [33].
als can tailor rehabilitation and daily activity programs Meanwhile, in this study, our analysis of functional
to the individual needs of each patient, fostering not status is limited to the Barthel index after the end of the
only the maintenance but potentially the improvement trials, while ignoring the trend of changes in the Barthel
of functional status. Such an approach ensures that care index before and after the intervention treatment in both
is not merely reactive but preventive and rehabilitative, groups. Due to the different evaluation indicators used
addressing both the immediate and long-term health in different trials for the various states of patients before
needs of the older adults. and after hospitalization, the emphasis on measurement
While our meta-analysis indicated that the efficacy of indicators also varies. Some trials focus on measuring
Hospital-at-Home (HaH) programs is equivocal with the index changes before and after admission, while oth-
regard to readmission rates and certain cognitive and ers focus on measuring the Barthel index after discharge.
functional outcomes, the approach nevertheless offers These may reduce the accuracy of our assessment of the
several distinct advantages that warrant further explora- improvement in functional status between the early dis-
tion and development. One of the most salient advan- charge group and the routine hospital care group. More-
tages of HaH programs in the aftermath of the COVID-19 over, the overlap of evaluation indicators for cognitive
pandemic is the minimization of exposure to nosocomial health in different experiments is also relatively low. If
infections for vulnerable older adults [26–28]. Traditional different indicators can be standardized, the evaluation in
hospital settings, by their nature, expose patients to a the mental state dimension will be more comprehensive.
variety of pathogens, posing an increased risk of acquir-
ing secondary infections. This is especially critical for
older adults who are often immunocompromised and Conclusions
thus more susceptible to infections. Implementing HaH In summation, while early discharge HaH presents a
programs can thus provide a safer, more controlled envi- viable approach for diminishing mortality and treat-
ronment, reducing the likelihood of cross-infections. The ment duration, its efficacy remains ambiguous in relation
healthcare industry’s experience in the pandemic under- to readmission, as well as cognitive and functional out-
scores the urgency for such decentralized healthcare comes. Thus, a prudent interpretation of these findings is
models that can offer quality care while mitigating risks essential.
associated with hospital-based treatment [29–31]. Abbreviations
HaH programs could also play a pivotal role in bridg- CI Confidence intervals
COPD Chronic obstructive pulmonary disease
ing healthcare disparities observed in rural areas and ED HaH Early discharge hospital at home
developing countries, which frequently face infrastruc- EQ-5D The 5-level EQ-5D version
tural constraints and shortages of healthcare profession- HaH Hospital at Home
MD Mean differences
als. The burden of chronic diseases such as heart disease, MeSH Medical Subject Headings
cancer, diabetes, and mental disorders is high in low- MMSE The Mini-Mental State Examination
income and middle-income countries and is expected Non-RCTs Non-randomized controlled trials
Lin et al. BMC Medicine (2024) 22:250 Page 14 of 15

OR Odds ratio Received: 3 November 2023 Accepted: 3 June 2024

PRISMA  Preferred Reporting of Items for Systematic Reviews and
Meta-Analyses
PROSPERO International prospective register of systematic reviews
RCTs Randomized controlled trials
RoB2 Version 2 of the Cochrane risk-of-bias tool for randomized trials References
ROBINS-I Risk Of Bias In Non-randomised Studies—of Interventions 1. Shepperd S, Iliffe S, Doll HA, Clarke MJ, Kalra L, Wilson AD, et al.
Robvis Risk of Bias VISualization Admission avoidance hospital at home. Cochrane Database Syst Rev.
SF-36 The 36-Item Short Form Health Survey 2016(9):CD007491.
2. World Health Organization. Decade of healthy ageing: baseline report.
2020.
Supplementary Information 3. Caplan GA, Sulaiman NS, Mangin DA, Aimonino Ricauda N, Wilson
The online version contains supplementary material available at https://​doi.​ AD, Barclay L. A meta-analysis of “hospital in the home.” Med J Aust.
org/​10.​1186/​s12916-​024-​03463-3. 2012;197:512.
4. Hernandez C, Casas A, Escarrabill J, Alonso J, Puig-Junoy J, Farrero E, et al.
Additional file 1: Table S1. PRISMA check list. Table S2. Characteristics of Home hospitalisation of exacerbated chronic obstructive pulmonary
included studies. Table S3. Meta analysis on early discharge HaH program disease patients. Eur Respir J. 2003;21:58–67.
in the past 20 years. Text S1. Search Strategy Used In PUBMED. 5. Leff B, Burton L, Mader SL, Naughton B, Burl J, Inouye SK, et al. Hospital
at home: feasibility and outcomes of a program to provide hospital-level
care at home for acutely Ill older patients background: acutely ill older
Acknowledgements persons often experience adverse. 2005.
We would like to thank the reviewers for their helpful comments, which 6. Suwanwela NC, Phanthumchinda K, Limtongkul S, Suvanprakorn P, Thai
greatly improved this article. T, Cross R, et al. Comparison of short (3-day) hospitalization followed by
home care treatment and conventional (10-day) hospitalization for acute
Authors’ contributions ischemic stroke. 2002.
LLL, CMY, CWB, and ZZX conceptualized the study. LLL, CMY, and CXW 7. Ward D, Drahota A, Gal D, Severs M, Dean TP. Care home versus hospital
designed the methodology and did the data collection. LLL did the data and own home environments for rehabilitation of older people.
analysis. TWM, CWB, and ZZX supervised the data collection, analyses, and/or Cochrane Database Syst Rev. 2008(4):CD003164.
interpretation. GYW and XX verified the data collection and evidence interpre- 8. Qaddoura A, Yazdan-Ashoori P, Kabali C, Thabane L, Haynes RB, Connolly
tation. ZZX acquired funding. LLL prepared the original draft of the manuscript SJ, et al. Efficacy of hospital at home in patients with heart failure: a
with the input of TWM, CWB, GYW, and ZZX. All authors edited and reviewed systematic review and meta-analysis. PLoS One. 2015;10:e0129282.
the final manuscript. All authors have read and agreed to the final version of 9. Gonçalves-Bradley DC, Iliffe S, Doll HA, Broad J, Gladman J, Langhorne
the manuscript and to the decision to submit. All authors had access to all the P, et al. Early discharge hospital at home. Cochrane Database Syst Rev.
data. LLL and ZZX have verified the data. All authors read and approved the 2017;(6):CD000356.
final manuscript. 10. Sterne JAC, Savović J, Page MJ, Elbers RG, Blencowe NS, Boutron I, et al.
RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ.
Funding 2019;366:l4898.
This work was supported in part by the Basic and Applied Research Funding of 11. Sterne JA, Hernán MA, Reeves BC, Savović J, Berkman ND, Viswanathan M,
the Guangzhou Science and Technology Plan (2024A04J4621), in part by the et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies
Sun Yat-sen University Basic Start-up Funding (51000–12230014), and in part of interventions. BMJ. 2016;355:i4919.
by the Guangzhou Science and Technology Programme (2024A03J074, and 12. Sterne JA, Hernán MA, Reeves BC, Savović J, Berkman ND, Viswanathan M,
2023A03J0286). et al. ROBINS-I: A tool for assessing risk of bias in non-randomised studies
of interventions. Res Synth Methods. 2021;12:55–61.
Availability of data and materials 13. Mooney K, Titchener K, Haaland B, Coombs LA, O’Neil B, Nelson R, et al.
The data used to support the findings of this study are included within the Evaluation of oncology hospital at home: unplanned health care utiliza-
article. tion and costs in the huntsman at home real-world trial. J Clin Oncol.
2021;39:2586.
14. Caplan GA, Coconis J, Board N, Sayers A, Woods J. Does home treatment
Declarations affect delirium? A randomised controlled trial of rehabilitation of elderly
and care at home or usual treatment (The REACH-OUT trial). Age Ageing.
Ethics approval and consent to participate 2006;35:53–60.
Not applicable. 15. Harris R, Ashton T, Broad J, Connolly G, Richmond D. The effectiveness,
acceptability and costs of a hospital-at-home service compared with
Consent for publication acute hospital care: a randomized controlled trial. 2005.
All of the authors agreed to submit and publish the final manuscript. 16. Mas M, Inzitari M, Sabaté S, Santaeugènia SJ, Miralles R. Hospital-at-home
Integrated Care Programme for the management of disabling health
Competing interests crises in older patients: Comparison with bed-based Intermediate Care.
The authors declare that they have no competing interests. Age Ageing. 2017;46:925–31.
17. Mas M, Santaeugènia SJ, Tarazona-Santabalbina FJ, Gámez S, Inzitari M.
Author details Effectiveness of a Hospital-at-Home Integrated Care Program as Alterna-
1
School of Public Health, Sun Yat-Sen University, Guangzhou, China. 2 Insti- tive Resource for Medical Crises Care in Older Adults With Complex
tute for Healthcare Artificial Intelligence Application, Guangdong Second Chronic Conditions. J Am Med Dir Assoc. 2018;19:860–3.
Provincial General Hospital, Guangzhou, China. 3 Bouvé College of Health 18. Shepperd S, Harwood D, Jenkinson C, Gray A, Vessey M, Morgan P. Ran-
Sciences, Northeastern University, Boston, MA, USA. 4 Institute of Global Health domised controlled trial comparing hospital at home care with inpatient
and Infectious Diseases, University of North Carolina at ChapelHill, Chapel Hill, hospital care. I: three month follow up of health outcomes. Br Med J.
NC, USA. 5 School of Data Science, City University of Hong Kong, Hong Kong 1998;316:1786–91.
SAR, China. 6 Faculty of Health Sciences, City University of Macau, Macao SAR, 19. Villani A, Malfatto G, Compare A, Rosa FD, Bellardita L, Branzi G, et al.
China. Clinical and psychological telemonitoring and telecare of high risk heart
failure patients. J Telemed Telecare. 2014;20:468–75.
20. Richards SH, Coast J, Gunnell DJ, Peters TJ, Pounsford J, Darlow MA.
Randomised controlled trial comparing effectiveness and acceptability of
Lin et al. BMC Medicine (2024) 22:250 Page 15 of 15

an early discharge, hospital at home scheme with acute hospital care. Br

Med J. 1998;316:1796–801.
21. Closa C, Mas M, Santaeugènia SJ, Inzitari M, Ribera A, Gallofré M. Hospital-
at-home Integrated Care Program for Older Patients With Orthopedic
Processes: An Efficient Alternative to Usual Hospital-Based Care. J Am
Med Dir Assoc. 2017;18:780–4.
22. Levi Y, Punchik B, Zikrin E, Shacham D, Katz D, Makulin E, et al. Intensive
inpatient vs. home-based rehabilitation after hip fracture in the elderly
population. Front Med (Lausanne). 2020;7:592693.
23. Mas MÀ, Closa C, Santaeugènia SJ, Inzitari M, Ribera A, Gallofré M.
Hospital-at-home integrated care programme for older patients with
orthopaedic conditions: Early community reintegration maximising
physical function. Maturitas. 2016;88:65–9.
24. Cai S, Intrator O, Chan C, Buxbaum L, Haggerty MA, Phibbs CS, et al.
Association of Costs and Days at Home with Transfer Hospital in Home.
JAMA Netw Open. 2021. https://​doi.​org/​10.​1001/​jaman​etwor​kopen.​2021.​
14920.
25. Donald IP, Baldwin RN, Bannerjee M. Gloucester hospital-at-home: a
randomized controlled trial. Age Ageing. 1995;24:434.
26. Moazzami B, Razavi-Khorasani N, Dooghaie Moghadam A, Farokhi
E, Rezaei N. COVID-19 and telemedicine: immediate action required
for maintaining healthcare providers well-being. J Clin Virol.
2020;126:104345.
27. Ohannessian R, Duong TA, Odone A. Global telemedicine implemen-
tation and integration within health systems to fight the COVID-19
pandemic: a call to action. JMIR Public Health Surveill. 2020;6:e18810.
28. Calton B, Abedini N, Fratkin M. Telemedicine in the Time of Coronavirus. J
Pain Symptom Manage. 2020;60:e12–4.
29. Hollander JE, Carr BG. Virtually perfect? Telemedicine for Covid-19. N Engl
J Med. 2020;382:1679.
30. Nundy S, Patel KK. Hospital-at-home to support COVID-19 surge - time to
bring down the walls? JAMA Health Forum. 2020;1:e200504.
31. Latifi R, Doarn CR. Perspective on COVID-19: finally, telemedicine at
center stage. Telemed E Health. 2020;26:1106.
32. Beaglehole R, Epping-Jordan J, Patel V, Chopra M, Ebrahim S, Kidd M,
et al. Series Alma-Ata: Rebirth and revision 3 improving the prevention
and management of chronic disease in low-income and middle-income
countries: a priority for primary health care. 2008.
33. Hamine S, Gerth-Guyette E, Faulx D, Green BB, Ginsburg AS. Impact of
mHealth chronic disease management on treatment adherence and
patient outcomes: A systematic review. J Med Internet Res. 2015;17:e52.
34. Mueller DB, Tivey D, Croce D. Health-technology assessment: Its role in
strengthening health systems in developing countries. Strengthening.
2017. https://​doi.​org/​10.​7196/​SHS.​2017.​v2.​i1.​50.
35. Leonard E, de Kock I, Bam W. Barriers and facilitators to implementing
evidence-based health innovations in low- and middle-income coun-
tries: a systematic literature review. Eval Program Plann. 2020;82:101832.

Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in pub-
lished maps and institutional affiliations.