Hospital-in-the-Home as a Model for Mental Health Care Delivery: A Narrative Review

The care of people with mental illness has undergone a revolution in the past 50 years, transitioning from institutional to community-based models, with a greater focus on supporting consumers to be actively involved in their care, in line with a recovery-oriented model (1, 2). Along with this, there has been a move to reimagine the role of inpatient services from that of long-term support to brief interventions in situations of acute deterioration or increased risk (3). Deinstitutionalization has now spread globally, spawning a diverse range of community-based mental health services to meet a variety of needs (2, 4, 5). These include crisis assessment and treatment teams, which provide mobile assessment and triage focusing on containment and referral to appropriate services (either hospital admission or less acute follow-up) (2); continuing care teams or community mental health teams, which provide case management and specialist therapies on an outpatient basis with the aim of transitioning care back to a primary care provider (4); assertive outreach teams, which provide long-term intensive care to those with complex needs requiring support to remain living in the community (4); and early intervention teams, which provide early identification and treatment for those at high risk for certain conditions (particularly psychosis) (5).

Beyond these services, specialist therapy clinics, day hospitals, and residential care facilities all provide additional community treatment options to consumers (4, 5). Although the advent of these community treatment options has led to many benefits, inpatient beds continue to provide certain elements of necessary care, particularly restrictive practice during periods of heightened risk. However, the use of acute inpatient beds is costly, and achieving care models that align with the varying individual needs of consumers is a challenge. In addition, with population growth, increasing structural adversity and inequity, and a rise in the need for mental health treatment, many health care systems have seen demand for these beds reach critical levels (2, 6–8). A sustainable approach to mental health care is needed that takes into account escalating costs associated with hospital admission along with principles such as recovery-oriented, trauma-informed, and least restrictive care (1, 2).

Advances in modern medicine and community health care capabilities (including remote monitoring) have seen other areas of medicine explore models of bed substitution such as hospital-in-the-home (HITH) (9, 10). Current HITH models either are limited to certain specialties (e.g., respiratory diseases, stroke, geriatrics, and rehabilitation) or are set up to cover a broad range of potential hospital presentations (9, 11). HITH has significant advantages both for consumers (increased satisfaction and reduced risks) and for health care systems (lower costs and greater efficiency) (9, 11–13). Although mental health HITH models have been trialed to various degrees (14–18), there has not been widespread adoption except in the United Kingdom, where the closely related Crisis Resolution and Home Treatment (CRHT) model is in use (18). As in the HITH model, CRHT teams provide short-term, home-based care at an intensity similar to that available on inpatient admission. They focus on assessment and crisis intervention (working with the consumer to help adapt to challenging circumstances), whereas mental health HITH models may take a broader approach to substituting for inpatient care (19).

For mental health HITH services to be an effective alternative to inpatient admission, it is vital that their planning is based on a solid foundation of evidence. To support the translation of mental health hospital care into HITH service provision, we aimed to undertake a comprehensive review of home-based mental health services that augment or substitute inpatient treatment. This review assesses the design, implementation, and outcomes of these services, which will provide direction for the establishment of future mental health HITH programs.

Methods

We conducted this review from May to August 2020, using a structured search strategy to identify all relevant articles for inclusion. A narrative review methodology was used because of the heterogeneous nature of the data. Synthesizing data in this manner was helpful for extracting and exploring key themes and ideas relating to the use of mental health HITH programs, especially where high-quality studies of these services were limited.

Study Eligibility

To be eligible for inclusion, studies had to have an HITH-type program as their primary intervention. This was defined as any program intended to provide intensive home care to consumers who would otherwise need hospital admission. Eligible studies included programs with 24-hour, 7-days-a-week capacity; multiple home visits per week; brief episodes of care (similar to inpatient lengths of stay); and multidisciplinary care with psychiatrist oversight. Individuals who were admitted to these services had to have an acute episode of mental illness, and the main treatment provided had to be related to mental illness (as opposed to assistance with personal care or complex general medical needs). As the terminology of mental health HITH programs is still poorly specified, study eligibility was based on HITH characteristics as determined by two independent reviewers. All study designs were eligible for inclusion. Gray literature (such as government reports and editorials) was also included.

Search Strategy

Electronic database searches were conducted in MEDLINE, PsycINFO, CINAHL, Embase, Scopus, Web of Science, and Google Scholar. We conducted an additional gray literature search in Google. We used a combination of Medical Subject Headings—or MeSH—terms and keywords related to the concepts of “HITH” AND “mental health” OR “psychiatry.” (Full details are available in an online supplement to this review.) A total of 2,277 results were obtained. Two independent researchers (M.T., W.Y.) conducted screening and review for eligibility of all articles and reports identified (see the flow diagram in the online supplement). A manual search of reference lists of eligible studies was also conducted. As a result of the initial and manual searches, 56 studies were identified.

Data Synthesis

The studies identified included both quantitative and qualitative data covering a wide range of outcome measures, including consumer satisfaction, quality of life, health and social function, and consumer feedback. The most common outcome measure was the effect on hospital admission rates and days in the hospital. Significant heterogeneity across study types was noted in both methodology and outcome measures. As such, a narrative approach was used to extract data relating to the themes in the research question of program design, implementation, and outcome. On the basis of common trends in the literature, material was further synthesized into the headings listed below. The data were also used to compile a proposed mental health HITH model.

Results

The literature search yielded literature from countries including the United Kingdom (19–40), Germany (17, 41–43), Taiwan (44, 45), Australia (1, 9, 14, 15, 46–48), and Spain (16, 49, 50). Types of literature included were government reports and guidelines (38), commentaries and editorials (25, 47), and qualitative (16, 22, 23, 28, 33, 34, 37, 44, 45, 51, 52) and quantitative (31, 42, 53) research. A full list of included papers is included in Table 1 (49, 50, 54–67).

Definition of Existing Programs

No consensus exists regarding the terminology of HITH-type mental health services. Furthermore, there are variations in the purpose and types of existing programs. As such, a broad range of home-based treatment programs were found to be eligible for inclusion.

CRHT teams are a relatively well-researched HITH-type mental health service providing an alternative to inpatient care in the United Kingdom (19). Research on CRHT teams has been included, as these teams are broadly comparable to HITH-type programs in that they substitute for inpatient mental health care. CRHT arises from several theoretical frameworks, including crisis intervention theory, which stipulates that crises are normal periods of transition within which clinicians can work to promote adaptive skills and coping (19). Other home-based treatment models included in this review are linked more directly to a desire to alleviate psychiatric presentations to emergency departments (EDs) and subsequent pressure on inpatient beds, with a recognition that inpatient admission can be unhelpful in certain circumstances (14–16). Despite this, both program types operate similarly in terms of their key characteristics and functions (as discussed below). Both can be distinguished from other home-based treatment models such as assertive outreach, which aims to improve engagement and care over the long term (instead of replicating inpatient care) (17, 38).

Design and Implementation

Service components and functions.

Despite the variation in the types of mental health HITH programs worldwide, several key service components were noted in the literature, including a multidisciplinary team with clinical oversight provided by psychiatrists (19, 25, 36, 38), low consumer-to-staff ratios (54), and a 24-hour service with the capacity for multiple home visits per day (15). The main functions of the identified services include comprehensive assessments, identifying and addressing crisis triggers, psychoeducation, symptom management, medication prescribing and supervision, practical support, and supporting transition back to other services (41). Consumers who were treated within HITH-type services are those who would have otherwise required hospital admission (19, 51). Although challenging in a multielement, 24-hour rostered service, continuity of care and maintenance of the therapeutic alliance remain priorities in the HITH model. Also of importance when working in the home setting is the involvement of family and caregivers, as well as the development of a mutually agreed-upon treatment plan (43). These commonly identified elements are also represented in a proposed HITH model (Figure 1).

Despite these common characteristics (14, 15, 19, 25, 36, 51), research evaluating many of these components is still lacking. In a systematic review of CRHT implementation, Wheeler et al. (36) found that the presence of psychiatrists in CRHT teams and longer hours of operation were associated with better hospital admission outcomes compared with CRHT teams that lacked these characteristics. There is a dearth of quantitative research on other key components, and little is known about what effect individual components of HITH services might have on outcomes (36).

Suitable consumers.

Consumers identified as suitable for HITH services include those who are accepting of care, adhering to treatment, and having appropriate social supports (14). There is evidence that a wide variety of disorders can be managed in the home environment, including acute psychosis (26, 55), although mania, hypomania, or polysubstance misuse may be less manageable because they cause rapidly fluctuating mental state and risks for consumers who present with these disorders (14, 26). Those with complex personality disorders treated at home may be at risk for dependence or maladaptive coping, although this risk could be offset by clear, predefined discussions exploring expectations of length of admission and frequency of contacts (26). Other consumers with less suitable conditions include those with high levels of suicidal ideation or risk, as well as those who poorly adhere to care (49).

Staffing and practical considerations.

Providing mental health care at home presents unique challenges. As such, a clear need exists for experienced team members (56) and adequate supervision (52, 57). Regular handovers between staff are also crucial to maintain care consistency (48). Specific staff training that may be required includes safety training and education on medication use and storage (38).

Staff recruited for HITH-type services were generally based at existing hospitals or day facilities (14, 20, 40, 42, 47, 58). A centralized management and staff base (termed “HITH base” in Figure 1) has been identified as a crucial component for effective care coordination (20), with the likelihood for loss of continuity of care in programs that lack this element (15). Operating out of an existing HITH infrastructure may increase the visibility of mental health services to the general hospital, although in one Australian service, staff tended to work most closely with the emergency and mental health departments to facilitate care (14). In England, despite the option for appointments at the day clinic or at home, a significant proportion of contacts were still in the consumer’s home (45%) (40). Overall, most services were delivered in metropolitan areas (14–16, 47, 48, 51, 54–56, 58), with a minority in more rural areas (17, 59).

One key function of the CRHT model in particular is its role as a “gatekeeper,” assessing and referring mental health care consumers to appropriate services, including inpatient care or home treatment (19, 25, 36, 38). However, as indicated by data on the actual implementation of CRHT teams, only about one in three CRHT programs perform this role (37). Outside of the United Kingdom, the implementation of a gatekeeper role has also been variable. One Australian program received referrals to home treatment directly from acute inpatient wards or EDs (15). Other services had consumers referred to them by mobile assessment teams (48). In Switzerland, this role was performed by a central triage unit (42).

Implementation barriers and obstacles.

Several challenges to HITH service implementation have been identified in the literature. Concerns regarding the appropriate management of risk for harm to self or others in community settings have been a barrier, with home care considered a higher risk than hospital care (25). Early psychiatrist involvement in the design and implementation of these programs is seen as a way to support their use and promote referrals. Early involvement of middle management staff who manage day-to-day operations may help to promote smooth initiation of these programs (55). Another barrier in terms of staff attitudes was identified in an Australian study by Singh et al. (15), who found that staff typically viewed their HITH service as a “transfer of care” or “early discharge” service rather than a genuine alternative to inpatient treatment; this view shifted only after extensive education and promotional work.

High start-up costs may be a disincentive, leading to studies into the cost-effectiveness of HITH-type programs (discussed below) (31). In light of this, directing funding to existing services may be preferable, particularly by way of increasing the crisis capacity of existing community teams or opening more beds (27, 31). The former may distract from the ability of community teams to care for chronically ill patients (27), whereas the latter is not necessarily sustainable for reasons already discussed (1, 46).

Specific challenges in implementing the CRHT model were discussed in an article by Onyett et al. (37). Fewer than half of the CRHT teams surveyed described themselves as “fully established,” with many reporting “high assessment loads, understaffing, and limited multidisciplinary input,” leading to fewer clients seen than anticipated. Solutions included having clearer referral protocols, increased staffing, and after-hours resources, particularly in terms of medical and psychiatrist cover (37). Three studies mentioned high levels of staff enthusiasm and job satisfaction, although robust data were lacking (25, 48, 51). A hypothesized risk for staff burnout requires further research (18, 25).

Outcomes

Hospital admissions.

Eighteen studies provided data on the effect of HITH on hospital admission outcomes (9, 18, 20, 24, 26, 29, 35, 36, 39, 42, 49, 51, 58–63). Most of the evidence suggests that HITH effectively substitutes for inpatient admission. However, the quality and reporting of this evidence vary significantly. In a 2012 meta-analysis of HITH programs across all specialties, the combined effect on readmission rates of four mental health programs did not reach statistical significance (odds ratio [OR]=0.29; 95% confidence interval [CI]=0.05–1.65) (9). However, three of the four studies had interventions that were not considered intensive enough to be included in this review. Two of the studies examined a long-term assertive outreach model (68, 69), and the third exclusively examined a day hospital program (70). A 2015 prospective cohort study of 896 patients in Spain compared treatment at home for mental disorders with treatment in an ED. They found that those referred for home treatment were 4.59 times less likely to require hospital admission (95% CI=2.89–7.30). However, consumers were not randomly allocated, and this result was likely influenced by selection bias (49). A 2019 randomized controlled trial (RCT) with 707 patients in Germany found that those who received treatment at home had a 30% reduction in total hospital days (42).

Johnson et al. (61) conducted an RCT of the effect of CRHT teams on hospitalization outcomes in 2005. In that study, 260 participants were randomly assigned to CRHT or to standard care (usual inpatient and community services). After 8 weeks, the experimental group was statistically significantly less likely to require hospital admission (OR=0.19; 95% CI=0.11–0.32). Robin et al. (58) followed up CRHT patients over 5 years (the longest follow-up period of any study). Compared with standard inpatient treatment, the CRHT group had fewer hospital admissions and days in the hospital; however, this effect was lost after 2 years.

Three key studies assessing hospital admission rates at a population level were identified. In Ireland, Iqbal and colleagues (59) reported a 50% decrease in admission rates at a local hospital after the introduction of a home-based treatment team. An analysis of admission data across 229 health districts in England by Glover et al. (39) revealed a 23% decrease in admission rates after the introduction of CRHT teams with 24-hour capacity (compared with only 10% for CRHT teams with limited operating hours). However, after reanalysis and controlling for confounding factors, no significant differences were found in admission rates in districts with and without CRHT teams (35). The authors hypothesized that CRHT may reduce informal (voluntary) admissions, but this reduction may be offset by an increase in compulsory admissions (35). This finding is supported by the findings of the RCT by Johnson et al. (61), indicating that CRHT had no significant effect on compulsory admissions (despite reducing admissions overall).

However, HITH or CRHT programs may not be appropriate for all consumer populations. Logistic regression analysis of data from a 12-month study of a HITH-type program disclosed that high suicide risk at treatment initiation and a history of past hospital admissions increased the risk for future admissions (62). Additionally, first-episode psychosis (FEP) has been highlighted as an area in which home treatment may be of particular benefit, but results have been mixed so far (29, 63). A study by Gould et al. (29) identified relatively high admission rates for those with FEP treated through CRHT (45% over 3 months) compared with 28% for those treated via standard community services (although it is possible that these presentations were less acute).

Symptoms and psychosocial functioning.

Thus far, evidence for the impact of HITH on clinical outcomes has been mixed. Singh et al. (15) found that reductions in clinical and functional impairment (Health of the Nation Outcome Scale; HoNOS) scores postdischarge from HITH were comparable to those after inpatient treatment. In an observational trial of 60 consumers treated in an HITH-type intervention and 58 matched consumers assigned to inpatient admission, Kilian et al. (17) observed that only the HITH group had decreased clinical and functional impairments, although both groups had decreased psychotic and depressive symptoms. A 2015 systematic review of CRHT teams reported limited evidence for any positive impact on symptoms (36). Similarly, in a recent RCT of 707 participants diagnosed as having a wide range of mental disorders, Stulz et al. (42) found no difference between inpatient care and home treatment in terms of reductions in HoNOS and symptom scores (on the Brief Symptom Inventory). A 2005 RCT comparing CRHT with standard inpatient and outpatient care by Johnson et al. (61) of 260 participants reported a significant reduction in symptom severity in the experimental group at 8 weeks but not at 6 months. These more rigorous studies tended to report fewer significant outcomes on clinical measures.

Consumer satisfaction.

Reported rates of consumer satisfaction, obtained through questionnaires and interviews across the HITH programs, were high (23, 28, 32–34, 50, 64). This may reflect general consumer preference for treatment at home rather than in the hospital. Consumers were frequently averse to hospital admission, and receiving treatment in one’s own environment was seen as less disruptive than receiving hospital treatment (15, 34). The ability to address crises and learn coping skills in an everyday setting was noted as an important benefit (43, 50, 64). Consumer satisfaction may also have been influenced by selection bias in studies with consumer choice regarding the treatment received (14).

Several other factors influencing consumer satisfaction have been examined. These include ease of access and higher availability of staff and support, consistent approaches by staff, feeling listened to and understood as a person, and having choice and active involvement in care (33, 50, 64). Hubbeling and Bertram (32) found that a perceived resolution of the specific stressor was most linked with consumer satisfaction and recovery. Khan and Pillay (23) studied the cultural factors that influence consumer preference for home treatment. These factors included stigma associated with inpatient treatment, hospitals not being tailored to meet complex cultural needs (e.g., language, faith, and dietary needs), and individuals feeling more comfortable in their own communities. Carpenter and Tracy (22) used semistructured interviews to explore consumer experiences of HITH and found that although feedback was generally positive, consumers reported staff arriving late or at inconvenient times as a cause of anxiety. There was a reported lack of consistency between clinicians and lack of peer support. A common theme identified in this study and others was the sense from consumers of being discharged too soon or abruptly and not being involved in this decision (22, 33, 36). As with an inpatient admission, early signposting and linkages to community supports were considered crucial for ensuring smooth transitions (33, 36).

Cost-effectiveness.

The cost-effectiveness of HITH programs for mental disorder treatment has not been robustly studied (17, 31, 40, 55, 65, 66). HITH-associated costs are generally higher than for other forms of community care. Analysis of RCT data by McCrone and colleagues (31) revealed that noninpatient costs of the CRHT group were US$988 higher per consumer than among those receiving standard care. However, in the same study, when inpatient costs for the two groups were factored in, CRHT was on average $3,137 less expensive per consumer, suggesting that costs are offset by reduction in days spent in the hospital (31). The finding that HITH or CRHT reduces costs when compared with inpatient admission was replicated by several other studies. Wasylenki et al. (55) reported savings of $371 per consumer per day for an HITH-type program treating psychosis relative to inpatient care. Kilian et al. (17) estimated savings of $8,388 per episode of home treatment, and Harrison et al. (40) found that a “home-option” program lowered costs by $1,844 per consumer per episode, compared with an inpatient sample. The same study found that the “day-hospital-only” group incurred the least costs. Although current evidence suggests that mental health HITH treatment is cost-effective, future research should factor in additional variables, such as the costs of establishing and maintaining these programs in preexisting services (18, 65). Regardless of the effects on costs, the ability for a home treatment program to reduce hospital use has been identified as a key factor in long-term sustainability (67).

Risks associated with home care.

Risks for self-harm or suicide associated with mental health HITH models have not been extensively researched (21, 41). One exception was a retrospective longitudinal analysis by Hunt et al. (30), conducted with National Health Service data from 2003 to 2011, comparing suicide rates for those receiving CRHT with suicide rates for patients receiving inpatient treatment. Average suicide rates were 14.6 per 10,000 episodes for those under crisis care, compared with 8.8 per 10,000 inpatient admissions and 7.8 per 10,000 people in contact with standard community mental health services. Of the consumers in the CRHT group who died by suicide, 49% had experienced a recent adverse life event, 44% lived alone, and one-third had been discharged from the hospital in the previous month. This observation suggests that CRHT use as an early discharge option carries some risks, notwithstanding selection biases (30). Across a smaller sample of 111 people treated in an HITH service over 1 year, there was only one reported instance of self-harm, with the authors noting that a change in risk status could be detected and acted upon promptly (15).

Effect on caregivers.

There is a dearth of research regarding the potential effects on caregivers who support consumers enrolled in HITH-type programs (18, 27, 41). Inpatient admissions are generally seen as an avenue for respite for caregivers, and this effect may be offset by home treatment (23, 41). Indeed, 8.1% of admissions to the hospital from home treatment in a study in the United Kingdom were due to caregivers struggling to cope with the demands of caregiving (62). Giménez-Díez et al. (50) found that patients tended to be more satisfied with home care than were caregivers, possibly reflecting a degree of caregiver burden. Also important to consider are the needs of vulnerable family members—including dependent children—who, in a study by Khalifeh et al. (28), generally had negative experiences of home treatment for their parents and expressed a preference for parental admission. Despite these findings, many programs did report high levels of caregiver satisfaction (14, 21, 47, 55). In particular, HITH service allows close collaboration with caregivers to provide psychoeducation and practical skills to help their loved ones in times of crisis (25, 51).

Discussion

Key components of mental health HITH programs are outlined in a working model depicted in Figure 1. Important themes relating to the design and implementation identified in this review included having clear referral criteria and pathways, adequate staffing, after-hours availability, and working collaboratively with consumers and caregivers. In working toward the standardization of terminology, a mental health HITH service can be defined as any service intended to provide mental health care, of a quality comparable to that of inpatient care, in the home as an alternative to hospital admission.

There is a need to move beyond the current focus on mental health HITH service as a means to reduce pressure on inpatient bed capacity and to consider HITH’s role in an integrated mental health system. A key theme identified in the literature is the function of mental health HITH services in promoting the recovery model of mental health care (51). These programs offer both consumers and caregivers the opportunity to work closely with clinicians in their everyday environment to develop skills in improving their mental health, particularly during times of crisis or exacerbation of symptoms (25, 43, 50, 51, 64). This also helps in equalizing the therapeutic relationship in order to promote supported decision making and active involvement of consumers and caregivers in treatment (50), which may be harder to achieve during inpatient care. Although services with goals similar to those of HITH already exist across mental health service structures, specific HITH teams may have a unique role in supporting these existing services when patients require more intensive care.

Although equalizing power dynamics can be beneficial, therapeutic and ethical boundaries may become ambiguous when providing care in the home setting. Challenges can arise in maintaining confidentiality, establishing boundaries, and preventing exposure of staff to unsafe situations (52). There is also a potential tension between promoting consumer autonomy and privacy versus differing consumer and clinician appraisals of need, increasing risks, and a perceived duty of care (56). The need for adequate supervision, and clear policies and procedures, in view of these ethical challenges cannot be understated (52, 57).

Notably, we found significant variances in and among countries in terms of implementation of mental health HITH programs. In particular, a relative lack of implementation in the United States was noted compared with that in countries such as the United Kingdom where the CRHT model is widely used (18, 19). Although beyond the remit of this review, further investigation could be targeted to understanding clinical services factors that influence the delivery of certain model types; demographic and environmental factors; and financial factors, including insurance and funding arrangements.

This review had two key limitations. First, researchers and clinicians use a wide range of terminology to describe HITH-type mental health programs. Despite using broad search terms, the initial search failed to yield a significant number of studies, whose number was increased only later in the manual searching of reference lists of the studies obtained in the initial search. Programs have also been established internationally with various governance structures, funding models, and legislative requirements. Therefore, the research included in this review was heterogeneous and not systematically comparable. Second, this review attempted to identify and analyze key themes and outcomes in HITH-type programs. However, because of this heterogeneity, this review could only broadly highlight common components that may be of benefit in establishing a mental health HITH service. Future use of standardized methods and outcome reporting in an evaluation of these programs may help provide a more nuanced view.

Conclusions

In this review, we have identified and discussed an established set of characteristics and functions of HITH-type mental health care models. Moderate evidence exists that these models reduce hospitalization rates and days, in keeping with the goal of HITH to provide an alternative to inpatient care. However, HITH cannot replace inpatient hospitalization for all care requirements, and there is a continued need for inpatient treatment among some consumers with acute mental illness symptoms and risks (15). Future research should aim to place HITH in the context of the wider mental health care system, particularly through direct comparisons with existing community-based services. There is also a need to move away from studies on hospital bed days and costs to further the research on the effects of HITH on symptoms, quality of life, and long-term illness outcomes for consumers, as well as the effects of HITH models on caregivers and staff.