While the constant increase in the population of APSDMHD is a major public health issue (14, 16, 21–24), the number of studies on the deinstitutionalization process toward this population (83 articles) seems very limited (8, 14, 25, 26). This paucity of data seems to be primarily related to the population concerned. Also, even though many studies focus on chronic mental health disorders in adults, there is a large body of literature indicating that there are very few clinical, cognitive, epidemiological, and therapeutic studies on APSDMHD (24, 26–31). This, even though this fringe of the population has been showing a steady increase since 1980 and is expected to continue to do so until 2040, particularly in westernized countries such as the United States (21, 32–34).

Concerning the available studies, many authors point out that the existing work does not allow a generalization of the results. Indeed, the published literature often presents certain methodological weaknesses (sample size too small or inclusion of patients with dementia) (27, 30, 31, 35). Similarly, few have proposed an assessment of the factors leading to longer hospital stays for APSDMHD-H (33, 36, 37). In this respect, the clinical evolution of APSDMHD is difficult to evaluate from a psychometric point of view because there are no tools yet available to study the changes in psychological symptoms during aging (16, 35, 37). Moreover, considering age-related predictive variables (emergence of dementia, somatic comorbidities, etc.) to minimize bias is extremely complex to achieve in the context of elderly psychiatry (11, 16, 38, 39). All these variables therefore make it difficult to conduct randomized clinical trials (RCTs) and indeed to obtain generalizable data (11, 30, 38, 39).

At the same time, the issues surrounding APSDMHD seem to be under-explored. For example, WHO consensus statements (2) or the publication of the European commission (40) addressed the issue of cognitive impairment in the elderly, depressive symptoms, or dementia, but not the aging of PSDMHD (11). Moreover, more recent WHO publications (3) still do not seem to have explored this issue further. In addition to their scarcity, work on the existing provision of care for these users seems to focus essentially on the medical approach (14, 26, 41–43). They also neglect the social and medico-social dimensions, which are nonetheless important in supporting APSDMHD, particularly when implementing a policy of deinstitutionalization (14, 26, 41–43). Finally, despite their importance in the daily life of institutions, the difficulties encountered by the managers (medical, social, medico-social, administrative, etc.) of the services devolved to the APSDMHD are rarely mentioned in the literature (9, 29, 31, 34, 38).

Faced with this observation, one of the answers, brought by the literature, seems to be found in the development of gerontopsychiatry (5, 27, 44–46). Unfortunately, this field developed at the interface of several disciplines (gerontology, psychology of aging, neurology, psychiatry, geriatrics) is still in its early stage and has so far provided very little epidemiological and clinical data (27, 30, 34, 47). Recently, a consensus seems to be emerging. It specifies the need to develop more research (epidemiological, clinical, therapeutic, etc.) on APSDMHD, but also more work for the states, institutions and practitioners concerned by this issue (9, 26).

Secondly, many articles associate the deinstitutionalization process with an economic logic. It should be remembered that APSDMHD generally have many atypical clinical characteristics, somatic comorbidities and sometimes cognitive impairments, and that the management of these individuals leads to higher costs of care (36, 48). Also, recalling that the proportion of APSDMHD is constantly increasing (28, 49), a mechanical increase in the costs associated with these individuals is estimated by several publications without providing any figures nor accurate data (16, 22, 28, 48–51). Only Bartels et al. suggest that health care support for APSDMHD-H incurs a cost that is two to three times higher than that of their counterparts without severe and disabling mental disorders (21). However, conducting a numerical evaluation to highlight the increased costs incurred by the APSDMHD-H population is quite complex (49, 52). Indeed, studies have difficulty integrating the overall cost of care for this population because the epidemiological databases on which health economics research is based rely on systems only including demographic factors and diagnoses (24).

In this context, some authors affirm that the implementation of deinstitutionalization policies for APSDMHD-H lead to a reduction in costs (17, 39, 49, 53, 54). Only the Canadian study by Reinharz et al. in 2000, focusing on individuals over 50 years of age with severe and disabling mental disorders, was able to shed light on the figures based on a retrospective study carried out between 1989 and 1997 (39). It appears that the average annual cost of care for an individual hospitalized in a long-term care facility was $34,455, while that of an individual who had benefited from a deinstitutionalization process was $31,696. The authors also showed that over the years, this difference had been increasing ($70,109 for hospitalized people and $56,095 for deinstitutionalized people in 1996-1998), suggesting that this increase in costs for APSDMHD-H will continue in the years to come.

Finally, while some authors rely on an economic argument in favor of the deinstitutionalization process, others point out that mental health services for the elderly are generally underfunded in relation to the needs of this population (34, 50, 52–56). In addition, the literature focusing on economic dynamics emphasizes (but is not evidence-based) that despite the proven effectiveness of community-based care for APSDMHD (9), community-based care could reduce unnecessary costs by being more efficient, better organized, identified, and visible (47, 51, 52). This is to limit costly recourse to hospitalization (39), rehospitalization (21, 30, 35), or placement in retirement homes (8, 12, 21, 34). These publications do not however provide precise figures.

A third category of results highlighted the presence of an argument linking the difficulties of the deinstitutionalization process to the phenomenon of stigmatization of APSDMHD. While the stigmatization of older people with mental health problems has been highlighted by many selected studies (8, 22, 25, 42, 44, 57, 58), many authors point out that APSDMHD are subject to a double stigma based on both the stigma of aging and the stigma of mental illness (8, 22, 42, 49, 59). Recently, the systematic review conducted in English and French by Clesse et al. was able to recall that aging individuals with mental disorders are perceived as “unsightly and ugly,” “carriers of difficult and/or violent behavior,” “perceived as a cost to society” and “with diminished or declining cognitive abilities” (8). These representations sometimes lead to the mechanism of self-stigma (8).

Present in the general population, the impact of this double stigmatization would be just as pronounced at the professional level. Thus, some works recall that within the psychiatric field, health professionals adopt a pessimistic and nihilistic stance induced by the idea that this population is “incurable” (22, 42), “unaware of its environment” (55, 59) and by long-term hospitalization (55). Outside the psychiatric field, these individuals are also negatively perceived by certain medical specialties fearing that they will have to manage complex situations correlated to the diagnoses, bed availability/problems and the difficulty of reorienting this population (14, 22, 46). Similarly, some authors recall the difficulties encountered by non-psychiatric institutions by highlighting the resistance of professionals who consider the APSDMHD to be inadequate to these services because of their behavior (34, 38, 60). More recently, a French study of 790 health professionals working with individual carrying psychiatric disorders showed that these professionals had the same representation as the general population (59). These professional attitudes can then lead to the hasty idea that these individuals supported by psychiatry present major adaptation difficulties and lead to professional counter-attitudes (42), apprehension (42), or even the emergence of discrimination in state, institutional and professional support (8, 12, 14, 59).

Finally, the presence of these stigmatizing elements induces the idea that accompanying APSDMHD is too complex or impractical, which often leads to the emergence of institutional resistance (34, 38, 42, 61). On this point, some authors have pointed out that stigma and discrimination mechanisms may have kept APSDMHD away from community support policies (12, 59, 62). Faced with these elements, many state or supranational institutions (2, 3, 63, 64), have recalled that stigmatization mechanisms have deleterious consequences on the organization and delivery of community services, as well as on the living conditions of older people with mental disorders (5, 22, 42, 57, 58). Nevertheless, the complex mechanisms inherent in the diffusion of the dynamics of discrimination within state, institutional and professional policies are still very little studied, and some authors still point out the need to produce more work in this field (8, 59).

Results demonstrated that geriatric psychiatric subpopulation are among the most successfully reoriented in community programs, leading mainly to a reduction of hospital use (60). In comparison to hospitalization, receiving care in the community significantly improves quality of life (particularly in the following dimensions: social skills, recreation and living situation) and life satisfaction (60, 62, 65, 66). It also decreases the yearly hospitalization duration (65), the use of psychopharmacology treatment and the need of case manager supervision (66). Receiving care in community also lead to significant improvements in communication and social contact, autonomy, global functioning, cognitive skills and psychiatric symptoms (60, 62, 65, 67–70). Improvements of participant's psychiatric, social and functional statues appears stable and progressive over time, with the absence of significant deterioration in overall functioning when transferred to community homes (60, 67, 68, 71, 72). Selected studies also demonstrate that APSDMHD express a clear preference for community care (70%) (73), perceived more independence (86%), privacy (93%) (65), and feel more satisfied (72). There is however a negative influence of age on relocation, indicating that younger patients (<70) could benefit more from deinstitutionalization programs (70).

It is as well mentioned that social anxiety symptoms are commonly reported among elderly patients with mental disorders after hospital discharge (71). This is interpreted as a consequence of long-time hospitalization characterized by a reduction in social functioning or communication skills, and the fact that patients in the community need to mainly manage their stress by themselves (71). A decline of global cognition and processing speed is also noticed after a few years in community, but it is mainly due to a genuine decline and aging (70). The most vulnerable patients are at greater risk of injuries than with full hospitalization (60, 67) requiring specific attention from staff.

Finally, it appears that a successful relocation requires careful planning and preparation (participants informed months before relocation, organization of counseling and visits, family members as active participants of the preparation process), an initial 24-h supervision, specific age related medical and psychiatric services, adequate housing supported by government funding, acceptance by local residents, sufficient funds to purchase daily necessities (even if this population is generally below the poverty line) and a routine measure of patient's own attitudes and preferences with detailed patient's problems and needs (65, 66, 71–73).

One of the remedies to institutionalization of APSDMHD-H is based on the development of community-based services (1, 7, 9, 24, 72, 73). There is however a limited supply of community care and a lack of alternatives to hospitalization (30, 32, 34, 42, 46, 55, 60). Similarly, it would appear that existing services are considered difficult to access and present problems of use (32, 50, 54, 74), by being, in particular, poorly defined and poorly identified (46, 52). Very little data are available in the West (42). For this reason, none of the selected items provide a detailed inventory of existing services and the obstacles encountered. Finally, the lack of innovative structures in terms of social reintegration (7, 27, 30, 60, 75) as well as the lack of specific training on community support for APSDMHD (7, 10, 16, 26, 30, 42, 76), shows that the community care offer for APSDMHD remains to be improved. Where structures are identified, they appear to focus primarily on continuity and coordination of care (46). They may then abandon the psychosocial dimension associated with the deinstitutionalization process (10, 18, 25, 32, 55, 57). There is also evidence that the needs of individuals with psychiatric support in terms of recreation, culture or social connection are poorly addressed (10, 26, 37, 42, 75, 77). Some publications also denounce the mechanism of ghettoization of PSDMHD in poor and disadvantaged neighborhoods (55, 56).

Finally, it seems that the support needs of APSDMHD are still difficult to assess because of the difficult overlap between mental health and geriatrics (32, 46, 78), leading to both specific multi-professional care (46) and improved coordination between services (42, 74, 75). However, the confusion still existing between gerontopsychiatry and psychogeriatry (16, 44, 78, 79) as well as the difficulties existing between psychiatry and gerontopsychiatry teams (44, 79) show that community support for APSDMHD remains very difficult in Europe (9, 16, 47) and in the USA (76). In the end, all these studies point out that the process of deinstitutionalization of the APSDMHD cannot be fully realized due to the lack of numerous and efficient community alternatives.

During our categorization, 37 articles indicated the emergence of a phenomenon related to the mechanism of deinstitutionalization of APSDMHD-H: the phenomenon of transinstitutionalization. Favored by the lack of available psychiatric hospital beds and the decrease in length of stay (42, 53, 75, 80, 81), the phenomenon of transinstitutionalization describes the migratory process by which a patient is transferred from one institution to another (52). The greatest paradox of the deinstitutionalization movement initially dedicated to APSDMHD-H has been the transfer to the sector concerned by aging (retirement homes, etc.) of the support for APSDMHD initially cared for in hospital (24, 34, 49, 54, 74, 75, 80–82). This, when many of these individuals could have benefited from community support.

A significant body of literature has highlighted the fact that APSDMHD-H have benefited very little from the deinstitutionalization movement (32, 34, 38, 42, 74, 83, 84). Again, there is few available quantitative data on this point. The work of Kermis (32) and Talbott (49) show that in the late 1980s in the USA, only 25% of APSDMHD lived in the community. In 2005, Grabowski et al. (56) estimated that the number of APSDMHD living in psychiatric hospitals was 54%. More recently and in France, this trend was confirmed in 2011 with only 7% of the APSDMHD lived in independent housing (84). The latter is in line with Jovelet's estimate (34) that only 13% of APSDMHD-H leave the psychiatric hospital in favor of inclusion in the community. Finally, these figures are also linked to 2009 data recalling that 44% of French APSDMHD were living in psychiatric hospitals (85).

Typically, in many westernized countries the address of the transinstitutionalization dynamic is the retirement home (16, 30, 34, 38, 49, 54, 75, 82, 83). As early as the 1980s, this movement was already denounced by Goldman et al. (78), Kermis (32), and Freiman et al. (52), for whom 30% of APSDMHD-H were directly referred to retirement homes. Today, current statistical data show that this dynamic is still at work (74). Thus, in the USA, between 2009 and 2011, 500,000 APSDMHD resided in institutions dedicated to aging (retirement home etc.) (54, 56). Similarly, in France, between 2011 and 2018 this figure was 40,000 or 28% of the population (34, 84). At the dawn of its 60 years of existence, some authors therefore consider that the deinstitutionalization movement has not proven to be efficient for APSDMHD (74, 81). Others view retirement homes today as new mental health care homes (34, 42, 54, 56, 75).

In parallel, scientific literature has identified some of the factors that have promoted and driven the transinstitutionalization mechanism for APSDMHD-H. The first factor is based on the preconceived idea that aging can lead to premature loss of autonomy in individuals suffering from psychosis (30, 34, 42, 52). Another factor is based on the fact that the association of psychiatric, cognitive and somatic pathologies implies a strong coordination of care that is difficult to achieve within a community-based approach (28, 30, 33, 34, 55, 86). A third factor seems to be related to the lack of effective therapies for APSDMHD, given that most of the care offered to this population remains focused on medication (18, 26, 53, 82, 87). In the absence of clinical studies, drug therapy is then routinely described as inappropriate (29). This is even though some authors point out that, taking into account the pharmacokinetic data and the risk of drug iatrogeny, doses of psychotropic drugs are progressively limited with advancing age (28, 29, 33, 34). A fourth factor is related to the economic interest of these guidelines, which are considered less costly for the state (34, 52, 56, 75). Similarly, referral may be facilitated by a funding system (less favorable to the problem of aging than to psychiatric problems) that is more advantageous for the states (16). Today, these elements tend to be reinforced by certain economic policies aimed at limiting the crisis of older age institutions in the face of the heterogeneity of the publics they cater for (30, 56).

Despite the increase in financial resources made by some countries, these are considered insufficient (15, 56, 75). Thus, the provision of these structures with staff and activities for their residents is often perceived as insufficient (30, 37, 42, 76, 81). Likewise, professionals do not seem to be really trained to accompany aging psychological problems (30, 34, 38, 42, 87). As a result, these institutions struggle to provide the necessary psychological care (24, 32). In the end, this dynamic could lead to an accentuated degradation of the residents accommodated (32, 34). All these elements are now considered inherent factors in the process of transinstitutionalization of APSDMHD.

The findings of our literature review allow us to make a few recommendations for researchers and professionals concerned with APSDMHD. First, it seems essential to encourage the conduct of epidemiological studies on APSDMHD and APSDMHD-H (16, 29, 76). As well, pharmacological studies that go beyond symptom reduction by questioning the recovery process would be very useful (14). Then, recalling that adapted cognitive and psychoaffective evaluations for this population are recently developed (94) while cognitive remediation and psychosocial rehabilitation programs allow significant cognitive and social gains, these latters would benefit from developing a specific adaptation to the problems of APSDMHD (5, 25, 77). Similarly, longitudinal studies based on homogeneous cohorts evaluated using valid and standardized tools should also be encouraged (30).

Secondly, the recognition of research work on APSDMHD would make it possible to federate the scientific community within the gerontopsychiatric stream and facilitate epistemological identification, which is still difficult (16, 47, 76). This perspective would allow the emergence of an identifiable theoretical trend that would ultimately facilitate training for students (76) and professionals (22, 46, 77).

Third, for each country, an inventory of existing community structures that includes a possibility of support for APSDMHD would allow for an effective evaluation of existing networks (associating psychiatry, geriatrics, and aging professionals), territorial disparities and the needs of this population (24, 28, 47, 58).

Fourth, studies on the mechanisms of stigmatization of APSDMHD and the institutional consequences of stigmatization would provide useful insights. The latter could potentially identify the barriers that caused the transinstitutionalization mechanism. In doing so, it would then be possible to develop appropriate training dynamics based on scientific advances in the field of APSDMHD, concerned about their well-being and guaranteeing equity in access to care.

One of the main strengths of this work is that it is the first to focus the body of literature on the deinstitutionalization of APSDMHD-H. Similarly, the association of a second language (French) has shed more light on the issue. Also, the systematic used methodology let us to gather all existing publications on this topic and provide an historical perspective. On the other hand, due to a lack of specific information on this topic, it has not been possible to point potential variations between geographical area regarding the deinstitutionalization of APSDMHD. Finally, the small number of writings, and paucity of qualitative and quantitative data did not allow for a selection based on methodology. This review then integrates a variety of writings including clinical and economic studies, but also some positions of clinical authors and literature reviews.