A Systematic Review of Specialty Dementia Care Units in Long-Term Care Settings

Abstract

Background:

Alzheimer’s disease and related dementias (ADRD) are common among nursing home residents. Yet, conclusive evidence regarding best care practices among this population is lacking. Objectives of this systematic review were to explore features of dementia specialty care units (DSCUs) in long-term care settings and examine benefits for residents, staff, families, and facilities.

Methods:

PubMed, CINAHL, and PsychINFO were searched to identify articles involving DSCUs in long-term care settings published in English with full text available between 01.01.2008–06.03.2022. Articles containing empirical data about ADRD special care in long-term care settings were included in the review. Articles focused on clinic-based or out-patient dementia care programs (e.g., adult day care) were excluded. Articles were categorized based on geography (U.S. vs. international) and study design: interventions, descriptive studies, or comparison studies (traditional vs. specialty ADRD care).

Results:

Our review included 38 U.S. articles and 54 articles from 15 international countries. In the U.S., 12 intervention, 13 descriptive, and 13 comparison studies met the inclusion criteria. Articles from international countries included 22 intervention, 20 descriptive, and 12 comparison studies. Results were mixed in terms of the efficacy of DSCUs. Promising DSCU features include small-scale settings, dementia-educated staff, and multidisciplinary approaches to care.

Conclusion:

Overall, our review did not find conclusive evidence regarding the benefits of DSCUs in long-term care settings. No rigorous study designs were found examining ‘special’ features of DSCUs and associations with outcomes among residents, family, staff, and the facility. Randomized clinical trials are needed to disentangle the ‘special’ features of DSCUs.

Brief Summary:

Overall, our review did not find conclusive evidence regarding the benefits of DSCUs in long-term care settings. No rigorous study designs were found examining ‘special’ features of DSCUs and associations with outcomes among residents, family, staff, and the facility.

I. INTRODUCTION

More than half of adults aged 65 years and over in nursing homes in the United States have a diagnosis of Alzheimer’s disease and related dementias (ADRD), cognitive impairment, or behavioral health disorders.¹ Worldwide, estimates of ADRD among older adults in long-term care settings range from 16.1–85.2%.² Family caregivers of persons with ADRD assist patients with activities of daily living and instrumental activities of daily living, medication and treatment management, and management of behavioral symptoms of ADRD, among other tasks.² Many family caregivers face the difficult decision of transitioning their loved ones to long-term care settings for advanced nursing and custodial care as symptoms of ADRD progress.

However, more than 30% of U.S. facilities from a random sample were challenged in the care of residents with ADRD and other behavioral health disorders, reporting unmet care needs, inadequate staff behavioral health education, inadequate infrastructure, and inadequate care coordination with community providers.³ These facility-level challenges are associated with increased prevalence of suicidal ideation, severe depression, and severe aggressive behaviors among residents with ADRD.⁴ Conclusive evidence regarding best care practices for older adults with ADRD in long-term care settings is lacking and is greatly needed to maximize quality of life among this population.

Dementia specialty care units (DSCUs) are often promoted as promising options for persons with ADRD, and are present in 14.9% of nursing homes nationally.¹ Ownership status, geographic region, and level of nurse staffing are predictors of DSCUs in nursing homes, and question whether access to specialty dementia care is equitable for older adults with ADRD across the U.S.⁵ These specialty care units, while not standardized or regulated, often feature specially designed wings for residents with ADRD, staff training in dementia care, and dementia-friendly activities and programming. Noted benefits of DSCUs include reductions in inappropriate antipsychotic use, pressure ulcers, feeding tubes, physical restraints, and hospitalizations.⁶ However, other studies do not find DSCUs to be associated with improved resident outcomes.^7,8 DSCUs in long-term care settings are being examined across countries. Yet, even within countries there is limited conclusive evidence of the mechanism of benefit, if any, from these units.

Several international systematic and scoping reviews highlight the need for further examinations of the possible mechanisms of benefits from DSCUs.^7,9–12 Lai et al. (2009) sought to focus on randomized controlled trials to examine behavioral problems of ADRD. The authors found no such studies but did identify eight non-randomized controlled trials. Among these studies, no significant benefit of specialty care units was found, questioning whether such units are necessary, or whether improving care practices overall would better serve residents with behavioral problems of ADRD.⁷ Adlbrecht et al. (2021) suggest that DSCUs encourage social interactions (although few in number) among residents when staff members initiate interactions among residents (residents were reluctant to initiate interactions on their own). Environmental design features of specialty care units promoting social interactions include common places and encouraging small groups, but specific features remain unknown and the relationship with staff interaction remains unclear.⁹ Additionally, Gilbert et al. (2019) examined organization and standardization of care among residents with dementia. They found no clear evidence informing quality nursing care among this population in DSCUs, but highlight the importance of resident-to-staff ratios, staff education and job satisfaction, and individualized nursing care.¹⁰ Ausserhofer et al. (2016) took a different approach and explored the impact of home-like care models on resident, family, and staff outcomes (e.g., resident physical functioning and satisfaction with care; family satisfaction, amount of assistance, family burden, and contact with residents; and staff-reported quality of work life, job satisfaction, motivation, burnout, and work-related mental health problems). While no improvement was found in family satisfaction, caregiver burden, or staff-reported burnout and job satisfaction, physical functioning improved among residents with ADRD living in specialty care homes.¹² Most recently, Harrison et al. (2022) examined the effects of home-like care settings for residents with ADRD. The authors were unable to find sufficient evidence in the studies reviewed to demonstrate the efficacy of environmental design on resident quality of life, behavior, mood, and depression using rigorous study methods.⁹

Theoretical framework

We adopted an expanded palliative care model approach to our systematic review. Patient-centered care provides care that is “respectful of and responsive to individual patient preferences, needs, and values and ensuring that patient values guide all clinical decisions”¹³. Palliative care expands patient-centered care to “provide relief from pain and other symptoms, support quality of life, and focus on patients with serious advanced illness and their families.”¹⁴ Our approach expands palliative care to include residents with ADRD, their families, nursing home staff, and long-term care communities as the unit of care (Figure 1). The focus of our model is on the quality of life for the four stakeholder groups. This includes stakeholders’ spiritual, physical, social, and cognitive-emotional well-being, and the healthcare services promoting these domains. Stakeholders include not only residents, families, and nursing staff, but also nursing home administrators and business operators as well as custodial, environmental, and food service staff, among others. In short, all who are involved in the care of older adults with ADRD in long-term care communities are included in our model.

Figure 1.

Conceptual framework. Adapted from: Cagle J. G., & Widera, E. W. (2014). Geriatrics and palliative care. In B. Williams, A. Chang, C. Ahalt, H., Chen, R. Conant, S. Landefeld, C. Ritchie, & M. Yukawa (Eds.), Current diagnosis and treatment: Geriatrics, 2nd edition. New York: McGraw Hill.

Given the increasing number of persons with ADRD worldwide likely to transition to long-term care settings and the uncertainty surrounding best care practices among older adults with ADRD in these settings, our objectives in this systematic review were to: 1) explore features of specialty dementia care units in long-term care settings and 2) examine benefits for residents, staff, families, and long-term communities.

II. METHODS

Data sources and search strategy

GB searched PubMed, CINAHL, and PsychINFO to identify articles involving specialty dementia care in long-term care settings published in English between 01.01.2008–06.03.2022. This starting date was selected because 2009 was the earliest publication date among the systematic reviews that came out in our initial search. Search term combinations included phrases related to ‘dementia’, ‘nursing homes’, and ‘special care units’. See the Appendix for the full list of search term combinations employed.

JO was primarily responsible for the initial screening of titles and abstracts. Articles were deemed ineligible during this stage if there was no mention of dementia specialty care in long-term care settings; DF and JC each reviewed a selected 10% of titles and abstracts deemed ineligible for confirmation. Among the titles and abstracts reviewed by DF and JC, confirmation of ineligibility was affirmed in all cases. During the full text screening, studies were selected for inclusion if they provided empirical data – whether quantitative, mixed methods, or qualitative – about dementia special care in long-term care. Articles focused on clinic-based or out-patient dementia care programs (e.g., adult day care) were excluded. Commentaries, editorials, systematic reviews, study protocols, and conceptual articles without novel data were excluded. During the full-text review, JO and JC screened all articles for eligibility. For selected studies, JO, DF, and JC read and extracted key study elements.

Given the unique characteristics of the U.S. health care system and its approach to caring for persons living with dementia, results were stratified into two categories: U.S. studies and non-U.S. studies. Studies were also further categorized based on (1) the primary unit of analysis (residents, family, staff, or facility); (2) whether the study reported data comparing dementia specialty care to non-dementia specialty care (e.g., usual care); and (3) whether a study was testing an intervention within a dementia special care environment.

III. RESULTS

Our search strategy initially identified 5,537 articles for screening (Figure 2). After exclusions, 92 articles were identified as eligible and included in our systematic review: 38 articles from the U.S. and 54 articles from 15 international countries.

Figure 2.

PRISMA diagram.

Intervention Studies

United States

Among the 38 selected studies from the U.S., 12 studies described interventions within the dementia special care environment (Table 1.A). One study (two articles) described a randomized controlled trial that evaluated the use of multidisciplinary geriatric team assessments using validated measure scales.^15,16 Residents were evaluated by a medical social worker, a dietician, a geriatrics advanced practice nurse, a physical therapist, and a geriatrician. Findings from this study suggest the involvement of such an assessment team may reduce fall risk among residents; however, the results were not statistically significant.

Table 1.A:

Summary of reviewed intervention articles in the United States (N=12).

Intervention	Study Design	Outcomes	Measures	Results Synopsis
Introduction of dementia specialty care unit (DSCU)8	Quantitative, Retrospective	Physical restraints, feeding tubes, psychotropic medications	On-line Survey Certification and Reporting System (OSCAR)	No associations
Presence of DSCU17	Quantitative, Retrospective	Quality indicator: Incidence of worsening behaviors	Minimum Data Set	Important to risk-adjust analyses for resident cognitive status and presence of DSCU
Memory care license18	Quantitative, retrospective	Hospital and nursing home admissions	Medicare data	Reductions in hospital and nursing home admissions
Multidisciplinary geriatric team asessment15,16	Randomized clinical trial	Unanticipated transition from assisted living	Mini-Mental State Examination (MMSE); Katz Activities of daily living (ADL) index; Behavioral Pathology in Alzheimer’s Disease Rating Scale; Berg Balance Scale	Reduced risk of unanticipated transitions in the intervention arm; results not statistically significant
Moderate intensity group exercise activity19	Quantitative, pre/post study	Negative affect	Philadelphia Geriatric Center Apparent Affect Rating Scale; MMSE; Katz ADL	Intervention associated with declines in negative affect
Free-form painting program21	Quantitative, pre/post study	Engagement, communication, cognitive-linguistic function, quality of life	Menorah Park Engagement Scale, Conversational Act Profile, ABCD, Quality of life-Alzheimer’s Disease (QOL-AD)	Non-significant improvements in outcomes
Dementia wander garden20	Quantitative, single blind pre/post study	Falls, antipsychotic use	N/A	Outcomes decreased with access to intervention
Tablet use25	Quantitative, survey	Feelings of social isolation, mood	Researcher created measures	Tablets increased mood, decreased feelings of social isolation, were beneficial to staff
Fit bit activity monitor23	Qualitative, pilot study	Acceptability and feasibility of telehealth monitoring	Montreal Cognitive Assessment; the Clinical Dementia Rating Scale; Bristol Activities of Daily Living Scale; Neuropsychiatric Inventory−Questionnaire; QOL−AD	Positive acceptance and feasibility of intervention
Montessori methods24	Qualitative, field notes	Barriers to implementing Montessori methods	Researcher created measures	Barriers: respect, interdependence, meaningful interaction, and structure
Environmental design: large vs. small dining settings22	Qualitative, case study	Resident patterns of socialization and interaction at mealtimes	N/A	Residents enjoyed meals more in the smaller dining settings

Three studies used administrative data to examine associations between presence of DSCUs and memory care licenses with facility-level quality indicators (e.g., incidence of worsening resident behaviors, medication use, hospital and nursing home admissions).^8,17,18 Use of physical restraints, psychotropic medications, feeding tubes, and hospitalizations were less likely among residents with dementia of long-term care communities with specialized dementia care (results not statistically significant); overall care in these communities was more supportive of resident well-being. These studies highlight the importance of risk-adjusting analyses for resident cognitive status and presence of DSCUs and question the utility of administrative data to capture the impact of DSCUs on care practices.

Moderate-intensity chair-based exercise, free-form art activities, and wander gardens were examined in three studies using pre/post study designs.^19–21 These interventions were associated with decreased negative affect, depression, and agitation among residents with dementia. Three qualitative studies examined Montessori methods (e.g., personalized activities based on respect, interdependence, meaningful interaction, and structure), telehealth, and dining environmental design.^22–24 Residents with ADRD preferred smaller home-like dining settings, which were associated with increased socialization.²² Nursing home staff appreciated telehealth monitoring, in part, because: objective resident measurements can be collected and residents are able to remain in the nursing home during medical visits avoiding disruptive care transfers.²³ Several keys to positive intervention implementation are nursing home staff who are empowered and educated in the intervention prior to implementation. Without these, barriers may impede rollout and sustainability of interventions.²⁴ One study in the U.S. examined whether tablet computers serve as a meaningful tool to connect isolated residents with family and friends; results demonstrated improved mood and decreased social isolation among residents.²⁵

International Countries

Of the 54 selected international studies, five intervention studies were conducted in the Netherlands, including three cluster randomized controlled trials,^26–30 one non-randomized controlled trial,^31–33 and one pre-post comparison study.³⁴ Interventions were designed to address resident challenging behavior (Grip on Challenging Behavior)^26–28, sleep (caffeine reductions)³⁴, psychotropic drug use (web-based tool)²⁹, staff education programs (TENSE)³⁰, and environmental design (DSCU size)^31–33 (Table 1.B). Findings from studies examining resident outcomes are generally positive (e.g., improved sleep with caffeine reduction³⁴ and behavioral functioning in smaller home-like settings,^31–33 decreased use of psychoactive drugs²⁹). However, findings from staff education studies are mixed. One such study, Grip on Challenging Behavior, which utilizes detection, analysis, treatment, and evaluation to monitor challenging behavior, was associated with increased job satisfaction.^26–28 However, The Educating Nursing Staff Effectively (TENSE) program had no significant effect on outcomes (e.g., stress, personal accomplishment).³⁰ Results regarding environmental design and small-scale homelike settings housing 7–8 residents were inconclusive but trended towards positive associations.^31–33

Table 1.B:

Summary of reviewed intervention articles in international countries (N=22).

Intervention	Study Design	Outcomes	Measures	Results Synopsis
Grip on Challenging Behavior26–28	Quantitative, cluster randomized controlled trial	Challenging behaviors, staff burnout	Cohen-Mansfield Agitation Inventory (CMAI) and Neuropsychiatric Inventory, Maslach burnout inventory: subscales of the Leiden Quality of Work Questionnaire	Reductions in challenging behaviors; no association with physical restraints; positive effects on job satisfaction; suboptimal intervention implementation
Caffeine reduction in the afternoon and evening34	Quantitative, pilot pre/post	Sleep, behavioral symptoms, psychotropic medications	Neuro Psychiatric Inventory- Nursing Home edition	Improvements in sleep scores; no improvements in other outcomes
DSCU size31–33	Quantitative, non-randomized controlled trial	Rest-activity rhythm, cognition, use of psychotropic drugs, memory, executive function, language and praxis functioning	Cognitive and behavioral functioning, Anatomical Therapeutic Chemical Classification system, actigraphy and observation scales	Improvements in cognition; no associations with other outcomes
Web-based tool for identifying BZRA prescriptions29	Quantitative, cluster randomized controlled trial	Prevalence of BZRA use	Dutch guidelines	BZRA use is consistently high and not in line with guidelines
TENSE training program30	Quantitative, cluster randomized controlled trial	Staff stress, work contentment, stress reactions	Utrecht Burnout Scale - C	No significant effect on outcomes
Marte Meo Counselling45,46	Qualitative, interviews and focus groups	Perceived learning	N/A	Increased learning, positive learning climate
Companion clowns40	Quantitative, pilot time series	Behavioral and psychological symptoms of dementia	Neuro-Psychiatric Inventory	Little to no effect on BPSD
Relocation43	Quantitative, prospective	Neuropsychiatric symptoms	Neuro-Psychiatric Inventory	Improved psychiatric symptoms after relocation to DSCU
Staff wearing street clothes35	Mixed, natural experiment	Quality of life	QOL-AD; MMSE	Quality of life improved when staff wore street clothes
Staff/resident shared lunchtime meals36	Mixed, natural experiment	Social interactions, support, autonomy	MMSE	Intervention improved outcomes
Shared housing vs. DSCU37	Quantitative, longitudinal	Quality of life	Barthel Index, neuropsychiatric symptoms of dementia, Cohen-Mansfield agitation inventory, Quality of life in dementia	No difference in quality of life
Segregated vs. partially segregated care38	Quantitative, cross sectional	Quality of life, quality of care	Researcher-created structured care and behavior assessment	Special dementia care was associated with more social staff contact, more involvement of volunteer caregivers, more involvement in home activities, and fewer physical restraints.
Acupressure and Montessori-based activities41	Quantitative, double blind randomized trial with crossover	Agitated behavior	Cohen-Mansfield Agitation Inventory, Ease-of-Care, Apparent Affect Rating Scale	Traditional and Western approaches address agitated behavior
Horticultural therapy39	Quantitative, matched pairs randomized controlled trial	Apathy, cognitive ability, quality of life, functional capacity	Apathy Evaluation Scale-informant version, MMSE, Quality of Life in Alzheimer’s Disease Scale, Barthel index	Reductions in apathy; improvements in cognitive function; no association with quality of life and physical functioning
Presence of dementia care program47	Quantitative, cross sectional	Provision of dementia care program	Researcher-created measures	Dementia care program more likely present along with other regulated care services
Dyadic interactions44	Quantitative, quasi-experimental single case study	Problematic person transfer situations	Dyadic Interaction in Dementia Transfer Assessment Scale, Pain Assessment in Advanced Dementia Scale, Resistiveness to Care Scale for Dementia of the Alzheimer’s Type	Improved transfer-related behaviors associated with dyadic interactions
Music engagement program42	Mixed, pre-post	Emotional well-being, social connectedness, depression	Cornell depression scale; WONCA charts for “feelings” and “social activities”	Improved mood, calmness, reduced aggression. Intervention is not sustainable after study exit.

Studies from eight other international countries examined resident quality of life,^35–39 behavior,^40,41 social connectedness,⁴² neuropsychiatric symptoms,⁴³ and staff transfer behaviors⁴⁴ and learning outcomes^45,46 using interventions ranging from horticultural therapy and staff clothing attire to counseling techniques. Two natural experiments examined resident quality of life in relation to staff behaviors of wearing street clothes vs. uniforms when caring for residents with ADRD³⁵ and staff sharing lunchtime meals with residents.³⁶ Overall resident quality of life improved when staff wore street clothes and nutrition increased when staff shared meals with residents. Wolf-Ostermann et al. (2012) found inconclusive results using a longitudinal design in Berlin examining shared housing vs. DSCU environments; the authors conclude that resident preference may be a better indicator of housing placement. A study in Hamburg also examined environmental design and found that resident quality of life (e.g., social contact with staff, increased home activities, decreased use of physical restraints, higher presence of psychiatrists) was improved among residents with ADRD living in DSCU environments compared to traditional housing.³⁸ Planting, handicraft, and dietary horticultural therapies were used among residents with ADRD in China; Yang et al. (2022) found improvements in resident apathy and cognitive function, but no difference compared to usual care in terms of quality of life and functional capacity.

Acupressure and Montessori-based methods,⁴¹ and companion clowns were used to examine associations with resident behaviors.⁴⁰ Physically nonaggressive behaviors, agitated behaviors, and aggressive behaviors were improved among residents with dementia receiving acupressure and Montessori-based methods, as were ease-of-care ratings and affect. Use of companion clowns was not associated with significant improvements in neuropsychiatric symptoms, but residents did experience improved aberrant motor behavior, euphoria, hallucinations, and delusions with companion clowns; however, companion clowns were associated with increased disinhibition, apathy, depression, and appetite disorders.

Social connectedness was evaluated among residents with ADRD in Australian nursing homes using the music engagement program (MEP).⁴² Although staff did not find it feasible to continue the MEP after intervention implementation (citing time constraints as a major hindering factor), residents with ADRD did experience improved depression scores, mood and calmness, and reduced aggression while in the MEP. Neuropsychiatric symptoms were assessed before and after relocation to a DSCU;⁴³ behavioral and psychological symptoms of dementia were significantly reduced 1 month after relocation.

Three international studies examined interventions among staff: one explored staff transfer behaviors⁴⁴ and two examined staff learning outcomes.^45,46 Interventions related to staff transfer behaviors were tailored to each care dyad; some interventions included eye contact, preparing the resident for movement, removing toilet handles, division of turning movements, and additional support for the resident. Overall, interventions were associated with improved verbal and nonverbal discomfort and staff transfer-related behaviors.⁴⁴ Marte Meo counseling is a technique where staff-resident care interactions are video-recorded, analyzed, and discussed with nursing home staff. Findings from these studies indicate that nurses gain additional knowledge about themselves and the resident by reviewing the care-transfer recordings,⁴⁵ but this is dependent on the learning climate in the DSCU.⁴⁶

One international intervention study was conducted in an assisted living facility and examined predictors of dementia specialty care programs; findings suggest that larger facilities and those offering more regulated care services were more likely to have dementia specialty care programs.⁴⁷

Descriptive Studies

United States

Descriptive studies conducted in the U.S. included surveys and interviews,^48–53 observational studies,^54–56 administrative data analyses,^57,58 and case and descriptive studies (Table 2.A).^59,60 Surveys and interviews conducted with residents, administrators, and caregivers of persons with ADRD in long-term care communities revealed the importance of social connections among residents for improving resident quality of life.^48,49 Additionally, findings demonstrated racial disparities between white and non-white residents in specialty dementia care (white residents were more likely to receive care in DSCUs than non-white residents)⁵⁰ and variations in transitions to DSCUs among different family members and type of long-term care facility.^51–53 Observations of residents with ADRD revealed increased risk of falls associated with increased hearing loss,⁵⁴ and potential approaches to enhance resident navigation (e.g., address difficulties locating one’s room, not recognizing the environment, entering others’ rooms by mistake, forgetting room numbers, among others)⁵⁵ and improve resident-to-resident interactions (e.g., modify situational and evolving triggers).⁵⁶

Table 2.A:

Summary of reviewed descriptive study articles in the United States (N=13).

Outcomes	Study Design	Measures	Results Synopsis
Social networks and quality of life48,49	Quantitative, surveys and interviews	Mini-Mental State Examination (MMSE), egocentric and sociocentric network analysis	Social integration was positively associated with quality of life; the association between social integration and cognitive status was inconclusive.
Special dementia care50,57	Quantitative, survey and descriptive	National nursing home survey	Unequal geographic distribution and racial disparities in special dementia care
Care transitions51,52	Qualitative, in-depth interviews	N/A	Variations in care transitions based on facility type and family caregiver experiences
Labeling in dementia special care53	Quantitative, surveys and interviews	Researcher-created measures	Labeling a resident as living in dementia special care unit (DSCU) was associated with reports of more problematic, inappropriate, and aggressive behavior than residents labeled as non-DSCU.
Falls54	Quantitative, observational	MMSE	Hearing loss associated with greater risk for chronic falls among residents with dementia
Spatial disorientation and wayfinding difficulties55	Qualitative, observational interviews	Cohen-Mansfield Agitation Inventory; Observed Emotion Rating Scale	Residents with dementia were challenged with spatial disorientation and wayfinding difficulties
Resident-to-resident aggressive behavior56	Qualitative, observational interviews	Cohen-Mansfield Agitation Inventory; Observed Emotion Rating Scale	Resident-to-resident aggressive behaviors are situational
Nursing home occupancy; Medicare occupancy or private pay occupancy58	Quantitative, longitudinal	N/A	Opening an SCU had a positive impact on overall occupancy
Relationship-centered care59	Qualitative, ethnographic case study	Ethnographic methods; participant-observation ethnography as informed by interpretive anthropology	Relationship-centered care is supported by reminiscence and communication
Cognition, physical function, behavioral symptoms60	Quantitative, descriptive	Resident Assessment Tool	Functional and behavioral symptoms did not significantly change; Cognition declined significantly over the 12-month window

Using administrative data and a national sample, Castle (2008) conducted a longitudinal study examining how presence of a DSCU impacts resident occupancy. The author found that opening a DSCU had financial benefits for the facility, as resident payer-mix and occupancy rates improved.⁵⁸ However, as Cornell et al. (2020) found, access to DSCUs is constrained by socioeconomic factors, leaving persons with dementia in lower-resourced communities with limited options for DSCU care.⁵⁷ Further care challenges arise in the end stages of ADRD and during COVID-19 restrictions (e.g., isolation protocols). These situations offer opportunities for staff to provide relationship-centered care.^59,60

International Countries

Three international descriptive studies used a cross-sectional approach to examine resident behavior and facility influences on resident care practices (Table 2.B). Findings indicate that behaviors and care practices (e.g., antipsychotic drug prescribing practices) vary within facilities by ADRD severity and across facilities by facility-level ‘norms of care’ (e.g., ways to address boredom, pain).^61,62 Furthermore, eldercare robots may be an appropriate means to supplement care in certain environments with healthcare monitoring (not hands-on care).⁶³ Additionally, some facilities in Australia utilized specialty dementia care programs for residents with challenging behavior, with the goal of stabilization and return to mainstream care.⁶⁴ Based on 10 years of data, the authors observed reductions in antipsychotics and concluded that such specialty programs may be beneficial for caring for residents with behavior challenges.

Table 2.B:

Summary of reviewed descriptive study articles in international countries (N=20).

Outcomes	Study Design	Measures	Results Synopsis
Antipsychotic use61,72	Quantitative, cross sectional and longitudinal	Global Deterioration Scale, Cohen-Mansfield Agitation inventory; Multi-Dimensional Dementia Assessment Scale	Antipsychotic use varied by dementia type, resident behavior, and facility practice patterns. Use of antipsychotic drugs is high and often inappropriate among residents with dementia in SCUs.
Resident behavior62,64,71,75	Quantitative – cross sectional; quantitative – retrospective; mixed – observational crossover; qualitative – focus groups	Cohen-Mansfield Agitation Inventory; Minimum Data Set (MDS) scales; researcher created measures	Associations between external factors and resident behavior, including staff uniforms. Resident aggression occurs as a process, not an event. A 12-month specialist dementia care program was associated with improved resident behavior.
Participation in care67–69,80	Quantitative – survey; qualitative – observation and interviews	N/A	Relatives advocate for residents’ care; residents participate in care according to their ability.
Nurse perceptions of care74,79	Qualitative, interviews	Researcher created measures	The nurse-resident relationship can be therapeutic; personal and organizational resilience promoted positive staff perceptions.
Staff mental health well-being65	Quantitative, observational survey	General Self-Efficacy Scale; General Health Questionnaire; Maslach Burnout Inventory	Housing residents with similar care needs together was associated with improved staff mental health and well-being
Social interactions70,76	Qualitative and mixed interviews	Therapeutic Environment Screening Survey for Nursing Homes; Professional Environmental Assessment Protocol	Environmental design and appropriate proximities enhanced social interactions among residents and staff
Robot interactions63	Mixed, cross-sectional	N/A	Healthcare robots have the potential to be entertaining, stimulating, and calming for residents with dementia; robots should be adapted for dementia settings
Television watching behavior77	Quantitative, survey	N/A	Television watching by itself was not a particularly good leisure activity; but watching TV with others who are close to residents was a good leisure activity
Advance care planning with family caregivers78	Qualitative, semi-structured interviews	N/A	Advance care practice (ACP) conversations can occur with family members of residents with dementia to anticipate future care decisions as dementia progresses
Behavioral symptoms66	Mixed, observational questionnaires	Neuro Psychiatric Inventory	Consuming caffeine was negatively correlated with apathy and aberrant motor behavior; positive association between caffeine consumption and getting out of bed at night
Nurse aid’s physical strain during person transfer tasks at DSCUs73	Qualitative, focus groups	Researcher created measures	Understanding cognitive decline and improving communication with residents with dementia was associated with less physical strain experienced by staff

Interviews and observational studies were conducted in several countries, exploring family members and nursing home staffs’ perceptions of care and social interactions with residents. These studies found that overall, a meaningful social community is associated with increased resident quality of life and sense of dignity as well as staff well-being.^65–80 Focus groups conducted with nursing home staff revealed that staff were better able to provide care to residents and mitigate resident-to-resident aggression when they understood the trajectory of ADRD.^73,75

Comparison: DSCU vs. TNH

United States

The remaining 13 U.S. studies compared DSCUs with traditional nursing home (TNH) care (Table 3.A). One study found that higher nurse staffing levels, being located in North Central and Middle Atlantic regions, and chain affiliation were predictors of DSCUs.⁵ The authors highlight the geographic disparity in access to specialty dementia care in long-term care communities across the U.S.

Table 3.A:

Summary of reviewed comparison study articles in the United States (N=13).

Outcomes	Study Design	Measures	Results Synopsis
Social connectedness86,92	Quantitative; cross-sectional, observational	Social connectedness index; Apparent Affect Rating Scale, Minimum Data Set (MDS)	Social connectedness scores were lower among residents in nursing homes with dementia special care units, where the proportion of residents with dementia is higher
Resident-to-resident elder mistreatment87	Quantitative, observational	Comprehensive Assessment and Referral Evaluation Diagnostic Cognitive Disorder Scale; Barret behavior index; performance activities of daily living, researcher observations; feeling tone questionnaire	Resident-to-resident elder mistreatment may be at higher risk in dementia special care units. Resident-to-resident elder mistreatment is more likely among residents who are less functionally impaired, show behavioral symptoms, and have milder dementia
Facility-level quality of care and resident-level care processes5,82–84,88,90	Quantitative, cross-sectional	Certification and Survey Provider Enhanced Reporting, Payroll-Based Journal; Nurse staffing (nursing hours per resident day), registered nurse skill mix, MDS, Cognitive Performance Scale, Burrows Depression Scale	Nursing homes with DSCUs were associated with better resident care, higher quality, and more deficiency citations.
Resident quality of life85	Quantitative, cross-sectional	Minnesota Nursing Home Resident and Consumer Satisfaction Survey; MDS	Demonstrated the importance of CNAs and activity staff in resident QOL; interpersonal factors (smaller facilities, better clinical outcomes, adequate staffing) more linked to resident QOL than facility characteristics.
Resident outcomes and facility processes of care81,91	Quantitative, cross-sectional	Mini-Mental State Examination (MMSE), Hamilton Anxiety Scale, Cornell Scale for Depression in Dementia, National Nursing Home Survey	Antipsychotic medications, and resident processes of care, were better in facilities with DSUCs
Infection prevention89	Quantitative, cross-sectional	Researcher-created measures	Infection control was the most challenging in dementia-specific communities

Another study determined that DSCUs do indeed provide ‘specialized’ care through improved care processes (e.g., reductions in physical restraints, improved continence management, increased influenza and pneumococcal vaccines, decreased use of feeding tubes) compared to TNH settings.⁸¹ Residents in DSCUs were hospitalized less frequently and experienced fewer pressure ulcers and weight loss compared to residents in TNHs; however, falls were more likely among residents in DSCUs.⁸¹

Seven studies controlled for the presence of DSCU in their analyses, including one study using instrumental variables and one study using propensity score matching. While these are more rigorous analytical approaches, results from these studies are mixed. In some cases, DSCUs were associated with better resident outcomes and quality of care (e.g., decreased use of feeding tubes, antipsychotic medications, and physical restraints; fewer hospitalizations and pressure ulcers).^82–85 Other studies reported less favorable outcomes among residents in DSCUs, including lower levels of social connectedness⁸⁶ and higher risk for elder-to-elder mistreatment,⁸⁷ deficiency citations,⁸⁸ infections,⁸⁹ and antipsychotic use.^90,91

Results from a pilot study using combinations of social network analysis methods demonstrated that staff-to-resident interactions were more frequent among residents in DSCUs, but this may be due to restrictive admission criteria (i.e., residents in DSCUs were more mobile than residents in TNH).⁹² The study also points to the question of whether social integration is an indicator of quality of life among residents of long-term care communities.

International Countries

Of 13 selected international comparison studies, five found little-to-no difference between residents living with ADRD in DSCUs and TNHs in terms of nutritional status, case conferences, mealtime experiences, and resident care and attention (Table 3.B).^93–97 Although not statistically significant, residents with dementia in DSCUs and TNHs were found to be different in terms of age, strength, and mobility in one study (e.g., residents in DSCUs were stronger and younger than residents in TNHs),⁹⁴ and in terms of content of case conferences (e.g., fewer case conferences regarding challenging behaviors were conducted in DSCUs compared to TNHs) in another study.⁹⁵ However, three studies reported worse outcomes among residents in DSCUs compared to residents in TNHs (e.g., higher rates of antipsychotic medications, social isolation, and burdensome transitions at the end-of-life).^98–100 In contrast, two studies reported higher levels of social interaction among residents in DSCUs and found associations with relaxation and activities.^11,101 Another study questioned whether DSCU and TNH are two separate and distinct populations¹⁰² (i.e., DSCU admission criteria contribute to selection bias).¹⁰³

Table 3.B:

Summary of reviewed comparison study articles in international countries (N=12).

Results Synopsis	Outcomes	Study Design	Measures
Limited difference93–97	Quality of life, nutritional status, case conferences, mealtime experience, care and attention	Quantitative; cross-sectional, observational	Quality of life in late-stage dementia scale; clinical dementia rating scale; physical self-maintenance scale; neuropsychiatric inventory questionnaire; Malnutrition Universal Screening Tool; Dementia Care Questionnaire; Researcher created measures; MMSE, CMAI, Cornell Depression Scale; Barthel functional capacity; CSDD depression
TNH better than DSCU98–100	Drug prescribing, friendship networks, care transitions	Mixed, quantitative; cross-sectional, interviews	computerized national pharmacy drug register; Global Deterioration Scale; observation; self-report; Nationwide registry data
DSCUs better than TNH11,101	Impact of DSCUs, engagement in activities and social interactions	Qualitative, quantitative; focus groups, observational	Researcher created measures; Maastricht Daily Life Observation (MEDLO) Tool; Naomi Feil’s Validation of a person-centered and emotion-oriented approach in care
Inconclusive – DSCU and TNH are ‘separate populations’102,103	Psychiatric and behavioral symptoms, responsive behaviors, quality of life	Quantitative and qualitative; interviews	Neuropsychiatric Inventory, Clinical Dementia Rating Scale, Lawton’s Physical Self-maintenance scale Mechanisms of dementia special care units

IV. DISCUSSION

To our knowledge, this is the first systematic review examining the recent existing empirical research on DSCUs in long-term care settings and their association with outcomes for residents, families, staff, and facilities. Other systematic and scoping reviews, except for Ausserhofer et al (2016), examined specific outcomes among residents with ADRD and did not consider benefits of DSCUs among families, staff, and facilities. Our systematic review is innovative and expansive in this regard. It is also timely given the projected number of older adults with ADRD likely to transition to long-term care settings in the coming years. Additionally, our study builds on previous reviews and includes nearly a decade’s worth of new studies. Several key results from our review for residents, families, staff, and facilities merit further discussions.

Our systematic review sought to explore the benefits of DSCUs on outcomes among residents, families, staff, and long-term care communities by expanding the palliative care model to include long-term care communities and staff members. Most studies (70%) included in our review examined associations between DSCUs and resident outcomes, including the use of physical restraints, psychotropic medication use, feeding tubes, mood and affect, healthcare transitions, engagement and communication, quality of life, falls, social isolation, behaviors, and spatial disorientation and wayfinding difficulties, among others. These outcomes are found in the literature to be indicators of well-being among residents with ADRD and are examined in many other studies. Very few studies (6%) included in our review examined associations among families and long-term care communities; those that did examined family satisfaction with care, and nursing home quality indicators and economic impacts. However, several included studies (24%) examined associations between DSCUs and staff outcomes including burnout, job satisfaction, mental health well-being, physical strain, and learning outcomes. As nursing homes are ‘home’ for residents and where staff provide 24-hour nursing care, it is not unsurprising that the majority of studies focused on resident and staff outcomes. However, to fully capture the essence of dementia specialty care in long-term care communities, family and community outcomes should be equally important and as available.

Although not directly comparable to DSCUs, clustered domestic models of care share similar features including small living units (15 or fewer residents), allocation of care staff to specific living units, residents’ participation in meal preparation, meals cooked in the units, self-service of meals by residents, and independently accessible outdoor areas.¹⁰⁴ Several studies have found associations between these models of care and reduced hospitalizations, system-level costs, and increased resident quality of life,¹⁰⁴ as well as higher consumer rated quality of care scores (including flexibility of care and access to outside gardens)¹⁰⁵ and staffing structures.¹⁰⁶ Future studies are needed to investigate outcome differences between these clustered domestic models of aged care and DSCUs.

Little progress has been made over the past 14 years informing the efficacy of DSCUs in long-term care settings (Figure 3). We remain without clear evidence regarding associations with resident, family, staff, and long-term care community outcomes. However, promising interventions, perhaps the most rigorous of the incorporated studies, from limited studies include licenses for memory care, moderate intensity group exercise, wander gardens, tablet computers, fit-bit activity monitors, small dining settings, caffeine reduction, Marte Meo counseling, staff appearance and interactions with residents, horticultural therapy, and dyadic interactions between staff and residents. While these interventions appear to be beneficial, few of the included studies utilized rigorous methods to examine possible mechanisms of change and those that did utilize randomized controlled trial designs did not find statistically significant results or were limited in generalizability. Furthermore, the included studies lack an understanding of exactly what is important to residents, families, staff, and long-term care communities regarding ADRD care and how to measure changes in outcomes. Interventions may show improved outcomes, but without groundwork to understand meaningful outcomes among these populations and appropriate measures to use, scholars, policymakers, and clinicians are left guessing whether the interventions are targeting the most appropriate outcomes.

Figure 3:

Distribution of studies by location and study design.

Descriptive studies indicate that relationships and social interactions among staff and residents are beneficial to quality of life for residents and job satisfaction among staff. Additionally, facilities with DSCUs reported improved occupancy rates and payer-mix, potentially incentivizing nursing home communities to adopt special care units for residents with dementia. Findings from these and other descriptive studies are limited in generalizability, however, due to the constrained samples and limited geographic representation.

Studies comparing DSCUs with TNHs generally show mixed results – with more studies finding limited differences or worse outcomes among residents in DSCUs compared to residents in TNH settings. However, it is worth considering the limitations of these studies, especially the impact of selection bias on results. As our study confirms, a formal definition of ADRD specialty care remains to be known. Without controlling for resident case-mix, it is unclear whether the comparisons groups are in fact ‘comparable’ or are separate populations.^94,102 Additionally, we are unable to separate the severity of ADRD among residents in DSCUs and TNHs in the included studies, nor can we understand the admission criteria and selection process for care in a DSCU. Further research is needed to understand the process of resident transfers into and out of DSCUs and the care provided to residents at each stage of ADRD.

Limitations

Our systematic review is limited to articles published on PubMed, CINAHL, and PsychINFO. Therefore, other relevant studies from grey literature or other publication databases may not be included in our review. Additionally, our review is constrained to the limitations of the included studies and generalization and meta-analyses were not able to be made. Furthermore, given the unique characteristics of the US health care system compared to international countries, it is unclear whether the conclusions found in our systematic review are generalizable to older adults with ADRD around the world.

V. CONCLUSION

Overall, our systematic review did not find conclusive evidence regarding the beneficial mechanisms of DSCUs and actionable care practice changes in long-term care settings. Studies examining the needs, wants, and concerns of residents, families, staff, and long-term care communities are needed to better understand meaningful outcome measures for these populations. Randomized clinical trials evaluating DSCUs vs. TNH are needed to disentangle the ‘special’ features of DSCUs.

KEY POINTS.

1. Overall, our review did not find conclusive evidence regarding the benefits of DSCUs in long-term care settings among residents with dementia, their families, the staff providing their care, and the facilities.

2. No rigorous study designs were found examining ‘special’ features of DSCUs and associations with outcomes among residents, family, staff, and the facility.

3. Randomized clinical trials are needed to disentangle the ‘special’ features of DSCUs.

APPENDIX

Databases searched: PubMed, CINAHL, PsycINFO

PubMed search strategy:

(Dementia[[MeSH Terms]) OR (alzheimer’s disease[MeSH Terms])) OR (dementia[Title/Abstract])) OR (Alzheimers[Title/Abstract])) OR (“cognitive impairment”[Title/Abstract])) OR (“cognitive impaired”[Title/Abstract])) OR (“cognitively impaired”[Title/Abstract]))

AND

(nursing homes[MeSH Terms]) OR (assisted living facilities[MeSH Terms])) OR (homes for the aged[MeSH Terms])) OR (palliative care[MeSH Terms])) OR (hospice care[MeSH Terms])) OR (terminal care[MeSH Terms])) OR (“long term care”[Title/Abstract])) OR (“nursing homes”[Title/Abstract])) OR (“skilled nursing facility”[Title/Abstract] OR “skilled nursing facilities”[Title/Abstract])) OR (“assisted living facility”[Title/Abstract] OR “assisted living facilities”[Title/Abstract])) OR (“residential care”[Title/Abstract])) OR (“care home”[Title/Abstract])) OR (hospice[Title/Abstract])) OR (“end of life care”[Title/Abstract])) OR (palliative[Title/Abstract])))

AND

(“special care unit”[Title/Abstract]) OR (“dementia specific”[Title/Abstract])) OR (“special care”[Title/Abstract] OR “specialized care”[Title/Abstract])) OR (“memory care”[Title/Abstract])) OR (“quality of life”[Title/Abstract])) OR (“dementia unit”[Title/Abstract])) OR (“memory clinic”[Title/Abstract])) OR (quality of life [MeSH Terms]))