The Therapeutic Design of Environments for People With Dementia: A Review of the Empirical Research

Abstract

Design of the physical environment is increasingly recognized as an important aid in caring for people with dementia. This article reviews the empirical research on design and dementia, including research concerning facility planning (relocation, respite and day care, special care units, group size), research on environmental attributes (noninstitutional character, sensory stimulation, lighting, safety), studies concerning building organization (orientation, outdoor space), and research on specific rooms and activity spaces (bathrooms, toilet rooms, dining rooms, kitchens, and resident rooms). The analysis reveals major themes in research and characterizes strengths and shortcomings in methodology, theoretical conceptualization, and applicability of findings.

Vernon L. Greene, PhD

Design of the physical environment is increasingly recognized as an important aid in the care of people with Alzheimer's disease and other dementias. Facility administrators and designers now view the design of long-term care, assisted living, and other environments as more than simply decorative. Design is regarded as a therapeutic resource to promote well-being and functionality among people with dementia. This article reviews and analyzes findings from empirical research on the therapeutic impacts of design in dementia care settings.

Since the early 1980s, numerous "design guides"— books and articles offering planning, architectural, and interior design recommendations—have been written to instruct architects and care providers on how to enhance safety, homelikeness, and so forth in dementia care facilities. At least four books of design guidance for dementia environments have been published to date (see Brawley 1997; Calkins 1988; Cohen and Day 1993; Cohen and Weisman 1991), along with numerous articles in scholarly and professional books and journals (see A, Note 1). Design recommendations for dementia environments are organized on a continuum by scale, as follows (after Cohen and Weisman 1991): (a) planning principles—broad decisions made when developing a dementia care facility (e.g., facility planning should accommodate a continuum of care); (b) general attributes—desired qualities of the overall environment of the facility (e.g., facility design should promote noninstitutional character); (c) building organization—desired arrangement of spaces within the facility (e.g., building design should support residents' sense of orientation); and (d) specific rooms and activity spaces—the design of particular rooms within the facility (e.g., design of bathrooms should preserve residents' dignity and privacy).

Design guides typically offer "hypotheses" for how the spatial organization and appointment of the physical environment may promote well-being for people with dementia. For example, to minimize the sensory overstimulation that afflicts many people with dementia, design guides recommend modifications such as designation of quiet rooms with soft colors, elimination of unnecessary clutter, and removal of paging systems (cf. Brawley 1997; Cohen and Weisman 1991). Frequently, design guidance is based on the practical experience of designers or facility administrators; other times, design guidance is research based, applying findings from clinical research on dementia in the form of design "solutions" (Weisman, Calkins, and Sloane 1994).

Not all design guidance requires empirical research findings to justify its recommendations. For instance, design guides frequently call for enhanced quality of life in institutional settings (e.g., design strategies to increase homelikeness and autonomy for residents). Such values—essential qualities of dignity, privacy, and so forth—are arguably "inalienable rights" (Lawton 1981, p. 245) that do not require empirical research for validation.

Other design guidance does demand verification through empirical research, however. Empirical research is needed to resolve situations in which conflicting design recommendations are offered. Research is also warranted when recommended design solutions are of unknown effectiveness or when design recommendations have major or controversial impacts for cost or quality of life. Research on design and dementia has been conducted in earnest since at least 1980, yet findings of many studies remain unknown among designers and facility administrators. In the following sections, we review and analyze existing studies of design and well-being for people with dementia to enhance the design of dementia facilities and to provide direction for future research.

Methods

Several strategies were used to identify potential studies for review. The first involved a key-word search of four major databases: Psych Abstracts, Medline, MAGS, and CAT (see A, Note 2). Potential studies were also identified by reviewing all issues (1980 or later) of several journals in gerontology and environment-design research (see A, Note 3). Finally, reference lists were inspected for all studies included in this review. In each case, articles, books, and chapters identified as potentially relevant (by title and by abstract if available) were collected and assessed for appropriateness.

Studies included in this review met the following criteria: a report of empirical research (see A, Note 4), published 1980 or later (see A, Note 5), written in English, with an emphasis on people with dementia or their families or staff caregivers, and with a substantial (though sometimes secondary) emphasis on the relationship between the design of the physical environment and the well-being of people with dementia, their families, and/or staff. For this review, the physical environment was loosely defined as the domain of relevance to architects, interior designers, facility managers, and/or administrators or caregivers undertaking environmental design or renovation. Thus, research on issues such as lighting, furnishings, and outdoor space was included. Research on microscale "product" design (e.g., prosthetic devices to facilitate eating) or on the sensory or social environment outside the purview of designers (e.g., incorporation of music and pets) was excluded, as were studies that did not examine actual environments or actual impacts. Well-being was defined broadly, to include factors such as residents' activities of daily living (ADL), physical well-being, cognitive function, and problem behaviors; family members' well-being; and staff well-being and job performance. Seventy-one research reports were included in the review. Because of this selective search strategy, we may have overlooked some relevant material in the review.

Research Design and Sample Size

Much research on design and dementia comprises small size samples. For example, more than 30% of the studies reviewed used samples of fewer than 30 participants; many included less than 10 participants. Sample sizes reflect the limited populations of residents at the single facility in which many studies were conducted, the high rates of resident mortality, and facilities' limited populations of residents in comparable stages of dementia. Although they raise concern for the validity and generalizability of findings, studies with small samples were included so as not to severely restrict the scope of this review. Research designs and samples are described in Table 1 .

Results

The rate of research on design and dementia is increasing: from 6 research reports from 1981–1985, to 17 research reports from 1986–1990, to 26 research reports from 1991–1995, and to 21 research reports already published since 1996 (see Table 1 ). This section summarizes findings from the research reports reviewed, according to the organizational framework presented earlier (i.e., planning principles, general attributes of the environment, building organization, and specific rooms and activity spaces). The summary is followed by a discussion and analysis of existing research on design and dementia.

Planning Principles

These studies examine broad decisions regarding the development of dementia care settings. Studies examined impacts for well-being following relocation of people with dementia to new environments, use of respite and day care environments and of special care units (SCUs), and exposure to various group sizes of residents.

Relocation to New Environments.

Findings are mixed regarding the impacts of relocating people with dementia to new environments (Robertson, Warrington, and Eagles 1993; Seltzer et al. 1988; see A, Note 6). When moved together as intact units of residents and staff, people with dementia appear to suffer few or no adverse impacts from relocation (Anthony, Procter, Silverman, and Murphy 1987; McAuslane and Sperlinger 1994; Robertson et al. 1993). The more pleasant environment of a new facility may partially explain the lack of negative impact for relocated residents (according to McAuslane and Sperlinger 1994). In contrast, residents with dementia who are moved individually appear to suffer higher rates of depression and mortality following relocation (Anthony et al. 1987; Robertson et al. 1993). This effect holds when residents undergo orientation to ease relocation. Staff members also report decreased job satisfaction (attributed to anxiety) prior to moving, which returns to premove levels of satisfaction following relocation (McAuslane and Sperlinger 1994).

Respite Environments.

Respite environments offer temporary care for people with dementia and provide relief to families. The impacts of respite environments on people with dementia appear to be related to individuals' functional levels and to the type of environment. In an examination of 37 clients of respite services, Seltzer and colleagues 1988 found that, following a 2-week respite stay, lower functioning individuals showed small improvements in ADLs, whereas higher functioning individuals showed a small decline in ADLs. Both effects were minor, and neither group revealed any changes in cognitive status following respite. In a study of 85 people with dementia, use of respite environments for approximately 2 weeks was associated with little deterioration and with improvements in cognitive function and mood (Ulla, Johanna, and Raimo 1998). Improvements were attributed to the therapeutic philosophy and care plan of the SCUs, and to the SCUs' homelike environments. Thus, findings largely support the use of respite as an alternative to home care alone, because negative impacts of respite are limited.

SCUs.

Generally, SCUs are segregated units that accommodate only cognitively impaired individuals, such as those with dementia. SCUs distinguish themselves by offering one or more "special" features, including dementia-appropriate activities, small groups of residents, special staff selection and training, family involvement, and specialized design (see also Berg et al. 1991). According to a survey of 31 SCUs in five states, the most typical, distinguishing environmental features of SCUs (compared to nonspecialized units) include smaller size units, fewer resident rooms, and more designated private rooms (Mathew and Sloane 1991). SCUs are further characterized by the presence of private dining rooms, separate and larger activity rooms, and access to the outdoors (Mathew and Sloane 1991).

The effectiveness of SCUs for people with dementia has been subject to debate. A complete review of this multifaceted research is beyond the scope of this article, in which we focus on the physical environment only (for reviews, see Maslow 1994; Teresi, Grant, Holmes, and Ory 1998; Weisman et al. 1994). General information on the impacts of SCUs warrant mention, however. Studies show associations between SCU environments and improvement or slowed decline in residents' communication skills, self-care skills, social function, mobility, and affective responses (Benson, Cameron, Humbach, Servino, and Gambert 1987; Greene, Asp, and Crane 1985; McCracken and Fitzwater 1989; Skea and Lindesay 1996). Additionally, SCUs are associated with reductions in behavior disturbances, abnormal motor activity, apathy, and hallucinations among residents (Annerstedt 1993; Bellelli et al. 1998; Benson et al. 1987; Greene et al. 1985; McCracken and Fitzwater 1989; Swanson, Maas, and Buckwalter 1993).

Other positive impacts of SCUs reported in these studies include reduced emotional strain among relatives and increased competence and satisfaction among staff (Annerstedt 1993; Wells and Jorm 1987; see A, Note 7). Segregation of dementia residents into special units also appears to benefit residents without cognitive impairments. Cognitively intact residents are found to suffer declines in mental and emotional status when living in close residential proximity to people with dementia (Teresi, Holmes, and Monaco 1993; Wiltzius, Gambert, and Duthie 1981).

Alternately, SCUs are reported to have little or no positive effect on residents' wandering, cognition, functionality, and behavior, or on staff job satisfaction or job pressure (Bellelli et al. 1998; Chafetz 1991; Holmes et al. 1990; Ramirez, Teresi, Holmes, and Fairchild 1998; Saxton, Silverman, Rica, Keane, & Deeley, 1998; Skea and Lindesay 1996; Swanson et al. 1993; Webber, Breuer, and Lindeman 1995). Reports of findings do not distinguish between SCUs with and without special environmental features. Many studies of SCUs include small sample sizes and lack comparison groups (see Table 1 for details.)

It is difficult to assess whether specialized design features in SCUs have any impact on people with dementia. First, SCUs are not comparable, because what is considered an SCU varies enormously (Maslow 1994; Teresi, Holmes, Ramirez, and Kong 1998). Second, most SCUs do not use extensive specialized design features (U.S. Office of Technology Assessment, 1992, in Maslow 1994). Further, special features used in SCUs (staffing, activities, design, etc.) are frequently treated by researchers as one "global" intervention (Weisman et al. 1994; see A, Note 8). Thus, potential impacts from individual design features (private rooms, minimal sensory stimulation, etc.) are obscured by simultaneous modifications in other arenas. When used, design features may not be identified by researchers as highly significant aspects of the special intervention (cf. Skea and Lindesay 1996; Swanson et al. 1993). For these reasons, the impact of specialized design cannot be easily distinguished in much existing research on the effectiveness of SCUs.

Day Care Centers.

Only one study was identified that specifically examined the design of day care centers in terms of therapeutic impacts. In this research, relocation of a day care center to an enhanced facility (including safety and surveillance features, an enclosed garden, and more space for day health programs and activities) was associated with positive and negative changes in staff stress and quality of care (Lyman 1989). Following the move, staff stress shifted from that prompted by space shortages to (lower) stress associated with specific spatial configurations (e.g., difficulty involving clients in activities in new, larger activity areas). Negative impacts on quality of care associated with limited space (e.g., insufficient space for clients to conduct specific activities as long as desired) were also reduced following relocation.

Group Size and Clusters of Residents.

Design guides suggest that units with fewer residents may reduce overstimulation among people with dementia by controlling noise and by limiting the number of people each resident encounters. This recommendation is supported by research findings, including those of a major survey of 53 SCUs in four states (Sloane et al. 1998). According to this study and others, larger unit sizes are associated with higher resident agitation levels and with increased intellectual deterioration and emotional disturbances (Annerstedt 1994; Sloane et al. 1998). Further, residents in larger units exhibit more frequent territorial conflicts, space invasions, and aggressiveness toward other residents (Morgan and Stewart 1998). In contrast, people with dementia residing in smaller units experience less anxiety and depression and more mobility (Annerstedt 1997; Skea and Lindesay 1996). Small group sizes are also positively associated with increased supervision and interaction between staff and residents (McCracken and Fitzwater 1989) and with social interaction and friendship formation among residents (McAllister and Silverman 1999; Moore 1999; Netten 1993). No consistent numbers are offered for what constitutes a "large" or a "small" unit.

Smaller facilities offer additional benefits for residents and staff. In a comparison of 28 residents of group living facilities (see A, Note 9) and 31 residents of traditional nursing homes, residents of group living displayed higher motor functions and slightly improved or maintained ADLs and required less usage of antibiotics and psychotropic drugs (Annerstedt 1993; see A, Note 10). In the same study, relatives with family members in group living units reported lower levels of strain and better attitudes toward dementia care than relatives of residents in nursing homes. Staff members also experienced benefits associated with group living facilities. Staff in group living units reported greater competence, more knowledge in dealing with dementia, and greater job satisfaction than did their counterparts in nursing homes (Annerstedt 1993).

General Attributes of the Environment

These studies investigate desired qualities of the overall facility environment. Studies have examined effects on well-being associated with noninstitutional character, levels of sensory stimulation, lighting levels, and design modifications for safety.

Noninstitutional Character.

Design guides frequently endorse the use of noninstitutional design features, such as homelike furnishings and personalization, to promote well-being among residents. This endorsement is supported by research findings, though studies often compare facilities in which many features vary (e.g., staff training, activity programming), in addition to environmental design. Noninstitutional environments characterized as having homelike or "enhanced" ambiance (personalized rooms, domestic furnishings, natural elements, etc.) are associated with improved intellectual and emotional well-being, enhanced social interaction, reduced agitation, reduced trespassing and exit seeking, greater preference and pleasure, and improved functionality of older adults with dementia and other mental illnesses (Annerstedt 1994; Cohen-Mansfield and Werner 1998; Kihlgren et al. 1992; McAllister and Silverman 1999; Sloane et al. 1998). Compared with those in traditional nursing homes and hospitals, residents in noninstitutional settings are less aggressive, preserve better motor functions, require lower usage of tranquilizing drugs, and have less anxiety. Relatives reported greater satisfaction and less burden associated with noninstitutional facilities (Annerstedt 1997; Cohen-Mansfield and Werner 1998; Kihlgren et al. 1992). Staff also prefer less institutional, enhanced environments (Cohen-Mansfield and Werner 1998).

Noninstitutional environments are not entirely beneficial, however. A higher degree of homelikeness is associated with greater restlessness, more disturbances (tied to greater assertion of independence), and increased disorientation and deterioration of diet (Elmstahl, Annerstedt, and Ahlund 1997; Kihlgren et al. 1992; Wimo, Nelvig, Adolfsson, Mattson, and Sandman 1993). Studies also show that mortality and decline rates for residents do not significantly improve in noninstitutional units when compared with traditional settings (Annerstedt 1994; Phillips, Sloane, Howes, and Koch 1997; Wimo et al. 1993). Further, noninstitutional design requires supportive caregiving to be effective. In an ethnographic study of one facility, "institutional" caregiving practices (characterized as inflexible and formal) were described as undermining the therapeutic potential of the homelike environment (Moore 1999).

Sensory Stimulation.

Residents face difficulties with sensory overstimulation, which may increase the distraction, agitation, and confusion associated with dementia. Sensory overstimulation may be exacerbated by the normal hearing loss that accompanies aging and the further hearing loss associated with dementia, both of which may increase confusion and reduce social interaction and self-esteem (Brawley 1997; see A, Note 11). (Visual deficits, discussed later, further increase overstimulation.) At the same time, sensory deprivation has been identified as a potential problem in many dementia care environments (Cohen and Weisman 1991). Design guides call for appropriate levels of sensory stimulation, striking a careful balance between environmental overstimulation and deprivation. Recommendations include removing unnecessary clutter, providing tactile stimulation in surfaces and wall hangings, and eliminating overstimulation from televisions, alarms, and so forth (cf. Evans 1989; Hall, Kirschling, and Todd 1986).

Researchers have identified characteristics and locations linked with high levels of sensory stimulation in environments for people with dementia. In an ethnographic study of one skilled nursing facility, overstimulation is associated with loud noises (loud talking, singing and clapping, etc.), with crowding and disruptive behavior from other residents, and with frightening experiences (e.g., scary movies, costumes;Nelson 1995). High stimulation—as measured by agitation levels—was found to occur in elevators, corridors, nursing stations, bathing rooms, and other residents' rooms, whereas low stimulation has been observed in activity and dining rooms (Cohen-Mansfield, Werner, and Marx 1990; Negley and Manley 1990). Detailed descriptions of these spaces were not provided by researchers.

Overstimulation may impair residents' ability to concentrate. Limited stimulation activity areas—made by hanging cloth partitions to eliminate views to ongoing activity—reduce distractions among residents by up to two-thirds (Namazi and Johnson 1992b). Use of partitions increased the ability to focus on a task among residents in all stages of dementia by eliminating some visual and especially auditory distractions (e.g., noise, talking).

Findings on the effects of low stimulation units are mixed. Use of a neutral design and color scheme, elimination of stimulation, and consistent daily routines have been shown to reduce behavioral disturbances, curtail use of physical and chemical restraints, and encourage weight gain (Bianchetti, Benvenuti, Ghisla, Frisoni, and Trabucchi 1997; Cleary, Clamon, Price, and Shullaw 1988). Similarly, in one quasi-experiment, 13 residents of an SCU that incorporated structured resident routines and reduced stimulation displayed fewer catastrophic reactions and more positive interactions, compared with nine residents in long-term care (Swanson et al. 1993). Reduced stimulation units have had little effect in regulating sleep patterns, decreasing urinary incontinence, or discouraging wandering, however (Bianchetti et al. 1997; Cleary et al. 1988; Swanson et al. 1993; see A, Note 12).

Design guidance argues that certain levels of sensory stimulation may be required to promote engagement in activities and interaction and to minimize withdrawal among people with dementia (cf. Calkins 1988). The positive impacts of sensory stimulation have received limited research. The experimental Weiss Institute of the Philadelphia Geriatric Center was designed to maximize positive sensory stimulation; this facility featured resident rooms opening directly to a central open space. The spatial configuration was intended to enhance residents' orientation and engagement in activities (Lawton, Fulcomer, and Kleban 1984). Indeed, in a postoccupancy evaluation of the Weiss Institute, residents were found to spend less time in their rooms and were more attentive to activity following relocation to this facility (Lawton et al. 1984). In a related study, a high stimulation environment (including orientation aids, recreational materials, and extensive reality orientation programs) was associated with increased morale among 16 staff members in one unit, compared with morale among 13 staff members in a traditional dementia unit (Jones 1988). The focus on increasing structure and resident orientation in the high stimulation unit suggests other possible explanations for enhanced staff morale in this unit.

Lighting and Visual Contrast.

People with dementia face particular visual deficits, including difficulty with color discrimination, depth perception, and sensitivity to contrast (Cronin-Golumb 1995). These deficits exacerbate normal changes in vision that accompany aging, such as irritation from glare and changes in color perception (Brawley 1997). Design guides for dementia environments recommend strategies to reduce glare, increase contrast where appropriate, and minimize confusion concerning depth perception. Design guides also recommend increasing overall light levels and exposure to bright light (cf. Brawley 1997).

Compared with other older adults, people with dementia are exposed to inadequate levels of bright light (described as light exceeding 2,000 lux; Campbell, Kripke, Gillin, and Hrubovcak 1988). In findings from two studies involving 24 and 10 residents, respectively, bright light treatment consistently regulated circadian rhythms and improved sleep patterns among people with dementia (Mishima et al. 1994; Satlin, Volicer, Ross, Herz, and Campbell 1992; see A, Note 13). Results are mixed concerning the impact of bright light on agitation (Lovell, Ancoli-Israel, and Gevirtz 1995; Mishima et al. 1994; Satlin et al. 1992).

Most often, research on the effects of bright light is conducted under laboratory conditions, requiring special equipment and the restraint of residents. The effects of bright light as a regular environmental feature have received limited attention. One quasi-experimental study was identified in which researchers examined the effect of ceiling-mounted light fixtures that provided high intensity illumination (790–2,190 lux; Van Someren, Kessler, Mirmiran, and Swaab 1997). Bright light administered in this fashion fostered behavioral improvements and increased circadian rest–activity rhythms among 22 people with severe dementia. Residents in facilities with low overall light displayed higher agitation levels (Sloane et al. 1998). Residents in units with inadequate lighting showed no difference in psychiatric symptoms compared with residents in units with ample lighting, however (Elmstahl et al. 1997).

Little research on the impacts of visual contrast in dementia care environments was identified, though this strategy is frequently recommended to enhance "legibility" or clarity of the environment. In one quasi- experiment, 13 residents with dementia ate more and displayed less agitation when dining arrangements incorporated brighter light and heightened color contrast (i.e., high contrast tablecloths, place mats, dishes; Koss and Gilmore 1998).

Safety.

Residents' attempts to leave facilities or homes present a major safety concern for staff and family caregivers. Design solutions to prevent unwanted exiting often do so by exploiting residents' cognitive deficits. For instance, in a study involving nine residents of a psychogeriatric ward, a full length mirror placed in front of the exit door reduced residents' exit attempts by half (Mayer and Darby 1991). A reverse mirror had a similar, but less significant effect. Impacts were attributed to residents' loss of memory of personal identities; accordingly, residents may have been distracted from exiting when engaged or frightened by the image of an approaching "stranger" in a mirror (Mayer and Darby 1991).

Another design strategy capitalized on the likelihood that, because of problems with depth perception, people with dementia may interpret two-dimensional patterns on the floor as three-dimensional barriers. In a quasi-experiment with eight residents, such two-dimensional grids successfully eliminated most exit attempts for some residents (Hussian and Brown 1987). In other studies, two-dimensional grids either increased or failed to decrease residents' exit attempts (Chafetz 1990; Namazi, Rosner, and Calkins 1989). Failure to reduce exiting was attributed to the presence of glass doors and adjacent large windows, which offered views to attractive, nearby outdoor spaces (see also Morgan and Stewart 1999). Attractive views were hypothesized to distract residents from two-dimensional grids or to entice residents to overcome their aversion to these optical illusions. In a study involving seven SCU residents, installation of closed, matching miniblinds that restricted light and views through exit door windows decreased exit attempts by half (Dickinson, McLain-Kark, and Marshall-Baker 1995).

Other design strategies also created optical illusions that reduced unwanted exiting. The addition of a cloth panel to camouflage a door knob or "panic bar" eliminated exit attempts for most residents (Dickinson et al. 1995; Namazi et al. 1989). (Both of these studies used fewer than 10 residents.) This effect held irrespective of the color or pattern of the cloth cover (Namazi et al. 1989) and with and without the use of miniblinds to cover windows (Dickinson et al. 1995). Disguising the door knob itself (with a knob cover or by painting the knob to blend with the door) reduced exit attempts to a lesser extent (Dickinson et al. 1995; Namazi et al. 1989).

Finally, conditioning residents to respond to attention-getting signage also reduced exit attempts. Three residents with dementia who were conditioned to develop negative associations with "supernormal" stimuli—in this case, large, colored, cardboard geometric shapes placed near exits (Hussian 1982–83; see A, Note 14)—wandered less into doors and stairways bearing those images.

Accommodating residents' exit attempts, rather than discouraging them, also generated positive outcomes. Unlocking doors to allow access into secure outdoor areas was associated with significant decreases in agitation in a quasi-experiment involving 12 residents (Namazi and Johnson 1992d). Reduced agitation was tied to increased autonomy as well as to outdoor usage.

Surveillance is considered essential by staff for maintaining safety in environments for people with dementia (Morgan and Stewart 1999). Design interventions may have unintended consequences for staff surveillance opportunities. In interviews with nine staff members and nine relatives associated with a newly designed SCU, staff reported that the new facility's low density, private resident rooms, enclosed charting spaces, and secluded outdoor area and activity spaces impeded staff surveillance and increased time spent locating and monitoring residents (Morgan and Stewart 1999). Ease of surveillance also has negative consequences. In an evaluation of the Weiss Institute, staff interaction with residents was found to decrease following occupation of this new facility (Lawton et al. 1984; see also Liebowitz, Lawton, and Waldman 1979). Because the facility's open design accommodated staff surveillance from the nurses' station, direct staff contact with residents may have been minimized.

Preventing falls among residents is another key safety concern (cf. Morgan and Stewart 1999; Pynoos and Ohta 1991; Scandura 1995). Design interventions have demonstrated some success in reducing residents' falls. A significant reduction in falls was reported in one SCU with the introduction of alternative furnishings that put residents closer to the ground (i.e., bean bag chairs, futons, and mattresses placed on the floor; Scandura 1995; see A, Note 15). In other research, environmental modifications introduced into home environments to reduce falls were judged effective by 12 dementia caregivers at a 7-month follow-up (Pynoos and Ohta 1991). These modifications included tub and stair rails, a nonskid bath mat, and a bath chair.

Building Organization

Studies of building organization examine the desirable arrangement of spaces within facilities. Issues investigated include residents' orientation and wayfinding, and the impact of providing outdoor spaces in dementia care facilities.

Orientation.

Disorientation—confusion regarding place, time, personal identity, or social situation—is common among people with dementia (Cohen and Weisman 1991). Design guides suggest numerous strategies to enhance orientation, including improvements for wayfinding (e.g., landmarks, signage) and provision of information from the environment (e.g., allowing views to accessible outdoor areas to increase residents' orientation to time of day and season).

Research confirms that residents' orientation depends, in part, on the physical environment. In a study of 79 dementia residents at 13 long-term care facilities, higher levels of orientation were associated with quiet environments (Netten 1993). Researchers theorized that disorientation followed residents' attempts to "shut out" noisy environments. Not surprisingly, wayfinding among residents was judged less successful in facilities with low lighting levels in public areas (Netten 1989).

Design strategies intended to enhance orientation appeared to aid at least some residents. Staff members reported that orientation among residents was supported by design modifications that included room numbers and use of distinguishing colors for resident rooms and doors (Lawton et al. 1984). In studies with eight residents, large signs improved resident orientation, when incorporated with orientation training (Hanley 1981); signs alone had minimal effect on residents' orientation, however.

The type of orientation device may make a difference, though research on this question is limited to one experiment involving 10 SCU residents. When displayed in cases outside resident rooms, personally significant memorabilia were somewhat more likely to help residents find their rooms than were displays without personal significance (Namazi, Rosner, and Rechlin 1991). Personally significant memorabilia were most useful for those with moderate dementia; higher functioning residents were able to orient with nonsignificant memorabilia as well, and lower functioning residents were aided by neither.

Orientation is further impacted by building configuration. Simple building configuration is associated with resident orientation, when residents are also provided with explicit environmental information (Passini, Rainville, Marchand, and Joanette 1998). In a quasi-experiment with 105 residents in several group living facilities, residents were found to experience greater spatial orientation in facilities designed around L-, H-, or square-shaped corridors, compared with facilities with corridor designs (Elmstahl et al. 1997). Corridor designs were also associated with higher degrees of restlessness and dyspraxia and with reduced vitality and identity (Elmstahl et al. 1997). Residents in facilities with more hallway space demonstrated less disorientation and less lack of vitality (Elmstahl et al. 1997).

In survey research with 104 residents in several homes, higher levels of orientation were identified in "cluster" facilities (comprised of small units of resident rooms and associated common spaces), compared with larger scale "communal" facilities (common spaces separated from resident rooms and shared by larger groups of residents; Netten 1989). In cluster facilities, higher levels of orientation were associated with complex decision points and longer corridors, which allowed meaningful choices between places residents used (Netten 1989). In communal facilities, heightened orientation was associated with short corridors and simple decision points, which allowed residents to travel only short distances without prompts and did not force residents to choose between spaces they did not use (Netten 1989).

Provision of Outdoor Areas.

Design guides recommend access to the outdoors to maintain homelikeness, to accommodate activities, and to increase residents' exposure to light and sun. Limited research findings support the value of outdoor spaces to reduce aggression among people with dementia. In a longitudinal study of five facilities with and without outdoor spaces, researchers found that violent episodes among residents decreased over time in facilities with outdoor environments, whereas violent episodes increased during the same time period in facilities without outdoor environments (Mooney and Nicell 1992). Residents walked outdoors more often (for short periods of time) in a facility with a special therapeutic garden (Mooney and Nicell 1992).

Specific Rooms and Activity Spaces

This research investigates the design of particular rooms within the facility. Studies examine the design of bathrooms, toilet rooms, dining rooms, kitchens, and resident rooms, as these impact well-being among people with dementia and others.

Bathrooms.

For people with dementia, bathing is an experience that frequently compromises dignity and autonomy. Design recommendations emphasize increasing independence and control in bathing (e.g., choice of shower or tub bath), promoting a more homelike bathing experience (e.g., less institutional design), and assisting caregivers during bathing (additional space, grab bars, etc.).

Bathing is regarded as among the most stressful tasks in caring for people with dementia (Kovach and Meyer-Arnold 1996; Pynoos and Ohta 1991; Sloane et al. 1995). Several studies examine aspects of bathing associated with high stress. Negative resident reactions are associated with unfamiliar or fearful equipment or procedures (bath tub lifts, specialized tubs, getting in and out of the water, high water levels in whirlpool baths), cold tub rooms (cold air or water temperature, chills from slow tub filling or draining), design features that impede bathing (poor lighting, inadequate mats or handrails), and distractions (noisy equipment, running water, or distracting activities outside the bathroom; Kovach and Meyer-Arnold 1996; Lawton et al. 1984; Namazi and Johnson 1996; Sloane et al. 1995). Some evidence suggests that baths may be less upsetting than showers for residents, though findings are mixed (Kovach and Meyer-Arnold 1996).

Perhaps because of their long-term positive association, natural elements had a calming effect when introduced during bathing in an experiment with 31 residents in five nursing homes (Whall et al. 1997). Nature sounds (e.g., animal and water noises) and pictures (e.g., birds), when provided along with favorite foods and distracting conversation, significantly decreased agitation during shower baths among residents with late stage dementia (Whall et al. 1997).

Toilet Rooms.

Incontinence is a major problem among people with dementia (Namazi and Johnson 1991b). Design guides emphasize the importance of maintaining independence in toileting whenever possible, such as by making toilets easy to locate and to identify (signage, visible locations, etc.). In some instances, the design of toilet rooms may exacerbate toileting problems. Staff report that small toilet rooms make assisting with toileting difficult and that wheelchair users are more likely to have "accidents" when the toilet room is occupied, preventing access (Hutchinson, Leger-Krall, and Wilson 1996).

Research findings, though limited, support the effectiveness of design interventions to facilitate toileting. One quasi-experiment involving 44 residents in two SCUs compared residents' responses to various forms of directional signage for toilet rooms, including the word "rest-room," "toilet," or a graphic of a familiar household toilet (Namazi and Johnson 1991b). Early and moderate stage dementia residents were most likely to locate and use public toilets in response to primary color signage affixed to the floor (responding to residents' typically downcast gaze) comprising a series of arrows and the word "toilet" (Namazi and Johnson 1991b). Further, frequency of toilet use increased dramatically when toilets were visibly accessible to residents (Namazi and Johnson 1991a), though this experiment included only 14 residents. Residents' use of toilets increased by over 800% when curtains surrounding toilets (in lieu of doors) were left open, making public and private toilets clearly visible when not in use (Namazi and Johnson 1991a). In particular, visibility increased toilet use among residents with more advanced dementia.

Dining Rooms and Kitchens.

Design guides offer many recommendations regarding dining and kitchen areas (cf. Calkins 1988; Cohen and Weisman 1991). Suggestions emphasize the importance of a familiar and normal dining experience, the need to locate dining and kitchen activity areas within each dementia unit or "household," and the value of reducing sensory stimulation to encourage eating. Research findings from an experiment with 22 residents support noninstitutional dining arrangements. Noninstitutional dining—in which residents dined "family style" at small dining tables in a coffee room, instead of from trays while seated in chairs in the corridor—was linked to increased social interaction and communication during dining and to improved eating behavior among residents (Gotestam and Melin 1987; Melin and Gotestam 1981). Institutional staff practices (e.g., assigned seating, institutional food service) provoked disruption and agitation in dining rooms with homelike design features (Moore 1999).

In an impact not anticipated by design guidance, relocating dining to the dementia unit of an SCU—from a remote, centralized dining room—significantly decreased residents' aggression (Negley and Manley 1990). Assaults were reduced by over 40% when residents were no longer crowded into elevators to reach the centralized dining room (Negley and Manley 1990). (Elevators had been sites of frequent violations of personal space, which caused altercations.) In this instance, assaults may have been further reduced by designating two dining areas on the dementia unit, thus separating higher functioning residents, more likely to be assailants, from lower functioning residents, more likely to be assault victims. In the same quasi-experiment, staff reported less anxiety and more time for assisting residents after moving dining to the dementia unit.

In a study on the design of environments to encourage independent snacking, installation in kitchenettes of small, accessible refrigerators stocked with snacks prompted only a minimal increase in residents' independent snacking (Namazi and Johnson 1992c). Transparent refrigerators, in which residents could clearly see snacks inside, were only slightly more effective than were conventional, dormitory-style refrigerators. Both styles of refrigerators may have been unfamiliar to residents. Staff provision of snacks was suggested as a possible impediment to residents' independent snacking (Namazi and Johnson 1992c).

Residents' Rooms.

Design guidance emphasizes the need for homelikeness, autonomy, and privacy in residents' rooms in dementia care facilities. The relative merits of private versus shared resident rooms is a matter of debate (Cohen and Day 1993); existing research provides limited guidance on this issue. Comparisons of facilities with and without private rooms typically incorporate other architectural and programmatic differences as well, thus obscuring the significance of resident room type (cf. Annerstedt 1994, Annerstedt 1997; Skea and Lindesay 1996).

Lawton and colleagues 1970 presented findings from a quasi-experiment involving 15 residents, which suggest that number of residents and room design may affect levels of social interaction. This study evaluated the renovation of a long-term care unit—from two institutional-looking group rooms (four and five residents, respectively), to six, less institutional-looking single rooms clustered around a common space. Following renovation, residents were found to spend comparatively less time in their rooms and more time in motion and to engage in less interaction, compared with residents before the renovation. Reduced interaction may reflect greater choice over interaction in private versus group rooms (Lawton et al. 1970).

Closet design was successfully used to enhance residents' independence in dressing. In a quasi-experiment with eight SCU residents, specially designed clothes closets were found to increase autonomy in dressing for those with middle stage dementia (Namazi and Johnson 1992a). By presenting preselected clothing in an appropriate sequential order (undergarments first, followed by blouse, pants, etc.), modified closets reduced staff members' physical assistance in dressing and enhanced residents' independence.

Discussion and Conclusions

From the research reviewed, four primary types of studies on design and dementia emerge. Studies are grouped according to their major focus (people/behavior vs. the physical environment) and their conceptualization of the physical environment (global or discrete). Environmental comparison studies compare two or more facility types for impacts on residents, staff, and family (e.g., SCUs vs. skilled nursing facilities.) Design feature studies assess the effects of specific environmental interventions (e.g., door modifications to reduce exit attempts). Studies of environmental services and policies examine organizational decisions and policies for dementia care environments (e.g., impacts of relocating residents to new environments). Studies of problem behaviors investigate resident conduct that creates difficulties in caregiving (e.g., stressful aspects of bathing). One study (Netten 1993) fit more than one type (see Table 2 ). Of these types, studies of design features (26 studies) and of environmental comparisons (24 studies) predominate (compared with 15 studies of problem behaviors and 7 studies of environmental services and policies). Studies should be evaluated according to the type of research they represent. For example, findings from studies of environmental comparisons should indicate which type of environment is preferred and why.

The following sections analyze findings from existing studies with respect to their implications for application and future research.

Recommendations to Enhance Applicability of Findings

The focus of this article on design application demands some recommendations (though tentative) concerning the therapeutic design of environments for people with dementia. On the basis of existing research findings, dementia care environments should consider the suggestions presented in B among others.

Application of findings is often impeded by studies' research design and/or methods. Confidence in findings is impaired by the frequent use of small samples and the absence of comparison groups. Additionally, many studies use nonequivalent comparison groups (e.g., residents in varying or unspecified stages of dementia, or residents with and without dementia who vary in other characteristics, such as mobility.) Studies do not always adjust reports of findings to account for baseline differences in severity of cognitive, behavioral, or physical deficits. Of the 71 studies we reviewed, only 45 made reference to the residents' stage of dementia at baseline. Further, the interrelations between design interventions has been largely overlooked. For example, SCUs often encompass multiple interventions (smaller unit size, homelike design, low level simulation, etc.). In evaluating the impact of SCUs, studies should consider both which design features are most essential and how various design features work together or detract from each other. These issues must be addressed to improve the validity and generalizability of future research findings.

Applicability of findings would also be enhanced by incorporating explicit hypotheses on the proposed relationships between physical environments and well-being to explain why design features are or are not successful. Finally, the applicability of future studies could be improved by thoroughly describing, in research reports, the physical context of the dementia environment and specific environmental modifications tied to well-being. These last two qualities are exemplified in research conducted by Namazi and colleagues at the Corrine Dolan Center, in Chardon, Ohio (see A, Note 16).

Recommendations for Future Research

Findings from existing studies substantiate the need for more attention to the therapeutic use of design in dementia. Further research should be designed to confirm findings from existing research, especially from small or exploratory studies. Future research should also support the call for therapeutic design of dementia environments and should elucidate the particular characteristics of effective design interventions.

Focus on Multiple Populations and Diverse Environments.

Of the 71 studies reviewed, 12 clearly addressed staff well-being or job performance as outcome measures of the design of dementia environments; only 7 studies investigated outcomes concerning family members' well-being or satisfaction. With greater focus on impacts for staff and relatives, research findings could provide a persuasive rationale for design interventions that might otherwise be neglected. For example, in addition to impacts for resident well-being, research should examine the impacts of noninstitutional design on staff morale and retention and on family visitation and satisfaction with care (cf. Chapman and Carder 1998; Hoglund, DiMotta, Ledewitz, and Saxton 1994; Regnier 1997). Improving staff and family well-being may also enhance caregiving.

In addition, studies should evaluate effective strategies for the therapeutic design of environments other than long-term care and SCUs. Environmental alternatives such as day care and assisted living often have resident populations, care practices and philosophies, physical environments, and regulatory realities that differ dramatically from the more "institutional" options that have been the focus of much existing research. Such environmental alternatives may present new opportunities and new challenges for therapeutic design interventions.

Target Research and Application to Stage of Dementia.

Research findings on the effects of design interventions reveal important differences in response according to residents' level of cognitive and behavioral function (see also Columbo, Vitali, Molla, Gioia, and Milani 1998; Mirmiran, Van Gool, Van Haaren, and Polak 1986). For example, interventions targeted to people in early or middle stages of dementia (e.g., closet design to promote independence in dressing; Namazi and Johnson 1992a) may prove useless for residents in more advanced stages, and vice versa. In developing research questions, researchers should carefully consider the stages of dementia during which design interventions are hypothesized to be of value (see A, Note 17). When possible, studies should include participants in different stages of dementia, and research reports should specify the stage of dementia for research participants.

Focus on Quality of Life, as Well as Problem Behaviors.

In the studies reviewed, impacts on problem behaviors were the most common outcome measure used (followed by impacts on resident ADLs, cognitive function, and social function). Because problem behaviors generate much caregiver burden, caregivers and administrators may especially appreciate this information. The emphasis on problem behaviors may also indicate, however, that many researchers and administrators do not fully appreciate the potential of environmental design to improve quality of life, beyond simply minimizing undesirable conduct. For greatest impact, design professionals and researchers must continue to educate administrators and families on the potential role of environmental design for improving quality of life in a comprehensive way. These recommendations, if implemented, will ensure continued progress in the study and design of therapeutic environments for people with dementia.

Notes

1. A small sample of design guidance includes Coons 1987, Hiatt 1987, Hyde 1989, Lawton 1979, Pynoos, Cohen, and Lucas 1988, Regnier 1997, and Zeisel, Hyde, and Levkoff 1994.

2. MAGS is the magazine and journal article database of over 1,500 scholarly and popular journals. CAT is the catalog of the holdings of the entire University of California library system. Search terms included "dementia" or "Alzheimer's" and the following: home, nursing home, special care unit, SCU, day care, assisted living, design, environment, safety, dignity, homelike, independence, security, wandering, activities, toilet, incontinence, kitchen, dining, resident room, privacy, bathing, continuum of care, aging in place, non-institutional, and barrier. Searches generating more than 400 references were discarded as overly broad.

3. Journals reviewed systematically included Ageing and Society, Environment and Behavior, The Gerontologist, Journal of Architectural and Planning Research, Journal of Environmental Psychology, Journal of Gerontology: Psychological Sciences, Journal of Gerontology: Social Sciences, and the Journal of Mental Health and Aging.

4. Case studies and nonpeer-reviewed work, such as dissertations, were excluded, as were articles that reported only design guidance, or that had only minor reports of research, or incomplete descriptions of research design.

5. This period was identified as the time of greatest productivity in research on dementia and design. An earlier article by Lawton and colleagues 1970 was also included. This article evaluates the impact of a renovation to the Weiss Institute of the Philadelphia Geriatric Center, a premier, experimentally designed dementia care facility.

6. See Borup 1983 for a review of the extensive research literature on the effects of relocation on older adults in institutional settings.

7. Wells and Jorm 1987 examine the use of an SCU for respite versus permanent care. The authors found beneficial impacts for family members associated with use of this SCU for respite care.

8. "Global" studies of SCUs with special environmental features include Annerstedt 1993, Bellelli et al. 1998, Benson et al. 1987, Chafetz 1991, Greene, Asp, and Crane 1985, Holmes et al. 1990, McCracken and Fitzwater 1989, Phillips et al. 1997, Skea and Lindesay 1996, Swanson, Maas, and Buckwalter 1993, Webber, Breuer, and Lindeman 1995. Table 1 describes only those studies of SCUs that specifically note environmental features.

9. Swedish group living facilities compare to both group homes and SCUs in the United States. These residences for 8–10 people are tailored—in design and in care plan—to the needs of people with dementia. Emphasis is placed on involving families and on making care more affordable than institutional alternatives (Annerstedt 1993).

10. Group living units in this study differ from traditional nursing homes in that group living units use noninstitutional design features and specialized dementia caregiving, in addition to small group size.

11. Evidence suggests that the prevalence and severity of hearing loss is greater among people with dementia; however, the cause and effect relationship between dementia and hearing loss is not well understood (cf. Gates et al. 1995; Rapcsak, Kentros, and Rubens 1989; Uhlman, Larson, and Koepsell 1986; Weinstein and Amsel 1986).

12. Similar changes in other facilities are associated with comparable results (cf. Hall, Kirschling, and Todd 1986).

13. Circadian rhythms refer to daily activity cycles based on 24-hr patterns.

14. Exaggerated, simple stimuli were hypothesized to be most effective, since subtle or complex stimuli may be difficult for people with dementia to comprehend (Hussian 1982–83). Conditioning occurred by reinforcing positive associations, such as a favorite food, with one shape, and negative associations, such as loud clapping, with another.

15. Changes in furnishings were accompanied by changes in care plans to reduce demanding tasks (e.g., bathing) in the evenings, when most falls occurred. Full research methods are not reported for this study.

16. Most of these studies are reported in a series published in 1991–92 in the American Journal of Alzheimer's Care and Related Disorders and Research.

17. Calkins 1997 provides an excellent example of stage-appropriate design guidance in her recommendations for design strategies to enhance care for people with late stage dementia.

Recommendations for the Therapeutic Design and Planning of Dementia Environments

• Incorporate small size units.

• Separate noncognitively impaired residents from people with dementia.

• Offer respite care as a complement to home care.

• Relocate residents, when necessary, in intact units rather than individually.

• Incorporate noninstutional design throughout the facility and in dining rooms in particular.

• Moderate levels of environmental stimulation.

• Incorporate higher light levels, in general, and exposure to bright light, in particular.

• Use covers over panic bars and door knobs to reduce unwanted exiting.

• Incorporate outdoor areas with therapeutic design features.

• Consider making toilets more visible to potentially reduce incontinence.

• Eliminate environmental factors that increase stress in bathing.

References