Introduction
In many parts of the world the provision of psychogeriatric inpatient units (PGUs) is limited (O'Connor and Melding, Reference O'Connor and Melding2006; Snowdon, Reference Snowdon2007). More units will be required over coming decades given the rapid aging of the population. This paper offers guidance, based on the existing literature, for those contributing to the design and commissioning of PGUs. In this review PGUs are defined as acute hospital wards/units which provide inpatient care for older people with psychiatric illnesses including psychosis and mood disorders as well as ambulant patients with challenging behaviors in dementia who require acute inpatient stabilization. Psychogeriatric long-stay and step-down wards are excluded from this definition although some of what is written about acute PGU design may apply to those settings. There is limited literature concerning the design of acute PGUs. Most of this literature is based on opinion and anecdote rather than firm evidence. This review therefore identifies gaps in the existing literature.
Methods
A search of the peer-reviewed literature was undertaken utilizing Medline (1950–2010), psycINFO (1806–2009), EMBASE (1980–2009) and CINAHL (1982–2009). Search terms included psychiatric, psychiatry, units, department, hospitals, ward, inpatient, environment, milieu, green house program, design, build, plan, construction, aged, old age, geriatric, psychogeriatric, violence, psychiatric intensive care, high dependency care, staff and staffing. References were also drawn from those cited in key articles. Articles were selected on the basis of relevance of abstracts or titles. Data were obtained from the literature pertaining to general adult inpatient psychiatry units where there was limited information from studies of units designed for older people. Selected non-peer reviewed literature, such as government reports and unpublished academic dissertations, was also reviewed. A systematic review of the literature (Higgins and Green, Reference Higgins and Green2008) was impossible because of the diversity of the literature pertaining to PGUs and to their design. Randomized controlled studies comparing different designs and assessing outcomes were virtually non-existent. Over 200 papers were reviewed and 130 were included.
Historical issues
Over the past 50 years concerns have been voiced about the design and environments of psychiatric wards. Osmond (Reference Osmond1957) urged planners to assess the clinical impact of psychiatric ward design in order to improve future designs. Goffman (Reference Goffman1961) raised concerns about the deleterious effects of the hospital environment on psychiatric patients and Stainbrook (Reference Stainbrook1966) maintained that the environment gives cues to the patient on how to behave: for example, that bare walls and crude furniture convey to the patients the staff expectations concerning regression and loss of control which patients may then feel obliged to fulfill. Prior to the 1990s PGUs in mental hospitals generally provided long-stay care but failed to meet patients' complex social, psychological and medical needs, and many older people with mental disorders were being neglected in the long-stay wards of institutions (Human Rights and Equal Opportunities Commission, 1993; Snowdon and Arie, Reference Snowdon, Arie, Draper, Melding and Brodaty2005). An inquiry in 1993 reported that many Australian public psychiatric facilities were gravely inadequate and often antiquated (Human Rights and Equal Opportunities Commission, 1993). Patients reported that psychiatric hospitals were alienating environments and new psychiatric units were often “poorly designed and cramped” (Human Rights and Equal Opportunities Commission, 1993, p. 277). Specific criticisms in the same report with respect to PGUs included institutional settings being depressing and intimidating for patients and their families, unsatisfactory dormitory-style accommodation, a lack of privacy, inadequate toilet and bathing facilities, and isolated facilities with limited access to public transport (p. 514). Such shortcomings were not fully addressed. An inquiry carried out in 2008 identified poor design, unsuitable patient mix and inappropriate environments lacking in privacy and security as deficiencies of acute inpatient psychiatric care in New South Wales (NSW), Australia (Garling, Reference Garling2008). A recent U.K. report expressed concerns about mental health wards being dirty, dangerous places with often more patients than beds (Mental Health Act Commission, 2008). The concept of positive health effects being related to space and the environment is not new and was espoused by the ancients. Osmond (Reference Osmond1957, p. 23) stated that “structure will determine function unless function determines structure.” Progress in this aspect of psychiatric care has been slow.
Problems in existing psychogeriatric units
Psychiatric facilities should enhance patients' dignity, self esteem and comfort as well as promoting autonomy and ensuring safety (New Zealand Ministry of Health, 2002; Hoskins and Marshall, Reference Hoskins, Marshall, Jacoby and Oppenheimer2002). Several problems in existing PGUs inhibit achieving these goals. It is clear from the literature that both design factors and deficiencies in care contribute to these problems but their relative contributions have yet to be determined. In this section we refer to the literature that best examines whether design factors have an adverse clinical impact.
Violence on PGUs
Violence is a significant issue for psychogeriatric inpatients. Shah (Reference Shah1995) found that 18% of patients with dementia were violent during their PGU admission. Stevenson et al. (Reference Stevenson, Khan and Perumal2009) found 43–80% of patients in a psychiatric intensive care unit for older adults had violent incidents and all had behavioral disturbance. Other studies have found violence to be a significant problem in related psychogeriatric settings (e.g. nursing homes) (Gjerberg, Reference Gjerberg1995; Kovach and Robinson, Reference Kovach and Robinson1996). There is evidence that rates of violence are high, if not higher, among older psychiatric inpatients compared to younger patients (Owen et al., Reference Owen, Tarantello, Jones and Tennant1998a; Reference Owen, Tarantello, Jones and Tennant1998b; Steinert et al., Reference Steinert, Wiebe and Gebhardt1999). There are significant safety concerns in general adult psychiatric units which admit both younger and older patients. Studies in these settings have found prevalence rates of violence of 8–31% and of aggression or potentially harmful experiences of 35–65% (Steinert et al., Reference Steinert, Wiebe and Gebhardt1999; Bowers et al., Reference Bowers2002; Frueh et al., Reference Frueh2005; Ketelsen et al., Reference Ketelsen, Zechert, Driessen and Schulz2007). Others report 2.54–13 assaults per bed yearly in psychiatric wards, compared with 0.37 assaults annually in non-psychiatric beds (Reid et al., Reference Reid, Bollinger and Edwards1985). Violence against patients and staff on psychiatric wards is increasing (Rippon, Reference Rippon2000; Kumar and Ng, Reference Kumar and Ng2001).
The specific factors relating to the prevalence of violence in PGUs have not been well studied. Studies of older patients awaiting respite admission and patients in both general psychiatric settings and nursing homes have found violence to be related to high social density, lack of privacy and loss of control (Shrivastava et al., Reference Shrivastava, Kumar and Jacobson1999; Kumar and Ng, Reference Kumar and Ng2001) as well as cognitive impairment, male gender, increasing age, involuntary status, and crowding (Owen et al., Reference Owen, Tarantello, Jones and Tennant1998a; Nijman and Rector, Reference Nijman and Rector1999; Neville and Byrne, Reference Neville and Byrne2006). Number of hospitalizations and alcohol abuse in men are also factors associated with aggression in general inpatient psychiatric settings (Steinert et al., Reference Steinert, Wiebe and Gebhardt1999; Ketelsen et al., Reference Ketelsen, Zechert, Driessen and Schulz2007). Female inpatients in mixed-gender psychiatric settings are vulnerable to threats, harassment and abuse by male patients (Barlow and Wolfson, Reference Barlow and Wolfson1997; Mezey et al., Reference Mezey, Hassell and Bartlett2005). Pollitt and O'Connor (Reference Pollitt and O'Connor2007) reported that 74% of psychogeriatric inpatients who experienced extreme fear and distress were women. PGUs admitting a higher proportion of mood disorder patients compared to those with dementia appear to have lower rates of violence (Anderson et al., Reference Anderson, Schultz, Buckwalter and Schneider2003).
Violence directed at staff by patients
A study of 6300 randomly selected nurses found that the highest risk of violence to staff was working in a nursing home/long-term care facility, followed by intensive care units, psychiatric units and emergency departments and that the risk of physical assault was highest when working with geriatric patients (Gerberich et al., Reference Gerberich2004). Data on the frequency of patients assaulting staff on PGUs are scant but Gerberich et al. (Reference Gerberich2004) suggest that there is a higher incidence of staff being assaulted when PGUs admit patients with challenging behaviors and dementia. Soares et al. (Reference Soares, Lawoko and Nolan2000) found that 70% of mental health nurses recalled being physically assaulted by a patient at work. Those who worked in noisy conditions with poor lighting and ventilation were more likely to be assaulted than those who worked in a more favorable environment (Soares et al., Reference Soares, Lawoko and Nolan2000). Physical assaults on mental health care workers by patients were the leading cause of occupational injuries to staff working in one U.S. community psychiatric hospital (Calabro and Baraniuk, Reference Calabro and Baraniuk2003). Staff perceived that their work environment rated more significantly than training, safety practices or occupational stress in contributing to violence risk (Calabro and Baraniuk, Reference Calabro and Baraniuk2003). Improving the work environment was rated by staff as the most important safety change needed.
Trauma related to psychogeriatric hospitalization
Trauma may result from patients' perception that they are at risk of verbal abuse or physical violence as well as experiencing or witnessing violence/abuse. Pollitt and O'Connor (Reference Pollitt and O'Connor2007) investigated the extent to which 50 older people (age range 66–91 years) with severe depression experienced trauma during a PGU admission. Almost all had spent time in wards with very disturbed and aggressive patients. Thirty-four percent of the patients interviewed post discharge reported concerns including: having felt frightened, unsafe, a sense of intrusion and lack of privacy, disgust and embarrassment, powerlessness, and feeling undermined at some time during their hospital stay. In 24% these experiences could be described as “traumatic”. Other concerns included loneliness, misplacement, being restricted in movements, and reluctance to have visitors. Sharing space with patients who were “disinhibited, intrusive, noisy, and aggressive and who lacked the capacity to communicate normally”, compounded respondents' sense of loneliness and isolation (Pollitt and O'Connor, Reference Pollitt and O'Connor2007, p. 121). Concerns related to fellow patients with dementia featured prominently. It is unclear from the study how much of the patients' perception of dangerousness was related to their own illness. The impact of different ward environments on patients' experiences was not systematically investigated. The authors suggested that while none met criteria for post-traumatic stress disorder (PTSD) there was a possibility that the most traumatized declined to participate.
Sequelae of traumatic experiences
There is evidence that general adult psychiatric patients who are traumatized by being hospitalized may experience lasting harm (Pollitt and O'Connor, Reference Pollitt and O'Connor2007). The sequelae of traumatic experiences in PGUs such as those described by Pollitt and O'Connor (Reference Pollitt and O'Connor2007) are not well studied. A small study by Wiltzius et al. (Reference Wiltzius, Gamberg and Duthie1981) showed that cognitively competent residents in a 38-bed nursing facility who had been randomly placed with a mentally confused room-mate became significantly more forgetful, depressed, lonely, anxious and irritable. Robins et al. (Reference Robins, Sauvageot, Cusack, Suffoletta-Maierle and Frueh2005) hypothesize that “sanctuary harm” may lead to reduced self-esteem and sense of self-worth, exacerbation of psychiatric symptoms and reduced participation in psychiatric care. Traumatic and harmful events may result in long-lasting emotional distress and distrust of the formal treatment system (Robins et al., Reference Robins, Sauvageot, Cusack, Suffoletta-Maierle and Frueh2005). Other possible sequelae for older patients – including PTSD, delayed recovery from psychiatric illness, functional impairment, avoidance of future readmission and physical injuries – deserve investigation.
Adverse impact of ward milieu on patients and staff
Ward milieu is thought to influence patient outcomes and behavior (Davis et al., Reference Davis, Glick and Rosow1979) as well as staff morale (Minde et al., Reference Minde, Haynes and Rodenburg1990). The twenty-first-century concept of a therapeutic milieu or “optimal healing environment” focuses on patient-centered care, continuous healing relationships, safety as a system priority and cooperation among clinicians within a holistic practice atmosphere (Mahoney et al., Reference Mahoney, Palyo, Napier and Giordano2009). Of note, at least one study suggests that a therapeutic milieu does not exist in contemporary acute inpatient psychiatric care (Shattell et al., Reference Shattell, Andes and Thomas2007).
In this section we limit our review to literature which examines the impact of the physical environment on ward milieu. Holahan and Saegert (Reference Holahan and Saegert1973) suggested that dismal, unattractive psychiatric ward environments contribute to lethargy and depression. Corey et al. (Reference Corey, Wallace, Harris and Casey1986) utilized the Ward Atmosphere Scale to evaluate the milieu in a 101-bed general adult psychiatric hospital. The problems identified by patients and staff included: metal hospital-style furniture which emphasized the sick role; furniture arrangements which inhibited socialization; crowding in dining areas; vinyl tile flooring which contributed to noise, coldness and an institutional feeling; a pale monotonous color scheme felt to be unstimulating; sparse ornamentation; and a lack of private areas and space for personal belongings.
The available psychogeriatric literature focuses on long-stay wards and the effect of redesign on such factors as hostility, tension, sociability, “spatial distribution of behavior” and self-maintenance skills in severely and chronically disabled older people (Stahler et al., Reference Stahler, Frazer and Rappaport1984). These studies tend to be limited in scientific rigor, report impressionistic or qualitative observations of outcomes rather than direct measurement of quality of life or symptoms and differentiate poorly between the physical and social environment. There is some evidence that redesign of units contributes to reduced length of stay (Lawson and Phiri, Reference Lawson and Phiri2000). However, there have been few systematic studies as to how and to what degree this occurs.
Designing the optimal physical environment
Key design issues include optimizing patient and staff safety, managing patients' physical health, designing a productive ward environment and creating a ward atmosphere which enhances recovery.
Safety and design
It is inevitable that patients at a high risk of violence to themselves or others will require management in PGUs. Yet several authors maintain that inpatient psychiatric units should be free from disinhibited, noisy, and aggressive patients and that this is critical in determining recovery (Owen et al., Reference Owen, Tarantello, Jones and Tennant1998b; MIND, 2004). Others emphasize the importance of wards being a safe haven, a center for therapy, and a place for containment (Shrivastava et al., Reference Shrivastava, Kumar and Jacobson1999). Pollitt and O'Connor (Reference Pollitt and O'Connor2007) conclude that feeling safe is crucial, particularly for those who are physically frail. The literature suggests several strategies to address this dilemma.
Provision of a segregated area for patients at risk of violence
PGUs should have access to a psychiatric intensive care unit (PICU) or high dependency unit (HDU) for older adults at risk of harming themselves or others. A PICU has been defined as a secure unit which is usually for compulsorily detained, acutely disturbed patients who cannot be safely managed in a general open acute ward (Beer et al., Reference Beer, Pereira, Paton, Beer, Pereira and Paton2008). HDUs serve a similar function. The facilities of PICUs and HDUs may include more space, a garden, a quiet area, a seclusion suite, activity and games room as well as a specific model of care (Thomas et al., Reference Thomas, Jones, Johns and Trauer2006; Beer et al., Reference Beer, Pereira, Paton, Beer, Pereira and Paton2008). In the U.K. existing PICUs generally exclude older adults, specifically those with a primary diagnosis of dementia (Stevenson et al., Reference Stevenson, Khan and Perumal2009). Stevenson et al. (Reference Stevenson, Khan and Perumal2009) describe a PICU for older adults. Only males were admitted, almost all had dementia and 50% were maintained on their pre-admission medication, which was attributed to the therapeutic effect of the PICU environment, including provision of adequate personal space within a safe, secure and low-stimulus setting. In the U.S.A. “behavioral intensive care units” (BICUs) provide brief inpatient assessment and treatment for patients with severe behavioral and psychological symptoms in dementia (BPSD) (Mintzer et al., Reference Mintzer1993).
Literature on the design of PICUs for older people is lacking. Some guidance regarding the design of such units is available from the general adult psychiatric literature. However, this literature makes little reference to the PICU care of patients over the age of 60 (Stevenson et al., Reference Stevenson, Khan and Perumal2009). Dix and Williams (Reference Dix and Williams1996) make several recommendations regarding high dependency care. They recommend an “extra care area” in which a single acutely disturbed patient can be nursed, in preference to a seclusion room. They also suggest that convex mirrors be fitted at ceiling level where corridors meet and the avoidance of numerous corners and corridors to promote non-obtrusive observation as well as the provision of a general activities room. Separate bedrooms with louvre-type observation panels, doors that open two ways, vandal-proof dimmer light fittings and homely and safe fixtures and duress alarms were also recommended. O'Brien and Cole (Reference O'Brien and Cole2003) recommend that close-observation areas in acute general adult psychiatric units need to be humane and well resourced with highly skilled nurses. The environment needs to be secure, structured, low stimulating, non-intrusive and supportive. Locked ground-floor units with a separate entrance are recommended. Ample space should be provided (at least 30 square meters per patient) including access to enclosed gardens (O'Brien and Cole, Reference O'Brien and Cole2003). Empirical research into the design aspects of these units has been neglected (O'Brien and Cole, Reference O'Brien and Cole2003) and the available literature provides little guidance on how to ensure that the design of PICUs/HDUs meets the complex demands put upon them. However, the evidence and opinion presented here suggests that the management of patients at risk of harming themselves or others, particularly those with challenging behaviors and dementia, is facilitated by the provision of such units. These units require a design that facilitates patient observation as well as the use of high nurse-to-patient ratios. There are cost implications to this approach but we were unable to locate a formal cost benefit analysis.
Provision of seclusion and restraint
Many PGUs make provision for agitated aggressive patients to be physically restrained or placed in seclusion for brief periods until their behavior subsides (O'Connor et al., Reference O'Connor, Horgan, Cheung, Fisher, George and Stafrace2004). However, systematic literature reviews regarding seclusion focus on its utilization in adult settings (O'Brien and Cole, Reference O'Brien and Cole2003) or nursing homes (O'Connor et al., Reference O'Connor, Horgan, Cheung, Fisher, George and Stafrace2004). The literature offers limited guidance on the benefits or otherwise of using seclusion and restraint on PGUs (O'Brien and Cole, Reference O'Brien and Cole2003). De-escalation together with appropriate use of medication has been recommended to limit their use and there is a move towards restraint-free care, mostly with positive outcomes (O'Connor et al., Reference O'Connor, Horgan, Cheung, Fisher, George and Stafrace2004). Persistently intrusive, restless or aggressive patients in PGUs adversely impact on their fellow patients, relatives and staff. O'Connor et al. (Reference O'Connor, Horgan, Cheung, Fisher, George and Stafrace2004) found that in some PGUs good clinical care can be delivered to even the most disturbed older patients without seclusion but they do not suggest eliminating restraint from PGUs.
Segregation of patients according to diagnosis
It has been suggested that cognitively impaired patients should be segregated from those who are not cognitively impaired because those with cognitive impairment require different management programs and staff with special training (Human Rights and Equal Opportunities Commission, 1993; Pollitt and O'Connor, Reference Pollitt and O'Connor2007). However, few empirical studies have investigated the evidence base for this assertion. In a small study, limited by the potential for observer bias, Craig et al. (Reference Craig, Patel, Lee-Jones and Hatton2000) examined the possible benefits of physically segregating inpatients with challenging behavior in dementia from those with functional illness. Their qualitative study involved 192 hours of observation, following four patients on each of two wards. One of the wards had a mix of functional and cognitively impaired (CI) patients and the other had a separation of these patients within the ward. Of the four patients observed on each ward, two had functional disorders, two had organic disorders and two were male. In the mixed ward setting functional patients and visitors reported that the CI patients were distressing, disruptive, noisy, frightening and intrusive. Functional patients tended to stay in their beds to avoid CI patients and experienced low levels of stimulation and interaction with staff and other patients. The mixing of CI and non-CI patients contributed to low staff morale. Staff were busy with “hands on” nursing of the CI patients and tended to have a limited understanding of patients' psychological needs. On the split ward more time was spent by nursing staff in non task-focused conversations with patients which was associated with greater job satisfaction. The split ward was quieter, calmer and had more patient-staff interaction and this enhanced the therapeutic milieu. A higher level of interaction between patients was also observed. The communal areas and corridor were able to be carpeted making them more attractive and possibly reducing noise. Currently there is insufficient evidence to mandate that there be segregated units for patients with dementia or for those with depression.
Gender separation
Some health services have separate male and female sections within PGUs. It is U.K. Department of Health policy to increase the provision of same-sex wards for psychiatric patients. A recent inquiry into hospital services in NSW Australia (Garling, Reference Garling2008) made a similar recommendation. Barlow and Wolfson (Reference Barlow and Wolfson1997) found that a majority of female psychiatric inpatients felt that female-only wards would improve safety. Garling (Reference Garling2008) concluded it was inappropriate for men and women to be in the same bedroom in an inpatient ward and recommended this practice stop immediately (Garling, Reference Garling2008). There is limited literature regarding gender segregation in psychogeriatric settings. Pollitt and O'Connor (Reference Pollitt and O'Connor2007) suggest that establishing single sex areas in acute PGUs may be beneficial. This could include same sex bedroom accommodation but unisex day areas. Currently there is insufficient evidence to mandate complete gender segregation in psychogeriatric settings.
Density, privacy and control
Crowding, lack of privacy and loss of control are related to violence on inpatient psychiatric units. Several architectural strategies have been suggested to address these issues (Osmond, Reference Osmond1957; Shrivastava et al., Reference Shrivastava, Kumar and Jacobson1999; Kumar and Ng, Reference Kumar and Ng2001). Patients' need for privacy can be addressed by providing single bedrooms (Osmond, Reference Osmond1957; Kumar and Ng, Reference Kumar and Ng2001) or, if this is not possible, placing a cap on the number of beds per ward (Osmond, Reference Osmond1957; Shrivastava et al., Reference Shrivastava, Kumar and Jacobson1999). Patient density can be reduced by providing direct access to usable outdoor space as well as providing access to open communal areas (Shrivastava et al., Reference Shrivastava, Kumar and Jacobson1999). Having separate areas for social interaction and for quiet reflection enhances patients' level of choice and control and allows patients to retreat physically when they feel threatened (Osmond, Reference Osmond1957; Shrivastava et al., Reference Shrivastava, Kumar and Jacobson1999). It is important to indicate clearly the intended use of each area/room (Gulak, Reference Gulak1991; Kumar and Ng, Reference Kumar and Ng2001). This can be achieved by using lighting to help define space, making the spaces that have special meaning to patients stand out, using adequate signage as well as making areas visually distinct so that their appearance highlights the fact that they have a specific use (Kumar and Ng, Reference Kumar and Ng2001).
Using design to prevent injury and self harm
Good visibility for staff to all parts of the ward and outside areas reduces the risk of both intentional and unintentional injuries (e.g. falls) (Morgan and Stewart, Reference Morgan and Stewart1999; Eagle, Reference Eagle2006). However, this may impinge on patients' privacy. Crucial to optimal visibility is the design and location of the nurses' station to enable direct line of sight to as much of the PGU as possible. Ideally, all potential “ligature points” should be eliminated because of the risk of suicide. PGUs need provision of ‘suicide proof’ (enclosed bottom) hand rails and grab bars, particularly in bathrooms and showers in order to maximize safety while maintaining patients' dignity and independence (Karlin and Zeiss, Reference Karlin and Zeiss2006).
Hamill and Matthei (Reference Hamill and Matthei1980) make the following specific recommendations to reduce self-harm risk: bedroom and toilet doors should swing out from rooms and door locks should be eliminated on patient wardrobes to prevent patients barricading themselves in. Mirrors should be made of stainless steel. Electrical fixtures should be tamper-proof. Hanging risk can be reduced by using break-away curtain rods and closed loop hangers on a continuous loop rod. They also recommend limiting shower temperature to prevent scalds. Eagle (Reference Eagle2006) recommends that electrical outlets should disconnect in response to tampering or have tamper-proof screws, and that edges and corners in patient areas should be rounded off and ceramic tiles be curved at the edge. She also suggests using high ceilings to prevent patients reaching light fixtures or fire sprinklers as well as using solid plaster drywall ceilings to restrict ceiling space access. Lockable water taps, especially with semi-private baths, may prevent drowning (Folks and Kinney, Reference Folks, Kinney, Copeland, Abou-Saleh and Blazer2002).
Reducing the perception of the ward as a dangerous environment
Visual aspects of design can reinforce perceptions of violence. Remen (Reference Remen1992) noted an increasing emphasis on “optimizing” design for safety. He refers to “hidden hazards” that tend to proliferate, such as multiple notices, signs or symbols that indicate the hospital is a dangerous place rather than friendly and welcoming. High fences that give an appearance that the unit contains dangerous individuals who must be quarantined from the general community are likely to impact negatively on patients' sense of physical safety. The physical environment should be designed to separate or contain those patients who represent an actual or perceived risk to other patients. Where this is not possible it has been suggested that a separate day area for quieter, more independent or vulnerable patients be provided (New Zealand Ministry of Health, 2002; Pollitt and O'Connor, Reference Pollitt and O'Connor2007).
Security
We were unable to find any guidance in the literature as to whether or not the doors on wards for voluntary PGU patients should be locked. The issue of whether psychiatric facilities should generally be locked continues to be debated. The prevalence of locked psychiatric units in the U.K. and Sweden ranges from 25–73% (Haglund et al., Reference Haglund, Knorring and Essen2005; Bowers et al., Reference Bowers, Nijman, Allan, Simpson, Warren and Turner2006). Gudeman (Reference Gudeman2005) stated that acute psychiatric units in general hospitals and psychiatric hospitals are locked because of community perception that the patients are dangerous, for the convenience of staff, and because of stigma and hospital-wide resistance. His opinion is that when units are unlocked few disasters occur and patients are less stigmatized and better able to integrate into the community. Haglund et al. (Reference Haglund, Knorring and Essen2005) found that the staff mentioned more disadvantages than advantages to having locked doors. The effect of locked doors on core outcomes, such as prevention of harm, use of psychopharmacology or staffing levels, has not been investigated. Within the PGU, patients' desires to lock their bedroom doors may conflict with the needs of staff to have access to them in case of mishap (Pollitt and O'Connor, Reference Pollitt and O'Connor2007). PICUs in the U.K. have evolved, somewhat paradoxically, in synchrony with the development of “open door” policies (Stevenson et al., Reference Stevenson, Menon and Gallacher2007). This supports the notion of segregation of patients according to risk/diagnosis. Patients at high risk of harming themselves or others should have access to a secure high dependency ward area.
Managing the physical health of PGU patients
Good design can play a significant role in minimizing falls risk, new incontinence (e.g. by considering distance to and location of the toilets) and preventable functional decline (e.g. including spaces in which people can move safely or have access to such spaces external to the unit). The literature regarding general hospitals provides guidance on using design to reduce falls, medical errors and infection rates, as well as on improving stress, sleep, privacy and patient satisfaction in medical settings (Ulrich, Reference Ulrich2006). What follows is guidance as it relates to PGUs.
Location
Good external visibility helps to reduce the stigma traditionally attached to mental health inpatient units (Hamill and Matthei, Reference Hamill and Matthei1980). Good links to public transport are important for staff and visitors.
Design issues related to medical co-morbidity and delirium
Medically unwell patients require lighting for bedside examination, adjustable hospital beds and outlets for oxygen and suction. Provision of hand basins close to patient examination areas encourages good staff hygiene. Glucose monitoring facilities and cardiac telemetry are highly desirable (Folks and Kinney, Reference Folks, Kinney, Copeland, Abou-Saleh and Blazer2002). Adequate storage space is required for patient lifters, medical equipment and supplies (Folks and Kinney, Reference Folks, Kinney, Copeland, Abou-Saleh and Blazer2002). Whilst patients with acute delirium are generally managed on medical wards, some PGU patients may have low grade delirium. Such patients are best managed in single rooms with appropriate lighting. “Dry erase boards” in each room for orientation are recommended as part of a comprehensive program for delirious patients (Inouye et al., Reference Inouye, Bogardus, Baker, Leo-Summers and Cooney2000).
Falls
Improving hospital facility design may decrease fall rates and fall-related injuries, although this has not been examined systematically (Tzeng and Yin, Reference Tzeng and Yin2008). Most falls occur when getting in or out of bed and are not witnessed (Eriksson et al., Reference Eriksson, Strandberg, Gustafson and Lundin-Olsson2009). In an integrated medical unit the location of falls appears to be associated with the ward design. Improved visibility from the nurses' station was associated with fewer falls in bed areas (Vassallo et al., Reference Vassallo, Amersey, Sharma and Allen2000). The design-regulation height and staff-working height for patient beds used in acute care units may contribute to patient falls and the severity of fall-related injuries (Tzeng and Yin, Reference Tzeng and Yin2008).
Some of the available falls literature is contradictory. Darkness, anxiety and not wearing shoes were associated with increased falls in patients with dementia in a PGU (Eriksson et al., Reference Eriksson, Strandberg, Gustafson and Lundin-Olsson2009). In a retrospective cohort study of falls in a PGU, six variables were found to be independently associated with falls but none was design-related (de Carle, Reference de Carle and Kohn2001). In a small prospective randomized study in an elderly care rehabilitation ward there was a strong trend (though not statistically significant) towards vinyl floor coverings being superior to carpet in falls prevention at the bedside (Donald et al., Reference Donald, Pitt, Armstrong and Shuttleworth2000). However, an observational study found that falls on vinyl were more likely to lead to injuries (Healey, Reference Healey1994). Concerns have been raised about odor accumulation, cleaning difficulties and infection risk with carpet, but have not proved to be problematic in practice when using appropriate carpet cleaning procedures (Healey, Reference Healey1994; Donald et al., Reference Donald, Pitt, Armstrong and Shuttleworth2000). Further studies are needed.
Designing a productive ward environment
“Lean thinking for the productive ward” is a method for organizing complex production processes so as to encourage workflow and reduce inefficiency (Ben-Tovim et al., Reference Ben-Tovim2007). ‘Redesigning care’, based on this methodology, has been applied in general hospitals which are notoriously disorganized workplaces (Ben-Tovim et al., Reference Ben-Tovim2008). An efficient hospital layout should promote staff efficiency by minimizing distance of unnecessary travel between frequently used spaces (Carr, Reference Carr2009). Locating the PGU close to the ECT suite, emergency department, acute geriatric medical wards, and radiology could improve access to medical care. This has not yet been formally evaluated in mental health settings. Both safety and efficiency can be improved, especially for rotating and relieving staff, by standardizing physical spaces in the hospital – for example, standardizing storage of medications, unit secretary workstations and equipment rooms (Ben-Tovim et al., Reference Ben-Tovim2008).
Optimizing ward milieu
The literature provides opinion and anecdote on the ideal ward milieu in the absence of empirical data (Fottler et al., Reference Fottler, Ford, Roberts, Ford and Spears2000; Eagle, Reference Eagle2006). Several authors claim that the experience of architecture is an important part of the healing process and emphasize settings which are welcoming and maximize patient privacy (Hair, Reference Hair1998; New Zealand Ministry of Health, 2002). An optimal environment should be conducive to relaxation, independent activity, and encourage appropriate social contact between patients and visitors. Public spaces should be welcoming (New Zealand Ministry of Health, 2002). Materials in a psychiatric unit should preferably be of high quality, to counter patients’ likely poor self-image and help build feelings of status and coping (Bergsland, Reference Bergsland2002).
Historically there has been a lack of patient involvement in hospital design (Kenny and Canter, Reference Kenny, Canter, Canter and Canter1979). However, several studies have found that the comfort of resting areas, overall cleanliness, décor, low noise, privacy and cheerfulness of facilities are positively related to patient satisfaction (Hall and Dornan, Reference Hall and Dornan1988; Fottler et al., Reference Fottler, Ford, Roberts, Ford and Spears2000). Others also report that the service setting is one of the more important determinants of customer satisfaction (Taylor, Reference Taylor1994).
Fottler et al. (Reference Fottler, Ford, Roberts, Ford and Spears2000) emphasize the importance of providing a setting that meets or exceeds patients’ needs for safety, security, competence and physical and psychological comfort: “an excellent healing environment will reinforce excellent clinical quality” (Fottler et al., Reference Fottler, Ford, Roberts, Ford and Spears2000, p. 91). They argue that the environment is important in promoting patient healing and well-being and improving staff attitudes. They hypothesize that environmental dimensions contribute to patient and staff psychological and physiological responses and behavior. They contend that this leads to outcomes of perceived quality and value, patient satisfaction, willingness to return and willingness to recommend. For example, the exterior environment provides a first impression and delivers a message about the organization, its services, and its quality long before the clinical encounter takes place. They stress the importance of homelike settings, as supported by the “Green House” model of nursing home care. This consists of small facilities (7–10 beds) with private rooms and bathrooms, a residential kitchen and a communal dining area and living room with a fireplace. Early experience with this care model demonstrated that there are significant benefits in quality of life for residents, although these effects may also have been due to the staffing model and philosophy of care (Rabig et al., Reference Rabig, Thomas, Kane, Cutler and McAlilly2006; Kane et al., Reference Kane, Lum, Cutler, Degenholtz and Yu2007). Although much of the above has face validity, it is yet to be fully supported by empirical data. Nevertheless, patient satisfaction is likely to be strongly linked to ward design and functioning.
Ambient environment (“atmospherics”)
Atmospherics, as defined by Kotler (Reference Kotler1973), are those physical and controllable environmental components that fulfill customers’ basic desires for comfort, convenience, safety, entertainment and information. They include temperature, humidity, air quality, smells, sounds, colors, lighting and textures. The interior built environment can influence patients’ perception of space, and therefore their behavior. Various interior design elements such as furnishings, wall coverings, floor coverings, artwork and window treatments as they relate to patients’ acceptance, aesthetics, maintenance, and budget should be considered (St Clair, Reference St Clair1987; Shrivastava et al., Reference Shrivastava, Kumar and Jacobson1999). Izumi (Reference Izumi1976) suggests the environment should be unambiguous, coherent and unobtrusive, rather than bland, sterile and stimulus-free. Whall and Conklin (Reference Whall and Conklin1985) caution that geometric patterns are not good for floor coverings as they sometimes resemble steps to those with visual problems. Ability to individualize bedrooms, e.g. having a bulletin board in each room, helps to maintain a less institutionalized environment (Hamill and Matthei, Reference Hamill and Matthei1980). However, claims that exposure to art in healthcare environments has been found to reduce anxiety and depression (Daykin et al., Reference Daykin, Byrne, Soteriou and O'Connor2008) cannot be justified because of confounding factors. There is evidence that minimal architectural changes (e.g. additional entrances/stairways to improve accessibility, brightening paint, changing lighting and separating living and dining areas) in a treatment environment can lead to significant improvement outcomes for staff, patients and families (Gutkowski et al., Reference Gutkowski, Ginath and Guttmann1992). Muted colors such as blue-grays and blue-greens should be avoided as they are difficult to distinguish by the elderly (Whall and Conklin, Reference Whall and Conklin1985).
Izumi (Reference Izumi1976) recommends soft, indirect and pervasive or full-spectrum lighting, while spotlight-type recessed lighting should be used sparingly and be carefully placed, so as not to focus directly on individuals. Ample natural daylight is highly valued by patients and designers. There is some evidence by way of natural experiment in a retrospective study suggesting that sunlight in patient rooms can promote the recovery of psychiatric patients with severe depression, even in the absence of seasonal affective disorder (Beauchemin and Hays, Reference Beauchemin and Hays1996). Fottler et al. (Reference Fottler, Ford, Roberts, Ford and Spears2000, p. 100) claim that “good lighting, open space, smiling employees who make eye contact with customers, and cleanliness make customers feel secure”.
Numerous authors have identified multiple windows with views of nature as a valuable design feature (Leather et al., Reference Leather, Pyrgas, Beale and Lawrence1998; Ulrich, Reference Ulrich, Cooper-Marcus and Barnes1999; Harris et al., Reference Harris, McBride, Ross and Curtis2002). Views can reduce psychological distress and recovery time in surgical settings and have been demonstrated to enhance staff functioning and job satisfaction (Karlin and Zeiss, Reference Karlin and Zeiss2006). Large low windows may improve sensory abilities and reduce delirium and paranoia. Outdoor gardens and other natural environments can serve as positive distractions and reduce stress and fatigue (Ulrich, Reference Ulrich, Cooper-Marcus and Barnes1999). A feeling of stimulation could be created in a geriatric unit through the use of brighter colors and visual stimulation cues. However, this may not be suitable for all psychogeriatric patients. In practice, the ideal level of stimulation for a depressed patient may be overstimulating for a manic or cognitively impaired patient (Grant, Reference Grant1998). Solutions arise from considering the ward as comprising separate areas that could meet these differing needs or providing separate units for patients with specific diagnoses. Many patients are extremely sensitive to their auditory environment and noise levels at times may be high. This calls for a high degree of control of the acoustics in the inpatient common spaces (Lawton et al., Reference Lawton, Fulcomer and Kleban1984; Bergsland, Reference Bergsland2002; Harris et al., Reference Harris, McBride, Ross and Curtis2002). Good air quality – with fresh air, good ventilation and neutral odors – is recommended as it can facilitate recovery (Karlin and Zeiss, Reference Karlin and Zeiss2006). Well-lit parking lots and pathways, low-cut hedges and the presence of uniformed employees have been proposed as important signals of a safe environment (Fottler et al., Reference Fottler, Ford, Roberts, Ford and Spears2000).
Spatial environment (“architectural features”)
The architectural features are the relatively permanent aspects of the hospital environment. They include layout, size and shape of the unit and may encompass equipment and furnishings, including the arrangement, size and shape of these objects, their accessibility to patients and the spatial relationships among them. The spatial layout should give people a sense of orientation so they know where they are at all times and should keep the walking required between services to a minimum (Fottler et al., Reference Fottler, Ford, Roberts, Ford and Spears2000). However, exercise areas and the presence of some longer distances may encourage physical activity (e.g. walking to the hospital coffee shop). A multifaceted intervention including graded exercise classes was found to benefit older people with depression in residential care (Llewellyn-Jones et al., Reference Llewellyn-Jones, Baikie, Smithers, Cohen, Snowdon and Tennant1999). This intervention used volunteers to assist the frail aged in walking outdoors using purpose-built walking tracks (Llewellyn-Jones et al., Reference Llewellyn-Jones2001).
Domestic scale and feel are recommended in order to make the inpatient experience more familiar and less confusing (Remen, Reference Remen1991; Fottler et al., Reference Fottler, Ford, Roberts, Ford and Spears2000; New Zealand Ministry of Health, 2002). Dinshaw (Reference Dinshaw2006) emphasizes the need for privacy and dignity. Some authors have specifically focused on providing “deinstitutionalizing” features such as the separation of living and treatment areas, elimination of long echoing corridors and highly reverberant spaces, easy access to safe outdoor areas, encouragement of greater patient/staff interaction through removing nurses’ stations, and providing residential features such as a laundry and galley kitchens (Hamill and Matthei, Reference Hamill and Matthei1980). Issues of crowding may be better managed in certain modern facilities (Bergsland, Reference Bergsland2002). Based on the work of Kumar and Ng (Reference Kumar and Ng2001) and following other qualitative research, the provision of sufficiently large indoor and outdoor spaces, sufficient activities areas, and limiting the number of beds should remain a priority. The quantity of space and the space scheduling requirements may be related to admission criteria. Facilities are required for social, therapeutic and individual activities including a therapy kitchen, occupational therapy garden and storage of appropriate location and size.
Bergsland (Reference Bergsland2002) reported that psychotic patients have personal safety zones that are many meters larger than when in a non-psychotic phase. Higgs (Reference Higgs1970) found significant clinical improvements among patients with schizophrenia following a move from an older ward to a recently constructed one. Holahan and Saegert (Reference Holahan and Saegert1973) found repainting, improving privacy and using new and rearranged furniture was associated with increased socializing. Whitehead et al. (Reference Whitehead, Polsky, Crookshank and Fik1984) found that reduced psychopathology among younger patients was facilitated by design solutions that encouraged social interactions such as the introduction of a ward kitchen, an open nursing station, and a dayroom module for crafts and other occupational therapy activities. Devlin (Reference Devlin1992) found that patients were less demanding on staff and staff felt more supported by renovations including new and rearranged furniture, plants, wallpaper, new bright colored paint, better lighting and curtains in a 40-year-old psychiatric facility. Gutkowski et al. (Reference Gutkowski, Ginath and Guttmann1992) considered that adding entrances and opening doors improved patients’ sense of freedom and orientation.
Private (single) bedrooms are recommended and should be provided if possible (Morgan and Stewart, Reference Morgan and Stewart1998; Harris et al., Reference Harris, McBride, Ross and Curtis2002). Historically, however, mental health units have had a mix of single and shared bedrooms. Appropriate bedroom size is an issue that has provoked much debate (Ittleson et al., Reference Ittleson, Proshansky and Rivlin1970). The major conflicting pressures in arriving at a solution have been economic (favoring larger shared rooms) versus therapeutic (favoring single or small shared rooms). Ittleson et al. (Reference Ittleson, Proshansky and Rivlin1970) attempted to assess the impact of shared rooms on the social interaction of patients on a psychiatric ward by comparing several units that comprised different sized rooms. There were clear confounding factors, which limit the reliability of the study findings. However, they found that being alone in a small room did not necessarily lead to withdrawal. Contrary to psychiatric folklore, it was the large multiple-occupancy room that provoked patients’ withdrawal.
Single and non dormitory bedrooms enhance privacy and autonomy and may promote participation in treatment activities. Pollitt and O'Connor (Reference Pollitt and O'Connor2008) found having a room to oneself was particularly valued. Private visiting areas increase privacy and intimacy (Karlin and Zeiss, Reference Karlin and Zeiss2006). There is a trend towards having only single bedrooms in new psychiatric inpatient units in NSW, Australia. But there may be some advantages to having some shared rooms available with the ability to “room in”. These include benefits from having some shared rooms for patients who are withdrawn, or for cases in which a higher-functioning roommate can provide aid to another with respect to orientation and structure (Folks and Kinney, Reference Folks, Kinney, Copeland, Abou-Saleh and Blazer2002). Ensuite bathrooms are preferable to large shared bathrooms (Harris et al., Reference Harris, McBride, Ross and Curtis2002; Pollitt and O'Connor, Reference Pollitt and O'Connor2008). They should be wheelchair accessible, include hand rails and have the facility for patients who may have fallen to alert staff (Folks and Kinney, Reference Folks, Kinney, Copeland, Abou-Saleh and Blazer2002).
Signs and symbols
Signs must be easy to read, clear and located in obvious places where they can direct and teach people how to use the service easily. Signs in psychiatric settings should convey sincere respect for the patient and sensitive concern for their well-being (Remen, Reference Remen1991). Dinshaw (Reference Dinshaw2006) suggests that signs for patients in geriatric settings should be clear and highlighted by a contrasting background while those that are only relevant to staff should not.
Other miscellaneous features
These include less permanent aspects of the hospital environment such as furnishings and their arrangement. Two small short-term randomized controlled trials found that rearranging seats (previously placed along the walls) around small tables and improving lighting improved communication and behavior in long stay patients with dementia in psychogeriatric units (Melin and Götestam, Reference Melin and Götestam1981; Götestam and Melin, Reference Götestam, Melin and Levi1987). Although not readily generalizable, the work of Stahler et al. (Reference Stahler, Frazer and Rappaport1984) found similar results for patients with schizophrenia. The most consistent recommendations are to reduce the institutional feel of the hospital to make it more home-like. This has been associated with enhanced emotional and intellectual well-being (Karlin and Zeiss, Reference Karlin and Zeiss2006). Carpeting some parts of a psychogeriatric ward may be appropriate depending on the patients admitted, especially in units where patients are generally continent (Karlin and Zeiss, Reference Karlin and Zeiss2006). Folks and Kinney (Reference Folks, Kinney, Copeland, Abou-Saleh and Blazer2002) recommend industrial grade carpet to soften falls in hallways and/or common areas but with vinyl to facilitate cleaning in patients’ bedrooms.
Staff satisfaction and design
Kenny and Canter (Reference Kenny, Canter, Canter and Canter1979) noted that staff satisfaction with the design features of a ward were dependent on the extent to which the features enabled them to carry out their role. Fottler et al. (Reference Fottler, Ford, Roberts, Ford and Spears2000) report that “no one wants to work in a dangerous, dirty, and depressing environment. Employees want an environment that is clean, organized, pleasant smelling and comfortable.” A pleasant work environment improves work performance and conveys the message that the organization is committed to quality and staff well-being. Reducing noise and improving air quality have been suggested to be effective in reducing violence to staff (Joseph and Rashid, Reference Joseph and Rashid2007).
The presence of a staff lounge, garden, or similar congregate space can improve morale and job satisfaction and encourage professional communication among the multidisciplinary team (Folks and Kinney, Reference Folks, Kinney, Copeland, Abou-Saleh and Blazer2002; Karlin and Zeiss, Reference Karlin and Zeiss2006). There should be appropriate facilities for staff training and staff showers and toilets. Staff offices should be congenial, sufficient in number and their location needs to be functional. A staff office near the entrance of the unit allows ingress and egress from the unit to be monitored.
Optimal design for patients with dementia
Most of the literature on this topic refers to long-term residential care of people with dementia, which is outside the scope of this review. However, some guidance for acute care in PGUs can be derived from this literature. Fleming et al. (Reference Fleming, Forbes and Bennett2003) suggest that the principles to help maintain the functional abilities of patients with dementia are that the environment should be safe and secure, small, simple, provide good visual access, reduce unwanted stimuli, highlight helpful stimuli and provide for planned wandering. They should also be familiar, provide opportunities for both privacy and community, provide links to the community and be domestic.
Special Care Units (SCUs) are segregated residential aged care facilities that accommodate only cognitively impaired individuals, especially people with dementia. Segregation can benefit residents without cognitive impairments (Wiltzius et al., Reference Wiltzius, Gamberg and Duthie1981; Teresi et al., Reference Teresi, Holmes and Monaco1993), as well as address the specific design needs for those with dementia. Morgan and Stewart (Reference Morgan and Stewart1999) found that a significant decrease in disruptive behavior occurred in a lower density special care unit for patients with dementia. Whilst the optimal design and effectiveness of SCUs has been debated, uncontrolled and quasi-experimental studies have shown associations between SCU environments and improved continence, ward orientation, social function, mobility affective responses, behavioral disturbance, exit-seeking, apathy, hallucinations and relatives’ visiting rates as well as slowed decline in residents’ communication and self care skills (Lawton et al., Reference Lawton, Liebowitz and Charon1970; Reference Lawton, Fulcomer and Kleban1984; Hanley, Reference Hanley1981; Benson et al., Reference Benson, Cameron, Humbach, Servino and Gambert1987; Hussian and Brown, Reference Hussian and Brown1987; Mayer and Darby, Reference Mayer and Darby1991; Namazi and Johnson, Reference Namazi and Johnson1991a; Reference Namazi and Johnson1991b; Reference Namazi and Johnson1992a; Reference Namazi and Johnson1992b; Reference Namazi and Johnson1996; Dickinson et al., Reference Dickinson, McLain-Kark and Marshall-Baker1995; Cohen-Mansfield and Werner, Reference Cohen-Mansfield and Werner1998; Morgan and Stewart, Reference Morgan and Stewart1999; Day et al., Reference Day, Carreon and Stump2000). On the basis of these studies helpful design features include: locks on exit doors; removal of obstacles for wandering patients; increased privacy; neutral wall colors; pleasant sounds, aromas and artwork; soundproofing; room doors and handrails of bright colors to facilitate their identification; easily visible and non-institutional bathroom facilities; access to outdoor areas and the creation of separate areas for specific activities. Further guidance on these issues can be found in Cohen and Day (Reference Cohen and Day1993), Day et al. (Reference Day, Carreon and Stump2000) and Hoskins and Marshall (Reference Hoskins, Marshall, Jacoby and Oppenheimer2002).
Design and the commissioning process
Several authors emphasize the importance of good communication between administrators, architects and clinicians (Gulak, Reference Gulak1991; Gutkowski et al., Reference Gutkowski, Ginath and Guttmann1992; Shrivastava et al., Reference Shrivastava, Kumar and Jacobson1999; Kumar and Ng, Reference Kumar and Ng2001). Gulak (Reference Gulak1991) notes, however, that this is rare in the design of psychiatric hospitals. Service users must be involved in the planning and delivery of proposed services. A “user group” has been recommended as a way to involve people with service user experience in planning and design in order to ensure that all those involved understand the end users of the facility (New Zealand Ministry of Health, 2002). A team approach is needed for this complex and challenging task. The user group aims to bring together those with different and often contradictory needs and ensures structural errors are not the result of allowing only the leaders to make design decisions (Beal and Keller, Reference Beal and Keller1989). It is recommended that there be careful selection of key personnel, team meetings to develop unit concepts prior to opening, and use of patient feedback and criticism as a guide (Malone and Holden, Reference Malone and Holden1987).
Gralton et al. (Reference Gralton, Pearson, Sutherland, Donovan and Lewis2001) undertook a pilot study to determine whether the perception of staff played an important role in the assessment of the ward environment. They generally found a lack of consistency in the way staff evaluated a range of factors as positive or negative. Interestingly, there were no consistent gender differences and older more experienced staff were not desensitized to poor conditions. This suggests that design should not be left to a few individuals and that patients and staff also need to be involved in design. Sivadon (Reference Sivadon1980) specifically states that psychiatrists must be involved in the design process.
Conclusion
This literature review indicates that there are significant problems with current acute PGUs and that the mix of patients in existing units is a significant issue. The available literature to guide the design of future acute PGUs is diverse and rigorous studies of the relationship between ward design and clinical outcomes, patient and staff safety, and satisfaction are lacking. Whilst such issues have been better studied in general hospitals (Ulrich et al., Reference Ulrich2008), the physical environment is the least studied but perhaps one of the most important and controllable dimensions of patient satisfaction (Fottler et al., Reference Fottler, Ford, Roberts, Ford and Spears2000; Nelson et al., Reference Nelson, West and Goodman2005). Research in the area of psychiatric inpatient unit design is still in its infancy (Karlin and Zeiss Reference Karlin and Zeiss2006).
There is a substantial literature on how to design residential facilities for people with dementia. However, research on the design of acute PGUs for patients with and without dementia has been neglected. Much of the existing literature is based on opinion and anecdote or at best based on observational studies. Despite its limitations the literature provides useful guidance on the design of acute PGUs. It must be acknowledged that whilst improved PGU design may assist, improving clinical outcomes also depends on implementing “patient centered” rather than “staff centred” models of care. Both patients and clinicians need to be involved in PGU design.
This review has identified significant gaps in knowledge about the design of acute PGUs. Systematic studies of the problems that staff, patients and their relatives encounter in existing units, including the possible lasting sequelae of traumatic experiences (e.g. PTSD) in PGUs, are needed. Further research is needed on patients’ perceptions of different PGU ward environments. The effects of segregating patients with challenging behaviors in dementia from those with functional psychiatric illness needs investigation. The clinical impact of locking the doors of acute PGUs is worthy of examination. Literature on the design of PICUs/HDUs for older people is lacking and more research is needed on the benefits or otherwise of using seclusion and restraint on PGUs. More needs to be known about the benefits or otherwise of gender segregation in psychogeriatric settings. Research is needed on how PGU design could reduce falls and fall-related injuries. Little attention has been paid to environmental sustainability or staff perceptions of the work environment and these should be further explored.
New units should be subjected to a post occupancy evaluation examining the strengths and weaknesses of their designs. Future research could usefully compare clinical outcomes and patient/staff perceptions before and after new units are built or redesigned. Quasi-experimental and pre-post research methodologies could be used to examine the usefulness of specific design features. Where possible, control groups/conditions should be used. Randomized controlled trials comparing different ward designs are likely to be impractical. Future studies would benefit from clearly describing the clinical populations of the wards studied. Certain design principles are common sense and do not require empirical validation.
In 1815 a British Parliamentary inquiry into madhouses identified overcrowding, inappropriate patient mix and insufficient numbers of suitably qualified staff as among the “basic evils” of inpatient psychiatric care of the time (Mental Health Act Commission, 2008). Despite undoubted subsequent advances in psychiatric inpatient care the strong parallels between these “basic evils” and current deficiencies of inpatient psychiatric care are striking. Osmond (Reference Osmond1957, p. 29) suggested a radically different approach to the provision of inpatient psychiatric care. It was his hope that “the best we can build now will be obsolescent in 15 to 20 years time”. He suggested adopting the plane designer's motto of: ‘Draw ’em, build ’em, test ’em, fly ’em, scrap ’em.’ We hope that this review will provide a strong impetus for researchers, clinicians and policy-makers to address the current deficiencies in acute PGU care and to improve knowledge about optimal PGU design.
Conflict of interest
None.
Description of authors' roles
J. Dobrohotoff (JD) and R. Llewellyn-Jones (RLJ) jointly initiated this review. JD performed the literature review. RLJ focused the review and also evaluated the literature. JD and RLJ identified gaps and key concepts and reviewed studies for methodological rigor and jointly drafted the paper.
