| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
Correspondence: Address correspondence to Margaret Calkins, PhD, IDEAS Institute, 8055 Chardon, Kirtland, OH 44094-9580. E-mail: mcalkins{at}IDEASInstitute.org
 |
Abstract |
|---|
 TOP Abstract MethodsDiscussionAppendixReferences |
|---|
Key Words: Construction • Design • Nosocomial Infection • Operational costs • Privacy
Our purpose in this exploratory project was to define as broad a range as possible of potential factors associated with different bedroom configurations, and to determine the extent of existing evidence, both empirical and anecdotal, that supports one bedroom configuration over another. In particular, our goal was to move beyond the relatively well-documented satisfaction-related outcomes to explore other factors that impact the life-cycle costs of private versus shared bedrooms.
The vast majority of research in nursing homes on this topic relates to psychosocial outcomes (preference and satisfaction). There is some, albeit more limited, research on clinical factors, although this is well studied in hospitals. Despite a growing interest in staffing issues, there is relatively little research that explores operational correlates of different room configurations on operational factors. Because of the lack of any previous comprehensive examination of the broad range of factors related to private rooms in nursing homes, for this project we drew on a framework developed by Chaudhury, Mahmood, and Valente (2005) to explore single- versus multiple-occupancy rooms in hospitals. Chaudhury and colleagues identified three clusters of factors: organizational costs (initial construction and ongoing operating costs), hospital management and patient care issues (infection control, patient transfer, and patient monitoring), and therapeutic impacts (privacy, stress, and family accommodation). We modified their framework slightly for this project, separating organizational factors from resident factors. Organizational factors can be further broken down into building-related issues (design and capital costs for construction and building operation) and operational issues (staffing issues, marketing or maintaining census, and time spent managing residents). Resident factors include psychosocial outcomes (well-being, satisfaction) and clinical issues (sleep, falls, nosocomial infections, etc.). This framework, shown in Figure 1, suggests that evidence (with greater weight on evidence-based outcomes than empirical outcomes) about resident factors should be fed into the decision-making process about design and operational issues (which also uses evidence-based and empirical information) to determine the ideal mix of bedroom configurations for a given project. Ideally, more research is then conducted on resident outcomes, which is fed back into the cycle again.
|
|
Methods |
|---|
|
TOPAbstractMethodsDiscussionAppendixReferences |
|---|
Before conducting a more thorough literature search, we conducted interviews with four nursing home administrators and four architects specializing in long-term care to flush out additional factors within each dimension that might not surface readily in the literature review. We then used these terms (from the initial search and the interviews) to conduct a systematic review of the literature. We conducted initial searches on Ageline and PubMed, and we included articles from 1970 to the present in our search. As we identified and abstracted articles, we also culled their references for related articles. We included only those articles that specifically addressed bedroom design or configuration, both empirical and anecdotal. We categorized articles by setting and type (empirical or descriptive). Because some of the topics identified in the interviews were not found in the literature, we held focus groups in three nursing homes, with staff, family members, and residents in attendance, to further probe the importance of these other factors. We selected a focus group format because it allows for discussion among different departments (nursing, social work, housekeeping, maintenance, and dietary), and this setup can encourage fertile discussions about topics that are sometimes infrequently thought about. We used a semistructured discussion guideline to allow for open-ended discussion and to ensure that all topics were systematically covered; this also allowed us to identify additional factors. Focus groups were run by two individuals, with one serving as facilitator and one as recorder. We identified several additional factors through the focus groups, and we conducted a second literature search (following the same parameters already described) for references on these factors.
We identified a total of 112 articles. Although we made efforts to focus on references specifically related to nursing homes (n = 55), some topics were only addressed in articles related to other settings (hospitals, n = 37; independent or assistant living, n = 7; multiple settings, such as articles on transfers, n = 7; and other or nonsetting specific, n = 6). It is worth noting that none of the published references differentiated a traditional shared bedroom, in which beds are side by side and occupants share one window and one bathroom, from what we refer to in this article as an enhanced shared bedroom, which is a relatively newer configuration in which each person has his or her own distinct territory and window and does not have to cross into the roommate's space to reach his or her own (see Figure 2).
|
Results of the Literature Review
We identified a total of 38 different factors within the four dimensions of the model. We identified most of the factors in at least one published reference, although there were several factors that we identified in focus groups that we did not find in the literature on either nursing homes or acute care settings (we discuss this issue separately). The vast majority of references related to the resident side of the model (psychosocial and clinical) as opposed to operational and building factors. The appendix lists the references, the setting (hospital, nursing home, etc.), and which factors we identified in each article. Because the purpose of this project was not a meta-analysis of extant research but rather an exploration of the broadest range of possible factors, we provide no other analysis of the articles.
Psychosocial Factors
There is strong evidence that, as a general cohort, older adults overwhelming prefer private rooms over shared rooms in residential settings, potentially even among people who thought they would prefer a shared room. A study by the American Association of Retired Persons found that individuals over the age of 50 preferred a private room by a ratio of 20:1 (82% vs 4%; see Baugh, 1996). These results replicate early research on the preference for private rooms conducted by Lawton and Bader (1970). The primary factors that influence this preference appear to be privacy (for self and when conversing with others), lack of control (over lifestyle and environment), and feeling uncomfortable being forced to be an "unwilling observer" to others, though several of these come from anecdotal resources and not empirical studies.
There is also evidence that seniors tend to express satisfaction with their current living situation, regardless of the objective quality of that housing (Pinquart & Burmedi, 2004). However, in a small-scale study conducted in Japan, Terakawa (2004) studied residents who moved from an older nursing home where all bedrooms were traditional shared bedrooms to a new nursing home where all residents had private rooms. The results indicated that even people who initially did not want a private room and expected not to like having a private room were completely satisfied with their private room by 8 months after the move. These results suggest that expressed opinion about satisfaction with or preference for a shared room may be based on being reasonably satisfied with a current situation (in a shared room) and may not be based on experience in both a private and shared bedroom. Other researchers have found that having a private bedroom is among the most desired changes of nursing home residents (Mosher-Ashley & Lemay, 2001). Residents who desired more privacy had lower life satisfaction than did residents who felt they had sufficient privacy.
There is a related concept of privacy with visitors, though the vast majority of research on this topic comes from acute care settings, where visiting, often with multiple people, occurs on a more frequent basis. Patients feel they have better visits with families in a private room, and they express higher satisfaction with this configuration (Chaudhury et al., 2005; Ulrich & Zimring, 2004).
Lack of control is another commonly cited factor that impacts preference for a private room in nursing homes. Common issues that cause conflicts between roommates include the television and radio (on or off, volume, and program selection); the time to get up and go to bed; having curtains open or closed; having the door to the hallway open or closed; heating, ventilation, and air conditioning levels; and the personalization or decoration of one's room (Foltz-Gray, 1995; Harris, McBride, Ross, & Curtis, 2002; Kaldenberg, 1999; Kane, Baker, Salmon, & Veazie, 1998; Ulrich & Zimring, 2004).
There is also some limited, mostly anecdotal, evidence about the positive benefits of sharing a room. Bitzan (1998) studied 31 nursing home residents who lived in shared rooms and found that 22% indicated an overall strong or positive emotional bond with their roommate, whereas 78% had a moderate or weak emotional bond with their roommate. Interestingly, even among those who indicated a positive emotional bond with their roommate, the majority did not enjoy spending time with their roommate, did not perceive their roommate to be sensitive to their feelings, and agreed they got along best when they kept their feelings and activities to themselves.
Clinical Factors
In clinical terms, the evidence is strong on iatrogenic outcomes, especially related to nosocomial infections. Pneumonia, the leading cause of death among nursing home residents, with overall mortality rates reported between 20% and 50% and as high as 80% in some studies (Zimmerman, Gruber-Baldini, Hebel, Sloane, & Magaziner, 2002), is the second most frequent nosocomial infection in nursing homes (Harkness, Bentley, & Roghmann, 1990). The vast majority of research suggests that there is a reduced risk of developing a nosocomial infection in a private room than in a shared bedroom (Fune, Shua-Haim, Ross, & Frank, 1999; Pegues & Woernle, 1993; Sharbaugh, 2003; Zimmerman et al.), although much of the research was conducted in acute care settings (Ben-Abraham et al., 2002; Berry, 2004; Boyce, Potter-Bynoe, Chenevert, & King, 1997; Chang & Nelson, 2000; Chaudhury, Mahmood, & Valente, 2004; Coleman, 2004; Drinka, Krause, Nest, Goodman, & Gravenstein, 2003; Ulrich & Zimring, 2004). Research conducted in nursing homes found that roommates of individuals infected with Influenza A had a 3.07 relative higher risk of acquiring the illness than did individuals in a private room (Drinka et al.). This statistic, combined with the 3.5% excess mortality rate associated with acquiring Influenza A, has serious life-threatening implications. Similarly, Pegues and Woernle found that 84% of nursing home residents who developed acute nonbacterial gastroenteritis during an outbreak lived in a room with a roommate, whereas only 16% of residents who became ill lived in private rooms. Beyond the potentially life-threatening consequences, there are also significant cost implications of nosocomial infections in nursing homes, which are estimated in one study to be in the range of $1 billion (Kayser-Jones, Wiener, & Barbaccia, 1989).
The empirical evidence of the negative impact on sleep in shared rooms in hospitals is fairly strong (Duffin, 2002; Ulrich & Zimring, 2004), although in nursing homes the evidence is weaker (Schnelle, Alessi, Al-Samarrai, Fricker, & Ouslander, 1999).
Falls prevalence was also hypothesized to be related to private rooms. However, we found no research that specifically linked the prevalence of falls to being in a private versus shared room in nursing homes. There were some suggestions, though no empirical evidence, that placing people who are at a high risk of falls in multibed rooms in hospitals might reduce the occurrence of falls, as roommates could remind individuals not to rise without assistance (Chaudhury et al., 2005; Tutuarima, van der Meulen, de Haan, van Straten, & Limburg, 1997).
Operational Factors
We identified two issues in the literature that relate to operational efficiency: the marketing of shared rooms, and the quality of staff–resident communications. However, empirical studies on both these topics are practically nonexistent, and virtually all of the evidence on this topic comes from interviews, focus groups, and a few descriptive articles. Duffin (2002) and Fisher (1995) both suggest that it is harder to market shared rooms, in part because of gender-matching issues and in part because of a preference for private rooms. However, we found no empirical studies to support these anecdotal descriptions.
Information on the quality of resident–staff communications comes primarily from hospital studies (Berry, 2004; Ulrich & Zimring, 2004). The Healthcare Insurance Portability and Accountability Act regulations, known as HIPAA, mandate the implementation of certain confidentiality procedures. Having a conversation with a resident about private medical matters is much more difficult when there is a roommate in the room, though this issue is certainly more relevant in a hospital setting than a nursing home, where HIPAA concerns are often focused on communication at the nursing station, not in the bedroom.
There were also some references that discussed the positive consequences of shared rooms in terms of staff efficiency, although again this literature was mostly conducted in hospital settings. Chaudhury and colleagues (2004) found that the only dimension that nurses in four hospitals rated private rooms worse than shared rooms was on walking distance from the nursing station. However, this may have as much to do with unit configuration as it does with the percentage of private rooms. Several studies have shown that radial units are much more efficient, from the perspective of walking distance and time spent walking, than corridor designs (Shepley & Davies, 2003; Trites, Galbraith, Sturdavant, & Leckwart, 1970), regardless of bedroom configuration, and these results may be translatable to a nursing home setting.
Building Factors
There are very few empirical studies exploring construction or ongoing building-related costs of nursing homes. The only relevant construction cost analysis that we identified was conducted by Chaudhury and colleagues (2005) of private versus shared rooms in hospitals. They calculated gross floor area per bed (for the whole unit, which includes all shared social spaces and staff support areas), and they estimated construction at $285/ft2 ($285/0.09m2). Using this format, they estimated the cost per patient room at $182,400 per patient in all private room configurations and $122,550 per patient in mixed (some private and some shared room) configurations, suggesting that all private rooms would cost substantially more to construct.
Results of Interviews and Focus Groups
In general, the interviews and focus groups reinforced the information we gleaned from the literature review, and we identified a number of additional topics. In addition, two of the focus group facilities had enhanced shared rooms, which staff felt impacted many of the topics of discussion. We found no mention in the published literature on this room configuration.
Psychosocial Factors
Staff and residents echoed the strong preference for private bedrooms found in the literature. In one facility that had a number of enhanced shared rooms, staff and residents alike said these rooms were perceived more like a "private room with a shared bathroom" than a shared room, with all the benefits thereof. Issues related to visiting appeared to be most critical during the death and dying process. Most family members want to be close to the dying relative but are sensitive to the fact that they are also in someone else's room. Families feel bad for the other resident and the encroachment of their family, and the resident who is not dying is also uncomfortable, having to intrude on what should be a private time for the family. Staff in the focus groups felt that being in a shared room sometimes kept as many family members from gathering or staying as long as they would have preferred.
Clinical Factors
Discussion of clinical factors in the interviews and focus groups related primarily to sleep and falls. Both residents and staff indicated that an individual is more likely to wake up when a staff member enters the room and provides care to a roommate than when the individual lives alone, although this may be mitigated in some enhanced shared rooms, depending on the level of acoustic separation between the residents. This can be a serious disruptor of sleep, because some individuals are checked every two hours. Staff members were uncertain how much of an impact frequently interrupted sleep had on residents the next day. In addition, several staff at different facilities indicated they were sure that there are more falls in shared rooms, though they had no hard data to support this. We identified several other factors as potential clinical outcomes related to private versus shared rooms in the focus groups that were not apparent in the literature, including the use of as-needed (known as PRN) and psychotropic medications, the rate of distressed behaviors by residents (particularly residents with dementia), and medical error rates. However, information on these topics from the focus groups was mixed.
Operational Factors
Not surprisingly, much of the discussion in the interviews and focus groups revolved around operational issues, as these are of primary concern to staff and administrators. Topics included increased time and effort for marketing and admissions, time spent dealing with families, time spent managing conflict, and time spent managing transfers, all of which appear to be greater with shared rooms than private rooms.
Focus group participants agreed with the limited literature about the increased difficulty of marketing shared rooms (which translates into greater costs). None of the focus group facilities had an open bed available in a private room, though there were several openings in shared rooms. When a private room becomes available, staff indicated that it is always filled immediately, often from someone in house who has been waiting. One focus group was held in the nursing home of a retirement community, and staff indicated that residents were leaving the campus to go to a different nursing home rather than move into a shared room, which represents lost income for the facility.
The management of roommate conflict had even greater cost implications. We found no empirical evidence related to the time spent managing roommate conflict in the literature, but the staff in the focus groups indicated that it could be substantial. Estimates of the average time spent (recognizing that on any given week it could be considerably higher) ranged from 2 to 25 hours per week. Apparently, it is not just the social workers and nursing staff who spend time on roommate issues. One housekeeper indicated she spends more time with residents in shared rooms who are upset by something than she spends with residents in private rooms, who seem to be upset less often.
If resolution of differences between residents is not possible, and the decision is made to relocate a resident, there are additional operational costs. Room-cleaning time and maintenance issues are greater at the time of relocation than routine room care is. All furniture must be removed and disinfected, and any maintenance issues (patching walls where personal belongings hung, repainting, and stripping and refinishing the floor) must be addressed. This also causes disruption to the remaining resident, who cannot access his or her room while it is being cleaned. In one facility, this process was estimated to add an additional 90 min of cleaning time over routine cleaning.
All these costs may be further compounded by the fact that, when a building is close to full, there may not be an appropriate empty room available into which the individual who is relocating can move. All facilities indicated that unanticipated resident relocation because of roommate problems can cause a domino effect, requiring one, two, or sometimes up to three other residents to also relocate. Each of these relocations also takes a substantial amount of staff time, as staff members explain to residents and families why it is best for someone, who may be relatively happy in her or his current location, to move. Often people do not want to move, forcing nursing staff to use their authority that it "is in everyone's best interest." This directly contradicts the principles of person-centered or self-directed care, as residents are given little or no choice or control in these situations. The time-management consequences, especially for nursing and social workers, can be substantial, though this remains undocumented. Finally, depending on where the individual(s) are relocated to (i.e., a different unit or household), staff may have to spend additional time getting to know the resident and his or her clinical needs and daily routines and helping the resident adjust to a new roommate. Thus, there are not only operational costs but also negative clinical correlates of this type of move.
We identified a few additional operational correlates in the focus groups. Several housekeepers indicated that private rooms take less time to clean than shared rooms, not just because there are two people in a shared room. In several facilities, housekeepers and direct care staff said that people in private rooms seem to "keep their spaces better." They speculated that there is a greater sense of ownership of the whole room as personal territory in a private room, whereas in a shared room, everything feels like common space, and people don't take as much care of it. There were also some cost factors related to lost income from rehabilitation residents who wanted to be discharged sooner because they were uncomfortable in shared bedrooms. Medicare Part A reimbursement rates are substantially higher, so an early discharge may mean both lost revenue and increased risk for people returning home before they are ready.
Results of the Bedroom Plan Analysis
In this project we conducted an analysis of 189 bedrooms to compare the construction costs of three bedroom configurations: traditional shared, enhanced shared, and private. Table 1 shows the average and range of the size of the three bedroom configurations.
|
Although the costs themselves are clearly higher for a private room, the significance of this difference remains unclear. In a private pay market, there is typically a difference in the cost of a shared bedroom and that of a private room. A large national study found that difference to be $23 ($167 for shared, and $190 for private; see Genworth Financial, 2005). Because there is no revenue data on enhanced shared rooms, we combined the data from the two shared configurations, for an average cost of $22,814 per person for shared rooms. Thus the difference in construction costs between a private and a shared room, per person, is $13,702. If a facility charges $23 more for a private room, the difference in costs (including debt) to construct a private room as opposed to a shared room can be recouped in less than 2 years (596 days). This assumes the shared room has two occupants. If, in fact, a bed remains unoccupied (possibly because potential residents choose to go to a facility that offers private rooms), then the revenue difference is not $23 per day, but $167 (if we assume there is one empty bed). In that case, the time it would take to recoup the cost of constructing a private room drops to 82 days, or less than 3 months. Stated another way, for every 82 resident days below full census, the facility could have built a private room with the lost revenue. After the 82 days, the facility is actually making more money on the private room that it would make on the shared room.
This analysis, of course, is based on the assumption of a cost differential of $23 between a private and a shared bedroom. If a facility is housing people who are on Medicaid, then the cost analysis changes. Generally speaking, Medicaid will not pay extra for a private room, unless it is medically necessary. The state of Michigan, however, has recognized the tremendous benefits of private rooms, and it now includes in their capital cost formula an additional $5 per patient per day for private rooms (up to 100 beds). Even with this minor increase, it would only take a facility 7.5 years to recoup the construction cost differential. If we assume that there is a 30-year mortgage, it means the facility is ahead, financially, for 22 years of the mortgage. This analysis is summarized in Table 2.
|
|
Discussion |
|---|
|
TOPAbstractMethodsDiscussionAppendixReferences |
|---|
One weakness to this analysis is that it was not possible to estimate the associated unit size differences caused by having more private rooms. It is argued that unit or household size and configuration (radial, open plan, hallway plan, or other variation) has a more significant impact on overall unit or household size than the number of private versus shared rooms. A study that expanded the plan analysis to include the whole unit configuration would shed light on this.
There is clearly a need for much more research in this area. Two or three potential topics for each domain of the framework are suggested here. In terms of psychosocial issues, researchers must analyze whether individuals who indicate they are satisfied with a shared room would be more satisfied with a private room if they had the opportunity to experience one. Consideration should also be given to what characteristics (of the individual or the situation) differentiate people who prefer a shared room from a private room. Surprisingly, there was very little information specific to the needs or preferences of people with dementia. In terms of clinical outcomes, the relationship of bedroom configuration to incidence of increased disruptiveness, distress, agitation, or aggression, particularly in individuals with dementia, requires more study. This area, in particular, should focus on the three different bedroom configurations (i.e., it should differentiate between traditional and enhanced shared rooms). There is also a need for greater understanding of the impact of the presence of a roommate on falls, because of the serious morbidity issues associated with falls.
Operational correlates of private versus shared bedrooms are not well addressed in the extant literature, although the focus groups indicated a number of issues worthy of further exploration. The issue with the largest financial impact relates to lost revenue from being unable to fill a shared room when an individual would have agreed to move into a private room. A related topic would be an exploration of the differential costs of marketing a shared room versus a private room versus an enhanced shared room. There is clear, albeit anecdotal, evidence that roommate conflict can occupy a substantial portion of staff time. Although having all private rooms might free up staff time, it will not necessarily reduce costs. The question is what staff members do with this time—whether this translates into better care. The focus groups suggested that maintenance and housekeeping costs are higher per person for shared rooms than for private rooms, but there is no concrete evidence to support this.
On the cost of construction side, an analysis of how unit layout relates to bedroom configuration and therefore costs would be of great benefit to the industry. This might also be tied to staff efficiency studies, such as tracking how much time is spent walking to destinations in units with different layouts.
Across all topics, attention should be given to differentiating between bedroom configurations. The vast majority of studies that we reviewed do not include bedroom configuration as a variable, and none have explored differential impacts of the enhanced shared bedrooms. A more detailed study of this should consider differentiating territory-enhanced rooms, where each person has her or his own territory but spaces are separated by a curtain (and thus lack auditory and olfactory privacy), from privacy-enhanced bedrooms, where each person has, in essence, a private bedroom with a solid door but shares the bathroom.
Currently, the Medicaid program serves as a disincentive to construct private rooms. Private rooms do cost more to construct, and there is, with few exceptions, no additional reimbursement to cover these additional costs. Given the need to control costs, it would not be inappropriate to suggest that additional reimbursement should equal (not exceed) the additional cost of construction plus debt service. An increase of $1.25 a day would cover the costs as assumed in this model in 30 years (the assumed length of the mortgage).
The evidence on preferences, satisfaction, and quality of life for residents living in private rooms in nursing homes is substantial. Virtually all other factors that impact life-cycle costs also trend toward better indicators for private rooms, although there is a need for better evidence to support this. Even the cost analysis suggests that, with a relatively minor increase in reimbursement, the differential construction and capital costs can be recovered. Unfortunately, some providers and designers, and well as the regulators and legislators who control Medicaid budgets, are not yet swayed by this evidence, and they are still building shared rooms. Over the next decade many nursing home buildings will be significantly renovated or replaced. There is a clear need for more evidence-based information, with widespread dissemination efforts, to support making more informed, evidence-based decisions.
|
Appendix |
|---|
|
TOPAbstractMethodsDiscussionAppendixReferences |
|---|
|
Footnotes |
|---|
1 IDEAS Institute, Kirtland, OH.
2 Department of Psychology, College of Wooster, OH.
Decision Editor: Linda S. Noelker, PhD
Received for publication April 13, 2006. Accepted for publication October 18, 2006.
|
References |
|---|
|
TOPAbstractMethodsDiscussionAppendixReferences |
|---|
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
