THE importance of quality of life (QoL)

Nursing Home Staff s Perceived Ability to Influence Quality of Life J Nurs Care Qual Vol. 21, No. 3, pp. 249 256 c 2006 Lippincott Williams & Wilkins, Inc. Robert L. Kane, MD; Todd Rockwood, PhD; Kathryn Hyer, PhD; Karen Desjardins, DrNP, MPH, ANP; Andrea Brassard, DNSc, RN; Charles Gessert, MD, MPH; Rosalie Kane, PhD; Christine Mueller, PhD, RN Nurses, certified nursing assistants, activity personnel, social workers, and physicians in 5 cities rated their ability to affect each of 17 quality of life (QoL) items for 2 hypothetical cases. Those closest to the residents feel the most empowered to make a difference. Overall, certified nursing assistants were consistently the most optimistic about their ability to influence QoL. Perceptions of ability to influence QoL were correlated with attitudes about nursing homes. These perceptions may be helpful in retaining such staff. Key words: magnitude estimation, nursing homes, nursing home staff, preferences, quality of life, values THE importance of quality of life (QoL) of nursing home (NH) residents is beginning to be appreciated. 1 Not only we can measure it, 2 but also there is evidence that changing NH staff behavior can influence it. 3 Because residents QoL is affected by how they are treated, the staff s attitudes about QoL are critical. 4 In order to address QoL, staff members must first believe that they can affect it. This study examines the extent to which NH staff members views about their ability to influence residents QoL are related to their attitudes about NH care. Moreover, From the University of Minnesota School of Public Health, Minneapolis, Minn (Drs Kane, Rockwood, and Kane); the University of South Florida Training Academy on Aging, Tampa, Fla (Dr Hyer); the Columbia University School of Nursing, New York (Dr Desjardins); the Department of Nursing, William Paterson University, Wayne, NJ (Dr Brassard); the St. Mary s/duluth Clinic Health System, Duluth, Minn (Dr Gessert); and the University of Minnesota School of Nursing, Minneapolis, Minn (Dr Mueller). Corresponding author: Robert L. Kane, MD, University of Minnesota School of Public Health, D351 Mayo (MMC 197), 420 Delaware St SE, Minneapolis, MN 55455 (e-mail: kanex001@umn.edu). Accepted for publication: November 10, 2005 because staff members who feel empowered to help residents may be more likely to feel positively about their work environment in general. 5,6 We also explored their views about NH care in general. Specifically, the study addressed 3 questions: 1. How do different types of NH staff members rate their ability to influence QoL? 2. How do these ratings differ with different types of hypothetical residents? 3. How do these ratings correlate with the staff members beliefs about NHs? Previous studies have shown that different types of healthcare professionals involved in the care of NH residents place varying levels of importance of dimensions of QoL. Quality of life ratings for persons with cognitive impairment were consistently lower than those for persons with physical disabilities. Registered nurse (RN) and certified nursing assistant (CNA) ratings were generally higher than those of the other NH staff members, attending physicians, residents, and their families. 7 The underlying conceptual model of this study holds that staff members perceptions about their ability to influence the QoL of the residents they care for is related to 249

250 JOURNAL OF NURSING CARE QUALITY/JULY SEPTEMBER 2006 characteristics of the residents and to their own training and experience. Their perceptions of their abilities to influence QoL are in turn linked to their beliefs about NH care. However, it is not possible to specify the direction of causation. Since most staff members have formed impressions about NH care, and, indeed, these perceptions have been influenced by and perhaps also influenced their decision to work in NHs, one might argue that the direction of causation runs from general attitudes about NHs to perceptions of influence. Those opting to work in NHs may evidence more positive attitudes about such care sites; conversely, work experience may shape attitudes. Different levels of NH staff are subject to different forces. Professional nurses may have a wider choice of working situations than CNAs, and hence those electing to work in NHs may be those who believe in the importance and value of such care; alternately, all NH staff members may simply be settling for a job. Certified nursing assistants have the most contact with residents, and thus may believe more strongly that they can affect their QoL. Resident characteristics can play a role in determining QoL. Staff members may be more inclined to think they can affect the QoL of residents who are more capable of responding; alternatively, they may believe they can affect those who are most vulnerable. Thus, this study did not set out a priori hypotheses about these relationships. Rather, this was an exploratory study, exploring what those relationships might be. METHODS This research was conducted in 4 cities in 3 different states: Florida, New Jersey, and Minnesota. (In Minnesota, we used 2 cities.) A site coordinator in each city selected NHs from which participants for the research were recruited. Most of the participating NHs had previously participated in the parent QoL study. 2 The city coordinator recruited a targeted number of individuals from 4 different groups who worked, resided, or had a family member residing in the nursing facilities. For each city, we sought to enroll 7 of each of the following: (1) RNs and licensed practical nurses (LPNs) (RNs/LPNs), (2) social workers and/or activity therapists (SWs/ATs), (3) physicians (MD), and (4) CNAs. To be eligible, physicians had to spend some part of their practice in a nursing facility. The selection of facilities and subjects for this research was done using convenience sampling procedures. Site coordinators selected the NHs that they collected data in, based on past relationships and ease of access. A self-administered questionnaire was used for all respondents. Participants were allowed 2 different methods of responding. They could complete the survey at the facility and return it to the site coordinator or they could take the survey home and return it by mail to the University of Minnesota. The majority of the surveys were completed at work and returned to the site coordinator. The survey had 2 sections. In one section, the provider was asked to rate his or her ability to affect (influence change for) each of 17 QoL items on a 10-point scale (1 = no influence, 10 = a great deal of influence) for 2 different resident types: (1) an NH resident who was physically impaired and cognitively intact and (2) a resident who was physically intact and cognitively impaired. The ordering of the 17 items for each resident type, as well as the order in which the scenarios were presented, was randomly generated for each questionnaire. The second section had 22 questions regarding a range of attitudes, beliefs, and perceptions about NHs. When these items were analyzed to determine whether underlying latent variables could be identified, the analysis generated 4 latent variables, shown in Table 1, which became subscales for subsequent analyses. The construct labeled place is composed of 5 items (α =.69) that include issues about NHs as places for care as well as perceptions about NHs as places (eg, warehouses). The work construct has 4 items A latent variable is one that cannot be assessed directly. It is assumed to underlie the constructs being measured.

Nursing Home Staff s Perceived Ability to Influence QoL 251 Table 1. Latent variable development for nursing home evaluation items Reverse coded. Response scale: 1 = strongly agree; 2 = somewhat agree; 3 = somewhat disagree; 4 = strongly disagree. (α =.62). Two of the items focus on work issues and the other 2 focus on residents. The third construct focuses on the outcomes associated with care provided in NHs (outcome, α =.60). These items focus on perceptions about the ability to influence outcomes and the relationship between NHs and outcomes. The final construct is residents (α =.60). The 3 items contributing to this construct focus on attributes and characteristics of NH residents. Finally, the raw data were screened to look for evidence of respondents who rated all items the same, especially those who rated everything with a 10 within each scenario. For scenario 1 (resident who was physically impaired and cognitively intact), 11 (6%) of the respondents answered with all 10s (great deal of influence: 10% of CNAs, 4% of RNs/LPNs, 4% of SWs/ATs, but none of the MDs). For scenario 2 (resident who was physically intact and cognitively impaired), 22 (12%)

252 JOURNAL OF NURSING CARE QUALITY/JULY SEPTEMBER 2006 Table 2. Mean (SD) ability to influence QoL for physically intact cognitively impaired residents Item CNA RN/LPN MD SW/AT F K-W Comfort 9.0 (1.5) 9.4 (1.1) 8.5 (1.6) 7.5 (2.1).01.01 Staff deal with pain 9.1 (1.3) 9.4 (1.0) 8.9 (1.4) 7.5 (2.1).01.01 Independence 8.7 (1.6) 8.4 (1.5) 6.7 (2.5) 7.4 (2.1).01.01 Privacy 9.0 (1.4) 9.0 (1.4) 7.4 (2.4) 8.2 (2.1).01.01 Choice 9.1 (1.3) 8.8 (1.6) 7.4 (2.6) 8.4 (2.3).01.05 Dignity 8.5 (1.8) 7.6 (1.9) 6.3 (2.5) 7.7 (1.8).01.01 Pleasurable activity 8.1 (1.9) 7.2 (2.0) 6.2 (2.6) 7.7 (2.1).01.01 Food and dining 8.4 (1.9) 7.1 (2.4) 6.0 (3.0) 6.8 (2.6).01.01 Do what they want 8.1 (1.9) 7.0 (1.9) 6.0 (2.5) 7.4 (2.0).01.01 Identity 8.5 (1.8) 7.8 (2.1) 6.0 (2.4) 7.5 (2.2).01.01 Relationships 8.4 (2.0) 7.6 (1.8) 6.1 (2.6) 7.5 (2.0).01.01 Personal safety 8.9 (1.5) 8.6 (1.5) 6.4 (2.7) 7.7 (2.2).01.01 Possessions safe 8.3 (2.0) 7.9 (2.1) 5.7 (2.7) 6.6 (2.8).01.01 Spiritual needs met 8.1 (2.1) 7.0 (2.2) 5.9 (2.9) 8.3 (1.9).01.01 Values respected 9.0 (1.5) 8.5 (1.9) 7.0 (2.8) 8.0 (2.3).01.01 Not belittled 8.8 (1.7) 8.7 (1.7) 7.0 (2.8) 8.4 (2.5).01.05 Anxiety/boredom 8.0 (1.9) 7.7 (1.9) 5.8 (2.7) 7.3 (2.3).01.01 QoL indicates quality of life; CNA, certified assistant; RN/LPN, registered nurse/licensed practical nurse; MD, physician; SW/AT, social worker/activity therapist; and K-W, Kruskal-Wallis test. Probability is for F statistic (Proc GLM in SAS). Probability is for Kruskal-Wallis test (Proc NPARWAY in SAS). Mean value is significantly different from CNAs on the basis of Bonferroni corrected t test. respondents answered with all 10s: 19% of CNAs, 7% of RNs/LPNs, 5% of MDs, and 8% of SWs/ATs. (These cases have been retained in the analysis; separate analyses showed that their inclusion did not alter the study results.) SAS software was used for the data analysis. Several types of analyses were performed: analysis of variance (ANOVA) (Proc GLM) and ordinary least squares (OLS) regression (Proc REG). We also used the Kruskal-Wallis technique to explore where differences occurred in the ANOVA models (Proc NPARWAY). A Bonferroni correction for multiple comparisons was applied when the resident rating was compared with each of the professional ratings. In the analysis of the interaction between the item content and provider group, 2 different data structures were used. The first analysis used a typical analytic database structure; in the second, the data were standardized. This normalization of the data allowed us to look at 3 issues: (1) the effect of provider type for each item, controlling for scenario; (2) the effect of scenario for each item, controlling for provider; and (3) an evaluation of provider type and perceptions of NHs. RESULTS The final participant pool included 26 RNs, 28 LPNs, 91 CNAs, 21 MDs, 12 SWs, and 14 activity directors. In general, females dominated all of the provider populations; the only exception was physicians (50% male). The average age in all of the provider groups was in the 40s, with CNAs on average being the youngest (41) and RNs/LPNs and MDs being the oldest (48). The response rates were 96% for RNs/LPNs and CNAs, 93% for SWs/ATs, and 75% for MDs. Tables 2 and 3 present the NH staff members ratings of their ability to influence items

Nursing Home Staff s Perceived Ability to Influence QoL 253 Table 3. Mean (SD) rating across all QoL items for each professional group controlling for resident scenario Item CNA RN/LPN MD SW/AT F K-W Comfort 9.1 (1.5) 9.4 (1.0) 8.7 (1.4) 7.5 (2.2).01.01 Staff deal with pain 9.2 (1.5) 9.5 (0.9) 9.0 (1.3) 7.8 (2.1).01.01 Independence 8.9 (1.6) 8.5 (1.6) 7.0 (2.3) 7.8 (1.9).01.01 Privacy 9.3 (1.2) 9.2 (1.2) 7.6 (2.3) 8.4 (1.9).01.01 Choice 9.2 (1.2) 9.1 (1.4) 7.6 (2.3) 8.8 (1.9).01.01 Dignity 8.6 (1.7) 7.8 (1.8) 6.8 (2.6) 8.1 (1.5).01.01 Pleasurable activity 8.3 (1.8) 7.6 (1.9) 6.5 (2.6) 8.2 (1.8).01.01 Food and dining 8.5 (1.8) 7.3 (2.2) 6.1 (3.0) 7.0 (2.4).01.01 Do what they want 8.4 (1.8) 7.4 (1.9) 6.3 (2.7) 7.9 (1.8).01.01 Identity 8.7 (1.7) 8.3 (1.8) 6.7 (2.5) 8.1 (2.0).01.01 Relationships 8.4 (2.1) 7.9 (1.8) 6.4 (2.7) 8.2 (1.8).01.01 Personal safety 9.0 (1.4) 8.8 (1.5) 6.7 (2.6) 7.9 (2.0).01.01 Possessions safe 8.5 (1.9) 8.2 (1.8) 6.1 (2.7) 7.3 (2.3).01.01 Spiritual needs met 8.3 (2.1) 7.3 (2.3) 6.4 (2.9) 8.6 (1.7).01.01 Values respected 9.1 (1.4) 8.8 (1.6) 7.4 (2.5) 8.4 (2.0).01.01 Not belittled 8.9 (1.6) 9.0 (1.4) 7.5 (2.5) 8.6 (2.0).01.01 Anxiety/boredom 8.2 (1.8) 7.8 (1.9) 5.9 (2.7) 7.6 (2.0).01.01 QoL indicates quality of life; CNA, certified assistant; RN/LPN, registered nurse/licensed practical nurse; MD, physician; SW/AT, social worker/activity therapist; and K-W, Kruskal-Wallis test. Probability is for F statistic (Proc GLM in SAS). Probability is for Kruskal-Wallis test (Proc NPARWAY in SAS). Mean value is significantly different from CNAs on the basis of Dunnett-Dunn corrected t test. for 2 different hypothetical residents. Table 2 shows the ratings for scenario 1 a physically impaired and cognitively intact resident. Among the 4 provider groups, ANOVA of the ratings demonstrates significant differences for all 17 items for both resident types. With both scenarios, CNAs consistently perceived that they had more influence in more of the QoL areas than any of the other NH providers, followed by the licensed nursing staff. Compared with the CNAs, MDs perceived that they had less ability to influence all of the items, except for the 2 items that focus on pain and discomfort. In addition, for the hypothetical resident with physical impairments and no cognitive impairment, MDs were less positive about independence as well. The SWs/ATs were lower than the CNAs on these 3 items, as well as personal safety and food and dining. They also differ on other selected items across the cases compared with the other NH providers for each resident type. The analysis in Table 3 pools the 2 scenarios to focus on differences among the provider groups. The differences are all significant. The MDs and social workers view themselves as having less influence on QoL compared with the CNAs on a variety of items. Table 4 shows the impact of scenarios physically intact cognitively impaired compared with physically impaired cognitively intact, controlling for provider type and the interaction between provider type and scenario. For 11 of the 17 items, the perceived ability to influence the item is significantly greater for the resident with cognitive impairment scenario than the resident with physical impairment scenario.

254 JOURNAL OF NURSING CARE QUALITY/JULY SEPTEMBER 2006 Table 4. Mean (SD) rating for each QoL item across scenarios controlling for raters (average of all professional groups for each of the scenarios) Item Physically impaired Cognitively impaired P Comfort 8.8 (1.6) 9.0 (1.6) NS Staff deal with pain 9.0 (1.5) 9.1 (1.5) NS Independence 8.2 (1.9) 8.6 (1.7) NS Privacy 8.7 (1.7) 9.1 (1.4).05 Choice 8.7 (1.8) 9.1 (1.4).01 Dignity 7.9 (2.0) 8.3 (1.8).05 Pleasurable activity 7.6 (2.1) 8.2 (1.9).01 Food and dining 7.5 (2.4) 7.8 (2.3) NS Do what they want 7.4 (2.1) 8.1 (2.0).01 Identity 7.9 (2.2) 8.6 (1.7).01 Relationships 7.8 (2.1) 8.2 (2.1).05 Personal safety 8.3 (2.0) 8.7 (1.7) NS Possessions safe 7.6 (2.4) 8.3 (1.9).01 Spiritual needs met 7.6 (2.4) 8.0 (2.2).05 Values respected 8.4 (2.0) 9.0 (1.5).05 Not belittled 8.5 (2.1) 9.0 (1.4).05 Anxiety/boredom 7.5 (2.2) 7.9 (2.0) NS QoL indicates quality of life; NS, not significant. Probability is for t test (Stepdown Sidak correction using Proc MULTTEST in SAS). [AQ1] Table 5 shows the results of a regression analysis to examine the relationship between the 4 attitudinal factor scores and the staff members perceptions of their ability to influence QoL in NHs, adjusting for profession and scenario. Only the relationships that were significant at P <.05 are shown. The factor involving NH residents, which is related to scenarios, is least likely to be related to perceptions of ability to affect QoL, whereas the factors relating to outcomes (which is most closely related to perceived ability to influence QoL) and place are most often related to perceptions of influence. No QoL element is unrelated to at least one factor. When we examined the scenarios separately, the pattern for residents who were physically intact but cognitively impaired differed little. Privacy, Do what they want, and Relationships were related to Home and Outcome. Values respected was related to Place and Outcome. In each case, perceived influence was correlated with more positive feelings about NHs. DISCUSSION The 4 types of NH staff providers differ considerably regarding their perceived ability to influence the 17 elements associated with QoL for both physically intact cognitively impaired and physically impaired cognitively intact residents, but their perceived influence is less for residents with cognitive impairment. Physicians are the most pessimistic, whereas CNAs are generally the most optimistic about their ability to influence QoL. In general, the licensed nursing staff and MDs feel more empowered to affect residents pain and discomfort, privacy, dignity, and sense of not being humiliated than they do arranging meaningful activities, making

Nursing Home Staff s Perceived Ability to Influence QoL 255 Table 5. Regression analysis evaluating the relationship between QoL elements and perceptions of nursing homes Factor score Item Place Home Outcome Residents Comfort 0.21 0.17 Staff deal with pain 0.17 0.26 Independence 0.11 Privacy 0.16 0.12 0.13 Choice 0.15 0.20 Dignity 0.15 Pleasurable activity 0.18 Food and dining 0.18 Do what they want 0.13 0.12 Identity 0.14 0.16 Relationships 0.14 0.12 Personal safety 0.21 0.14 Possessions safe 0.19 0.11 Spiritual needs met 0.14 Values respected 0.20 0.14 Not belittled 0.19 0.17 Anxiety/boredom 0.20 0.13 QoL indicates quality of life. Only variables significant at P <.05 are shown. The analysis is adjusted for professionals and scenarios. the dining experience pleasant, and avoiding boredom; however, there is no obvious reason why some of these areas should be more readily influenced than others. Nursing home staff has the potential to influence all these parameters of daily life; those in more regular contact with the residents, namely the CNAs, have the greatest opportunity to influence their QoL. The finding that all professionals felt more empowered to help residents with cognitive impairment is somewhat surprising. In related work, cognitive impairment was associated with lower QoL importance ratings. Perhaps, residents with cognitive impairment are seen as needing more or being more amenable to staff attention. Staff members may view cognitively frail residents as more vulnerable, and thus more easily affected by staff members actions. The professionals attitudes about NHs as places for care, work-related issues, ability to influence outcomes, and attributes of residents are significantly related to their perceptions of their ability to affect the QoL of NH residents. As noted earlier, it is hard to say in which way the causal arrow should go. Those with more positive attitudes regarding the 4 factors may be more likely to select nursing work, and hence to feel they can make a difference. Conversely, those who feel they can make a difference may derive more satisfaction from their work and feel more positive about it. Before professionals can be motivated to change their approach to NH care, they must believe that such changes can make a difference. A portion of these beliefs is directly related to the way they view nursing care in general. Physicians seem to be the most

256 JOURNAL OF NURSING CARE QUALITY/JULY SEPTEMBER 2006 pessimistic, perhaps because they are not as directly involved in daily NH life. Previous calls for improving NH care have looked to physician leadership. 8 They may not be in as good a position as hoped to provide the impetus for reform until they can believe that their efforts will be more successful. In contrast, the CNAs, who work closest to the NH residents, see themselves as having more impact on these residents QoL than other staff members. Therefore, they may be more receptive than expected to efforts designed to improve care and QoL for residents. This finding also underscores the importance of involving CNAs in the assessment and care planning of residents. The CNAs insights into the residents needs and preferences and their belief that they can positively influence residents QoL should be capitalized on in developing strategies to improve quality of care and life for residents. The professional nurses lie in between the physicians and CNAs. Their enthusiastic support will be needed as they are accountable for the care provided to residents. Some limitations of this study must be acknowledged. The sample is a convenience sample, and hence generalization is limited. Although the sample size is small, it has allowed for the analyses to be undertaken. There was some indication of a modest response pattern to overuse the top rating, but this did not appear to influence the results. Implications These findings offer some hope about how to make the working situations for nursing personnel more rewarding. It is unlikely that society will ever be willing to pay them their full worth. Those who opt for this work do so in large measure because they derive a sense of service from their work. We should build on the CNAs demonstrated belief that they can make difference by designing approaches to collecting and displaying information that can show just how much of difference they make. At present, most of NH care is associated with residents declining over time. Such evidence might also inspire more professional nursing staff to appreciate the difference that good care can make. The evidence of effective care is less often found in changing that pattern compared with slowing the rate of decline (be it in QoL or functional outcomes). Unfortunately, the current approaches to documentation fail to capture the gap between what would have happened in the absence of good care. Models for displaying the effects of good care by comparing observed and expected outcomes are available and can be constructed from measures like the minimum data set. Efforts to use this approach will not only improve staff morale but can also be used in the aggregate to make a stronger case for investing in better long-term care. REFERENCES 1. Kane RA. Definition, measurement, and correlates of quality of life in nursing homes: towards a reasonable practice, research, and policy agenda. Gerontologist. 2003;43(special ed II):28 36. 2. Kane RA, Kling KC, Bershadsky B, et al. Quality of life measures for nursing home residents. J Gerontol A Biol Sci Med Sci. 2003;58A(3):240 248. 3. Cox CL, Kaeser L, Montgomery AC, Marion LH. Quality of life nursing care. An experimental trial in long-term care. J Gerontol Nurs. 1991;17(4):6 11. 4. McGilton KS. Enhancing relationships between care providers and residents in long-term care. Designing a model of care. J Gerontol Nurs. 2002;28(12):13 21. 5. Lerner MJ, Simmons CH. Observer s reaction to the innocent victim :compassion or rejection. J Pers Soc Psychol. 1966;4:203 210. 6. Chou S-C, Boldy DP, Lee AH. Staff satisfaction and its components in residential aged care. Int J Qual Health Care. 2002;14(3):207 217. 7. Kane RL, Rockwood T, Hyer K, et al. Rating the importance of nursing home residents quality of life. JAm Geriatr Soc. 2005;53(12):2076 2082. 8. Ouslander JG, Osterweil D, Morley J. Medical Care in the Nursing Home. New York: McGraw-Hill; 1991. [AQ2]

Title: Nursing Home Staff s Perceived Ability to Influence Quality of Life Authors: Robert L. Kane, Todd Rockwood, Kathryn Hyer, Karen Desjardins, Andrea Brassard, Charles Gessert, Rosalie Kane, and Christine Mueller Author Queries AQ1: Should the sentence No QoL element is unrelated to at least one factor be changed to Each QoL element is related to at least one factor. AQ2: Reference updated per PubMed. OK?