Front-line management, staffing and nurse doctor relationships as predictors of nurse and patient outcomes. A survey of Icelandic hospital nurses

Transcription

1 Available online at International Journal of Nursing Studies 46 (2009) Front-line management, staffing and nurse doctor relationships as predictors of nurse and patient outcomes. A survey of Icelandic hospital nurses Sigrún Gunnarsdóttir a, *, Sean P. Clarke b,c, Anne Marie Rafferty d, Don Nutbeam e a Landspitali University Hospital, Office of the Chief Nursing Executive, Eiríksgata 19, Reykajvik, Iceland b Associate Director, Centre for Health Outcomes and Policy Research, Philadelphia, USA c Assistant Professor of Nursing, School of Nursing, University of Pennsylvania, Philadelphia, USA d Professor and Head, Florence Nightingale School of Nursing & Midwifery, Kings College London, UK e Pro-Vice-Chancellor and Head, College of Health Sciences, University of Sydney, Australia Received 15 March 2006; received in revised form 8 November 2006; accepted 28 November 2006 Abstract Objective: To investigate aspects of nurses work environments linked with job outcomes and assessments of quality of care in an Icelandic hospital. Background: Prior research suggests that poor working environments in hospitals significantly hinder retention of nurses and high quality patient care. On the other hand, hospitals with high retention rates (such as Magnet hospitals) show supportive management, professional autonomy, good inter-professional relations and nurse job satisfaction, reduced nurse burnout and improved quality of patient care. Methods: Cross-sectional survey of 695 nurses at Landspitali University Hospital, Reykjavík. Nurses work environments were measured using the nursing work index revised (NWI R) and examined as predictors of job satisfaction, the Maslach burnout inventory (MBI) and nurse-assessed quality of patient care using linear and logistic regression approaches. Results: An Icelandic adaptation of the NWI R showed a five-factor structure similar to that of Lake (2002). After controlling for nurses personal characteristics, job satisfaction, emotional exhaustion and nurse rated quality of care were found to be independently associated with perceptions of support from unit-level managers, staffing adequacy, and nurse doctor relations. Conclusions: The NWI R measures elements of hospital nurses work environments that predict job outcomes and nurses ratings of the quality of patient care in Iceland. Efforts to improve and maintain nurses relations with nurse managers and doctors, as well as their perceptions of staffing adequacy, will likely improve nurse job satisfaction and employee retention, and may improve the quality of patient care. # 2006 Elsevier Ltd. All rights reserved. Keywords: Work environment; NWI R; Front line management; Staffing; Job satisfaction; Burnout; Quality of care What is already known about the topic? * Corresponding author. Tel.: address: sigrungu@landspitali.is (S. Gunnarsdóttir). Supportive work environments are important for the quality of working life for nurses, and improved patient outcomes /$ see front matter # 2006 Elsevier Ltd. All rights reserved. doi: /j.ijnurstu

2 S. Gunnarsdóttir et al. / International Journal of Nursing Studies 46 (2009) There are gaps in the literature about the link between nurse work environment and nurse and patient outcomes, notably in countries other than the United States. What this paper adds Re-emphasizes the important role of front line nurse managers and nurse doctor relationships for successful hospital care. Provides further support for the idea that perceptions of staffing adequacy are linked to perceptions of quality. The NWI R was successfully adapted into Icelandic and resulted in empirically and conceptually acceptable subscales that demonstrated all expected associations. 1. Introduction Health policy experts and nurse leaders have recognized that a deepening global nurse shortage is leading to new challenges for health systems all over the world (OECD, 2005). Poor work environments for nurses (including inadequate staffing levels) are associated with nurse job dissatisfaction and burnout, which in turn often increase turnover rates (Aiken et al., 2001; Bauman et al., 2001; Hayes et al., 2006). Inadequate staffing levels, in turn, are associated with increases in adverse events in hospital care (Aiken et al., 2002b; Estabrooks et al., 2005; Lang et al., 2004; Needleman et al., 2002; Tourangeau et al., 2006). Since the current shortage of nurses (and particularly, hospital nurses) is a major international problem, identifying methods for creating more supportive nurse work environments for nurses across health care systems and cultures is critical (McClure and Hinshaw, 2002). High-quality patient care is the goal of nurses and their employers alike. However, many aspects of hospital work, such as work processes, staffing and workload issues, and irregular work hours create stresses for nurses. Research suggests that supportive leadership behaviour and management approaches have positive impacts on nurse job satisfaction and the prevention of burnout (Aiken et al., 2001; Laschinger and Finegan, 2005; Laschinger et al., 2001). Recent research also shows that patient care quality improves with higher nurse educational level, richer nurse skill mix and lower proportions of temporary staff and when nurses and doctors have good relationships (Estabrooks et al., 2005; Tourangeau et al., 2006). Furthermore, research shows that patient care quality improves when managers collaborate closely with and foster clinical autonomy in their nursing staffs (Laschinger et al., 2001; Nedd, 2006). Among the best-known model for effective management in hospital nursing is the Magnet hospital model. Scholarly writing about the Magnet model has identified a connection between supportive nurse manager and executive leadership and positive nurse and patient outcomes (Aiken, 2002; Kramer and Schmalenberg, 2002). Based initially on descriptive studies about organizations that were particularly successful in recruitment and retention of nurses (McClure et al., 2002) writing in this area now also draws upon research on organisational empowerment (Laschinger et al., 2001) and transformational leadership (Bass, 1998) more broadly. The roles of leadership behaviour in the apparent success of Magnet organizations have not yet been fully explored (Övretveit, 2004) and there are still methodological problems to be solved in this area of inquiry (Buchan, 1999; Kramer and Schmalenberg, 2003, 2004). Also unresolved is the question of the extent to which the elements associated with the Magnet concept are applicable in countries with cultures and health care systems outside the United States (McClure and Hinshaw, 2002). Iceland is a Nordic welfare state with a population of around 300,000. By a number of standards, Icelanders enjoy superior health status and quality of life (Halldorsson, 2003). Nursing is an autonomous profession under Icelandic law (Health Services Act, 1990), and a high proportion (70%) of Icelandic nurses workforce hold baccalaureate or higher degrees in nursing, which compares favourably to the educational attainment of the nurse workforce in other developed countries. Icelandic nursing education, research and administration draw from traditions in Scandinavia, Western Europe and North America. However, Iceland s unique culture and health care system are different enough that it provides opportunities to test ideas regarding nurse work environments seen in other countries. Interestingly, no prior research has been conducted about the impact of nurses work environments on outcomes in Iceland. On the basis of the literature on determinants of nurse and patient outcomes and in the context of Icelandic hospital nursing it was hypothesized that there would be a positive association between nurses perception of supportive work environmental factors and better nurse and patient outcomes. This paper describes a survey of nurses conducted at Landspitali University Hospital, Reykjavík (LSH), a 900- bed hospital that is the largest tertiary health centre in Iceland. LSH, like many other hospitals around the world, is experiencing a shortage of qualified and experienced nurses (Sigurdardóttir et al., 1999). The development and validation of an Icelandic adaptation of the nursing work index revised (NWI R), a tool used extensively in research and administrative practice internationally, is described. Also discussed are findings that the work environment elements shown in prior work to be associated with important nurse and nurse-reported outcomes are similarly linked with these variables in Icelandic nurses. 2. Methods 2.1. Participants Subjects in this cross-sectional study included nurses working in direct patient care in all specialities at the Landspitali University Hospital, Reykjavik (LSH) including medicine, surgery, intensive care, emergency care, pedia-

3 922 S. Gunnarsdóttir et al. / International Journal of Nursing Studies 46 (2009) trics, psychiatry, obstetrics and geriatric care. All eligible nurses were approached to participate in the study, a total of 930 nurses working in 98 clinical wards. Nurses who held more than 40% of a full-time equivalent position were eligible for the study. Nurses on maternity leave, extended sick leave or study leave were excluded Measures Subjects completed an eight-page questionnaire battery identical to that used in the UK and the US in the International Hospital Outcomes Study (IHOS; including the NWI R, and measures of burnout, job satisfaction, nurse-rated quality of care, as well as demographic questions) (Aiken et al., 2001; Aiken and Patrician, 2000). The questionnaire was carefully adapted for use in Iceland and translated into Icelandic by a panel of bilingual experts. It was subsequently back translated and pilot-tested before the survey was fielded (Hilton and Skrutkowski, 2002). Changes to the English version included slight changes in the phrasing of questions and additional nurse specialities to make the questionnaire relevant for the context of Icelandic nurses. Neither the ward on which nurses worked, nor their marital status were included, because of concerns about anonymity in this small population Nursing work index revised Various aspects of nurses work environments, the major independent variable of interest in this study, were measured with an Icelandic linguistic and cultural adaptation of the NWI R. The NWI R was originally developed for research on the attributes of institutions with excellent recruitment and retention records (i.e. Magnet hospitals) (Kramer and Schmalenberg, 2004). The most recent version of the NWI was used for the present study (the NWI R) (Aiken and Patrician, 2000). Among the adaptations made to improve its relevance to LSH nurses involved specific terms for senior management and quality assurance programs. Respondents were asked to rate each of 52 items on a 4-point scale to indicate the degree to which they agreed that specific work environment attributes were present in their current job. The scales were anchored with strongly agree and strongly disagree options, with higher scores representing stronger agreement Dependent measures Analyses of three dependent measures are presented here: job satisfaction, burnout, and nurse-rated quality of care. Nurses were asked a single question that inquired about their satisfaction with their current jobs. This single item job satisfaction question has been used previously in a large international nurse research program (Aiken et al., 2001) and is supported by studies indicating that single items are useful in measuring job satisfaction as a global construct (Wanous and Hudy, 2001). Burnout was measured using the Maslach Burnout Inventory (MBI), a well-validated standardized instrument widely used in health care research (Maslach et al., 1996). Respondents were asked to complete the entire tool. Factor analysis of the items in the Icelandic translation showed the expected factor structure (three subscales, including the depersonalization and personal accomplishment subscales). However, emotional exhaustion, the 9-item subscale most commonly used in organizational research, was analysed here. The emotional exhaustion subscale address aspects of the burnout syndrome that nurses and others are most familiar with (Maslach et al., 1996) and this subscale has been the most consistently linked with organizational factors in previous research (Maslach et al., 1996; Aiken et al., 2002a). Scores range from 0 to 54, with higher scores suggesting worse depletion of physical and emotional energy. Cronbach s alpha for this subscale in the present data set was Nurses were asked also to rate the general quality of nursing care provided on their wards on four-point scale from excellent to poor. This question was adapted from the IHOS and has been used in a number of reports from that study (Aiken et al., 2001) Data collection Questionnaires with letters of information attached were distributed to the clinical wards by the principal investigator and handed to the study subjects or to a contact person for the study at each ward. Before data collection nurses were informed about the study through the hospital newsletter and in presentations at staff meetings. Data collection took place over a 3 months period from September to December A 75% response rate (695 of 930 nurses approached) was achieved. A number of techniques to maximize response rates were used regarding the design of questionnaire and for the recruitment of study participants (Edwards et al., 2002). Among these were professional design and the use of twocolour printing for the survey, a user-friendly format, and clear instructions for returning the questionnaires in brown envelopes using the hospital s internal mail system Validity and reliability The study used well-established and previously-validated measures and previously validated measured. Attempts were made to ensure reliability by carefully translating, pre-testing and piloting the survey instrument and data collection procedures (Edwards et al., 2002; Polit and Hungler, 1999) and rigorous analysis. Response rate was high (75%). Sufficient reliability for all scales was supported Ethical consideration Participation in the study was entirely voluntary and the questionnaires were anonymous and confidential, with no ID numbers or codes. The protocol for the study was reviewed and approved by The National Bioethics Committee in Iceland and the Research Ethical Committee at the London School of Hygiene and Tropical Medicine.

4 S. Gunnarsdóttir et al. / International Journal of Nursing Studies 46 (2009) Data analysis Statistical analyses were performed using SPSS for Windows, Version The factor structure of the Icelandic version of the NWI R was examined using principal axis factoring with oblique rotation. Direct oblique rotation was used instead of varimax rotation since the items could not be assumed to cluster in theoretically independent factors (Tabachnick and Fidell, 2001). Visual inspection of the scree plot was used to determine the final number of factors in the solution. Items with high missing values (>10%) and with low loadings (<0.3) or high loadings on two factors were dropped from the final structure (Reise et al., 2000). Internal reliability coefficients were calculated for the derived subscales and means and standard deviations were reported. Next, nurses job outcomes, and their ratings of the quality of patient care on their units were analysed descriptively. Lastly, to examine the proposed hypothesis, regression models predicting the effects of work environmental factors on each outcome measure were fitted. In the regression analyses the five NWI R measures were put in individually, and then simultaneously, as potential predictors of job satisfaction, burnout scores, and high ratings of quality care. In both sets of models, a series of control variables including nurse demographics and job characteristics (i.e. age, education, family, health, fulltime work, shift work, job title, nurse speciality) were entered as a block in an attempt to control for confounders in determining the potential impact of work environmental factors. This procedure was repeated for each of the three outcome measures (i.e. job satisfaction, burnout measure and patient quality). Results are reported for each factor in relation to each outcome individually. A second set of models, indicating the predictive strength of each factor when all work environment characteristics were statistically controlled, was intended to examine whether any of the specific work environment factors were distinctive in their association with the outcomes beyond their properties as general indicators of perceptions of the quality of the work environment. Odds ratios for reporting very high nurse job satisfaction and excellent quality of patient care were computed from logistic regression models. For the emotional exhaustion (burnout) scores were treated as continuous dependent variable, and beta coefficients for the impacts of work environment factors were calculated from linear regression models. Variables were screened for multicollinearity before being entered simultaneously in models (Tabachnick and Fidell, 2001). The statistical significance ( p) level for the analyses was set at alpha= Results 3.1. Sample characteristics Of the final sample of 695 nurses working with direct nursing in all specialities at LSH, a majority of the study participants (64.2%) were between 31 and 50 years of age, 22.1% were older than 51 years, 13.6% were between 20 and 30 years of age, and 5.5% had a master s degree. A review of hospital records (LSH, 2003) indicated that the study cohort was representative for nurses at LSH as a whole with respect to the distribution of ages and nurse specialities. A large majority of the participants worked full-time (81%) and had worked at the hospital for more than 6 years (80%) Structure of the NWI R (Table 1) The factor analysis of the NWI R items revealed five factors that clustered around the following themes: nurse doctor relations, unit-level support, staffing, philosophy of practice, and hospital-level support. Table 1 lists the factor loadings for items on the primary factors and indicates that the majority of items showed loadings over 0.5 for the scales they were assigned to. The final number of items in each scale ranged from 4 to 9 items. Items in the nurse doctor relations subscale refer to collaboration, good working relationships, teamwork and the quality of care given by physicians. The unit-level support subscale includes items that relate to the support of nurses by front-line managers, praise, recognition and active orientation and staff development programs. The third subscale on staffing contains items relating to number of nurses and staff, adequate support services and time to discuss issues in patient care with other nurses. Items in the philosophy of nursing subscale relate to written nursing plans and nursing diagnoses, clear nursing philosophy and opportunities to work on specialized units. The hospital-level support subscale includes items dealing with support from senior management, involvement of nurses in hospital affairs, quality assurance programs and support for nurses pursuing degrees in nursing (see Table 1). Subscale scores were found to be correlated with each other, but not at a level that created problems with multicollinearity (Pearson coefficients of ) Descriptive findings Table 2 shows that nurses in this study rated items dealing with positive relations with doctors the most highly (i.e. as being the most present) of all in the tool. Following closely were the subscales dealing with support at the unit level and philosophy of nursing practice with mean item scores averaging 3 or lower on a four-point rating scales. The work environment aspect that received the lowest ratings was support from senior management at the hospital level, with scores averaging a little over 2 on the four-point scale. The mean emotional exhaustion scale score of 13.7 (SD 7.8) suggests that nurses reported relatively low levels of burnout relative to published scores for nurses in other countries (Edwards and Bernard, 2003; Jamal, 2000). Table 3 further shows that the majority of nurses (82%) reported being very satisfied or moderately satisfied with their jobs. The great majority of the nurses (95%) rated the quality of patient care as excellent or good, with the modal response (69%) being

5 924 S. Gunnarsdóttir et al. / International Journal of Nursing Studies 46 (2009) Table 1 Factor analysis a of the NWI R items Factors (NWI R sub-scales) Loadings b 1. Nurse Physician Relations (4 items) Collaboration between nurses and 0.81 physicians Physicians and nurses have good working 0.71 relationships A lot of nurses and physicians team work 0.60 Physicians give high quality of care Unit-level support (8 items) Ward management supportive of nurse 0.76 Ward manager good manager and leader 0.73 Ward manager backs up nurses in decision 0.65 making Praise and recognition for a good job 0.52 Active staff development education 0.42 programs Good induction program 0.36 Support for new and innovative ideas about 0.33 patient care Flexible shift patterns are available Staffing (4 items) Enough registered nurses to provide 0.87 quality patient care Enough staff to get work done 0.74 Adequate support service allow me to 0.54 spend time with my patient Enough time and opportunity to discuss 0.52 problems with other nurses 4. Philosophy of practice (5 items) Written nursing plans for all patients 0.65 Use of nursing diagnosis 0.51 Nursing care is based on a nursing rather 0.46 than medical model A clear philosophy of nursing throughout 0.45 the patient care environment Opportunity to work on a highly 0.32 specialized patient care ward 5. Hospital-level support (9 items) Senior managers consult with staff on daily 0.67 problems and proceed Senior management that listens and 0.65 responds to employee concerns Staff nurses are involved in the internal 0.58 governance of the hospital Staff nurses have the opportunity to serve 0.57 on trust committees A director of nursing who is highly visible 0.51 and accessible to staff Nursing staff are supported in pursuing 0.46 degrees in nursing Active quality assurance audit programs 0.42 Nurses participate to control costs 0.36 Nurses participate in selecting new 0.35 equipment a Extraction method: principal axis factoring. Rotation method: oblimin with Kaiser normalization. b Factor loadings on items according to pattern matrix. Table 2 Mean item scores with standard deviations (SD) and Cronbach alpha for Icelandic NWI R a subscales Mean (SD) Observed range Cronbach alpha NWI R sub-scales (range) Nurse physician work 3.0 (0.5) relations (1 4) Unit level support (1 4) 2.9 (0.5) Staffing (1 4) 2.6 (0.7) Philosophy of nursing 2.9 (0.5) practice (1 4) Hospital level support (1 4) 2.3 (0.5) a Nursing work index Revised. good. Upon examining the distributions of the satisfaction and quality of care variables, it was decided to analyse the likelihood of nurses giving the strongest possible positive responses on these two items in relation to the NWI R subscales Work environment factors as predictors of nurse and patient outcomes As can be seen in Table 4, each of the five NWI R subscales taken alone was a statistically significant predictor of both nurse job outcomes and of nurse-rated quality of patient care. However, only unit-level support and staffing were significant independent predictors of nurse job satisfaction after controlling for the effect of the five factors together as well as the background variables. A one-point increase in the average ratings a nurse gave to unit-level support and staffing were associated with approximately a five-fold increase in the odds of reporting high job satisfaction. Staffing was the sole statistically independent predictor of emotional exhaustion once all other factors were controlled, with a one-point increase on this score associated with a mean drop of 3.45 points on the scale. Lastly, unitlevel support and nurse doctor relations were the only Table 3 Nurse assessment of quality of patient care and satisfaction with present job Rating Number (%) Job satisfaction a Very satisfied 236 (34.4) Moderately satisfied 328 (47.7) A little dissatisfied 79 (11.5) Very dissatisfied 44 (6.4) Quality of Excellent 181 (26.4) patient care b Good 471 (68.7) Fair 33 (4.8) Poor 1 (0.1) a Satisfaction with present job. b Nurse assessed quality of patient care.

6 S. Gunnarsdóttir et al. / International Journal of Nursing Studies 46 (2009) Table 4 Associations of Icelandic NWI R subscales with nurse job outcomes and nurse perceptions of quality of care Very satisfied with current job (OR (95% CI)) a Emotional exhaustion scores (ß (SE)) b Alone With all subscales Alone With all subscales Excellent nurse-rated quality of patient care (OR (95% CI)) a Alone With all subscales Nurse physician relations ( ) *** ( ) (0.63) *** (0.66) ( ) *** ( ) * Unit-level support ( ) *** ( ) *** (0.64) *** (0.80) ( ) *** ( ) * Staffing * ( ) *** ( ) (0.47) *** (0.51) *** ( ) *** ( ) Philosophy of practice ( ) *** ( ) (0.65) *** (0.70) ( ) *** ( ) Hospital-level support ( ) *** ( ) (0.66) *** (0.74) ( ) *** ( ) Significance: * p<0.05; *** p< a Logistic regression modelling controlling for nurse background variables (characteristics and specialities). b Generalized linear modelling controlling for nurse background variables (characteristics and specialities). Lower scores indicate less distress. significant predictors of nurse-rated quality of patient care when all work environment and background variables were taken into account. A one-point increase on either of the scales was associated with nearly a doubling of the nurse s odds of reporting excellent quality of care on her unit in the fully adjusted models. 4. Discussion This paper indicates that an Icelandic adaptation of the NWI R, a tool employed extensively in nursing administration research in English-speaking countries, shows a factor structure comparable to that found by researchers working with it elsewhere. The aspects of nurses work environments it measures appear to form stable clusters of items across the countries in which it has been tested. The five NWI R scales identified here are similar to a factor solution for the NWI R published by Lake (2002) and thus support its validity. Three of the five subscales (i.e. nurse doctor relations, staffing adequacy and philosophy of nursing) are nearly identical to the corresponding scales in Lake s solution. Furthermore, four of the eight items of unit level support scale are the same as in Lake s nurse manager support scales and four of the nine items in the hospital level support scale are the same as in Lake s nurse participation in hospital affairs scale. Lake s solution, also known as the PES (Practice Environment Scale), forms the basis for the work environment tool adopted by the National Quality Forum ( retrieved 18 October 2005). The five NWI R subscales identified here are somewhat different from a scoring scheme based on autonomy, control, and nurse physician relations discussed in Aiken and Patrician s review (2000). Of further note is that the NWI R subscales derived here predicted job satisfaction, burnout levels and nurserated quality of patient care in Icelandic hospital nurses. This further highlights not only similarities between nursing in Iceland and in other countries, but also remarkable consistencies worldwide in the work environment features associated with professional nurses satisfaction with their jobs, feelings of emotional exhaustion and impressions of the care being delivered in their practice settings (Aiken et al., 2002a). Overall, this study joins a growing body of literature establishing the NWI R as a psychometrically sound instrument suited to the study of work environments for nurses in nurses internationally. A number of points regarding the descriptive findings bear mention. The Icelandic nurses in this study generally felt supported by their front-line managers, and felt that their working relationships with doctors were positive. On average, they assessed the staffing levels as adequate and the underlying philosophy of nursing supportive to their everyday practice (they tended to agree with statements to the effect that these elements were present in their current jobs). These findings suggest that LSH has important characteristics of supportive work environment as framed by magnet hospital model (Aiken, 2002; Kramer and Schmalenberg, 2002) and organizational empowerment (Laschinger et al., 2003; Laschinger et. al., 2006). However, perceptions of support from senior nurse management at the hospital level appeared somewhat lower. This may reflect a variety of factors, notably expectations on the parts of staff about the role of senior management in this particular setting or in this particular culture. In light of the hospital merger that occurred not long before the survey was conducted (a restructuring that required major organizational changes and increased nurse

7 926 S. Gunnarsdóttir et al. / International Journal of Nursing Studies 46 (2009) workloads), the relatively low levels of burnout and high levels of job satisfaction were somewhat unexpected. However, the high level of job satisfaction was supported by a previous study on Icelandic nurse job satisfaction (Biering and Flygenring, 2000). That so few of the nurses rated the quality of patient care on their units as excellent was somewhat surprising (see Table 4) given that a quality audit carried out around the same time showed more positive findings for the quality of patient care at the hospital (Icelandic National Audit Office, 2003). There may be cultural explanations for this finding, for instance a reluctance of Icelandic nurses to rate care as excellent. The main findings of this study suggest that support from unit-level managers, staffing adequacy, and nurse doctor relations have strong and consistent associations with nurse and patient outcomes. The implications for executives and managers in Iceland and other countries are clear. Managerial support at the unit level was a key predictor of both job satisfaction and quality of patient care. This finding is supported by previous studies in other countries about the significance of unit mangement for nurse job satisfaction (Aiken et al., 2002a; McNeese-Smith, 1999; Upenieks, 2002), for the prevention of nurse burnout (Aiken et al., 2002a; Janssen et al., 1999; Laschinger and Finegan, 2005; Laschinger et al., 2001), and for the quality of patient care (Aiken et al., 2002a; Laschinger et al., 2001). This may mean that supportive and empowering relations with frontline managers lead nurses to have positive attitudes towards their positions that may enhance their ability to provide good patient care. The exact roles of front-line nurse managers and methods for supporting nurses in their multifaceted work deserves further attention in research and practice in Iceland and abroad. Staffing levels were an important independent predictor of emotional exhaustion. This finding was consistent with literature that has established that low staffing capacity and heavy work demands (i.e. overload) are key causes of nurse burnout (Aiken et al., 2001; Laschinger et al., 2001). Relations with doctors were an important predictor of nurse ratings of quality of patient care, suggesting that fostering positive relations between the nursing and medical staffs serving the specific clinical populations may yield benefits in terms of morale and patient outcomes. These findings are specific to one Icelandic hospital (albeit the largest one in the country and its only university health care centre) at a single point in time, and are based on nurses self-reports regarding their work environments, experiences of their jobs, and perceptions of quality. Despite these limitations, they are however consistent with a large number of studies in other countries from other time periods. Although cross-cultural research on nursing practice often reveals interesting differences across countries, it appears that non-english speaking countries could learn much from successes in nursing administration practice in some of the largest health care systems because many of the core notions, for instance, those embedded in the Magnet hospital concept, seem to translate well. Furthermore, it seems equally possible that the profession in countries such as the U.S. could gain insights into human-resources challenges and practice management challenges in nursing by looking to research findings and data from high-performing health care organizations abroad. Acknowledgements This study has been supported by the British Council Chevening Scholarship and by grants from the Icelandic Centre for Research, the Icelandic Nurse Association Scientific Fund, the Icelandic Nurse Association Continuing Education Fund, the Landspitali University Hospital Scientific Fund and the Landakot Hospital Scientific Fund. References Aiken, L., Superior outcomes for magnet hospitals: the evidence base. In: McClure, M.L., Hinshaw, A.S. (Eds.), Magnet Hospitals Revisited: Attraction and Retention of Professional Nurses. first ed.. American Academy of Nursing. American Nurses Publishing, Washington, DC, pp Aiken, L., Patrician, P., Measuring organizational traits of hospitals. The revised nursing work index. Nursing Research 49 (3), Aiken, L., Clarke, S., Sloane, D., Sochalski, J., Busse, R., Clarke, H., et al., Nurses reports on hospital care in five countries. Health Affairs 20 (3), Aiken, L., Clarke, S., Sloane, D., 2002a. Hospital staffing, organization and quality of care: cross national findings. International Journal for Quality in Health Care 14 (1), Aiken, L., Clarke, S., Sloane, D., Sochalski, J., Silber, J., 2002b. Hospital nurse staffing and patient mortality, nurse burnout and job dissatisfaction. Journal of the American Medical Association 288 (16), Bass, B.M., Transformational Leadership. Lawrence Erlbaum Associates, New Jersey. Bauman, A., Brien-Pallas, L.O., Armstrong-Stassen, M., Blythe, J., Bourbonnais, R., Cameron, S. et al., Commitment and Care. The Benefits of a Healthy Workplace for Nurses, their Patients and the System. A Policy Synthesis Canadian Health Services Research Foundation. The Change Foundation. Biering, P., Flygenring, B., Study of Stress and Job Satisfaction in Icelandic Nurses. Icelandic Nurse Association, Reykjavik. Buchan, J., Still attractive after all these years? Magnet hospitals in a changing health care environment. Journal of Advanced Nursing 30 (1), Edwards, D., Bernard, P., A systematic review of stress and stress management interventions for mental health nurses. Journal of Advanced Nursing 42 (2), Edwards, P., Roberts, I., Clarke, M., DiGuiseppi, C., Pratap, S., Wentz, R., et al., Increasing response rates to postal questionnaires: systemic review. British Medical Journal 324,

8 S. Gunnarsdóttir et al. / International Journal of Nursing Studies 46 (2009) Estabrooks, C.A., Midodzi, W.K., Cummings, G.G., Ricker, L.K., Giovannetti, P., The impact of hospital nursing characteristics on 30-day mortality. Nursing Research 54, Halldorsson, M., In:Bankauskaite, V. (Ed.), Health Care Systems in Transition Iceland, first ed. WHO Regional Office for Europe on behalf of the European Observatory on Health Systems and Policies, Copenhagen. Hayes, L.J., O 0 Brien-Pallas, L., Duffield, C., Shamian, J., Buchan, J., Hughes, F., Laschinger, H.K.S., North, N., Stone, P.W., Nurse turnover: a literature review. International Journal of Nursing Studies 43, Health Services Act, No 97/ September Icelandic Parliament, Reykjavík (1990). Retrieved February 14, 2005 from hhttp:// Hilton, A., Skrutkowski, M., Translating instruments into other languages: development and testing processes. Cancer Nursing 25 (1), 1 7. Icelandic National Audit Office, The Merger of the Reykjavík Hospitals: Performance Assessment. Ríkisendurskoðun [Icelandic National Audit Office], Reykjavík. Jamal, M., Job stress and burnout among Canadian managers and nurses: an empirical examination. Canadian Journal of Public Health 91 (6), Janssen, P.P.M., de Jonge, J., Bakker, A.B., Specific determinants of intrinsic work motivation, burnout and turnover intentions: a study among nurses. Journal of Advanced Nursing 29 (6), Kramer, M., Schmalenberg, C., Staff nurses identify essentials of magnetism. In: McClure, M.L., Hinshaw, A.S. (Eds.), Magnet Hospitals Revisited: Attraction and Retention of Professional Nurses. American Nurses Publishing, Washington, DC. Kramer, M., Schmalenberg, C., Magnet hospital nurses describe control over nursing practice. Western Journal of Nursing Research 25 (4), Kramer, M., Schmalenberg, C., Development and evaluation of essentials of magnetism tool. JONA 34 (7/8), Lake, E.T., Development of the practice environment scale of the nursing work index. Research in Nursing & Health 25, Lang, T.A., et al., Nurse-patient ratios: a systematic review on the effects of nurse staffing on patient, nurse employee and hospital outcomes. Journal of Nursing Administration 34 (7/8), Laschinger, H.K.S., Finegan, J., Empowering nurses for work engagement and health in hospital settings. JONA 35 (10), Laschinger, H.K.S., Shamian, J., Thomson, D., Impact of magnet hospital characteristics on nurses perceptions of trust, burnout, quality of care and work satisfaction. Nursing Economics (19), Laschinger, H.K.S., Almost, J., Tuer-Hodes, D., Workplace empowerment and magnet hospital characteristics: making the link. Journal of Nursing Administration 33 (7/8), Laschinger, H.K.S., Purdy, N., Cho, J., Almost, J., Antecedents and consequences of nurses managers perception of organisational support. Nursing Economics 24, LSH, Staff Records. Office of Human Resources. LSHhospital, Reykjavík. Maslach, C., Jackson, S.E., Leiter, M.P., Maslach Burnout Inventory Manual, third ed. Consulting Psychologists Press, Palo Alto. McClure, M.L., Hinshaw, A.S., The future of magnet hospitals. In: McClure, L.M., Hinshaw, A.S. (Eds.), Magnet Hospitals Revisited: Attraction and Retention. McClure, M.L., Poulin, M.A., Sovie, M.D., Wandelt, M.A., Magnet hospitals: attraction and retention of professional nurses (the original study). In: McClure, M.L., Hinshaw, A.S. (Eds.), Magnet Hospitals Revisited: Attraction, Retention of Professional Nurses. first ed. American Academy of Nursing. American Nursing Publishing, Washington, DC, pp McNeese-Smith, D., The relationship between managerial motivation, leadership, nurse outcomes and patient satisfaction. Journal of Organisational Behaviour 20, Nedd, N., Perceptions of empowerment and intent to stay. Nursing Economics 24, Needleman, J., Buerhaus, P., Mattke, S., Stewart, M., Zelevinsky, K., Nurse-staffing levels and the quality of care in hospitals. New England Medical Journal 346 (22), OECD, Tackling Nurse Shortages in OECD Countries. OECD, Paris. Övretveit, J., The Leaders Role in Quality and Safety Improvement: A Review of Research and Guidance. Association of County Councils, Stockholm. Polit, D.F., Hungler, B.P., Nursing research: Principles and methods. Lippincott, Philadelphia. Reise, S., Walker, N., Comrey, A., Factor analysis and scale revision. Psychological Assessment 12 (3), Sigurdardóttir, S., Finnbogadóttir, A., Moller, A., Einarsdóttir, E., Ingadóttir, E., Sigurdardóttir, G., et al., Nursing Shortage (Icelandic). Icelandic Nurse Association, Reykjavik. Tabachnick, B., Fidell, L., Using Multivariate Statistics. International Student Edition. Allyn and Bacon, Boston. Tourangeau, A.E., Cranley, L.A., Jeffs, L., Impact of nursing on hospital patient mortality and related policy implications. Quality and Safety in Healthcare 15, 4 8. Upenieks, V.V., Assessing differences in job satisfaction of nurses in magnet and nonmagnet hospital. JONA 32 (11), Wanous, J.P., Hudy, M.J., Single-item reliability: a replication and extension. Organizational Research Methods 4,