Age-dependent relationships between work ability, thinking of quitting the job, and actual leaving among Italian nurses: A longitudinal study

Transcription

1 Available online at International Journal of Nursing Studies xxx (2008) xxx xxx Age-dependent relationships between work ability, thinking of quitting the job, and actual leaving among Italian nurses: A longitudinal study Donatella Camerino a,b, *, Paul Maurice Conway a,b, Beatrice Isabella Johanna Maria van der Heijden c,d,e, Madeleine Estryn-Béhar f, Giovanni Costa a,b, Hans-Martin Hasselhorn g a Department of Occupational Health, University of Milan, Milan, Italy b IRCCS Ospedale Maggiore Policlinico Mangiagalli e Regina Elena Foundation, Milan, Italy c Maastricht School of Management, Maastricht, The Netherlands d Open University of the Netherlands, Heerlen, The Netherlands e University of Twente, Enschede, The Netherlands f Service Central de Medecine du Travail Hôpitaux Hôtel Dieu AP-HP de Paris, Paris, France g Department of Safety Engineering, University of Wuppertal, Wuppertal, Germany Received 17 January 2008; received in revised form 11 March 2008; accepted 12 March 2008 Abstract Background: Qualified nurses commonly report several work-related problems, which may threat their health and work ability, and may have an impact upon retention. Objective: The aim of the present study was to investigate whether perceived work ability is a predictor for different types of thinking of quitting and for actual exit from the workplace. Design: A prospective study has been performed based on 3329 Italian registered nurses which responded to both measurements of the Nurses Early Exit Study. Institutions were selected by means of a stratified sampling procedure. All the 7447 nurses working in that healthcare organizations were recruited for the study. Baseline questionnaire was collected from 5504 nurses. Response rate at follow-up was 63.4%. Two hundred and fifty-five nurses left the job in the course of the investigation. Methods: Several multiple logistic regression analyses were performed controlling for indicators of work-related well-being, socio-demographic, and labour-market characteristics. Results: As a whole, our results show that among nurses younger than 45 years work ability is a significant predictor of different types of thinking of quitting, but not of actual exit from the workplace: a lower perceived work ability was associated with a higher desire to undertake further education (OR: 3.05; 95% CI: ) and/or to change workplace (OR: 4.03; 95% CI: ) or profession (OR: 6.67; 95% CI: ). In contrast, among the older nurses (>45), only actual exit was predicted by lower work ability scores (OR: 7.14; 95% CI: ), along with the perception of a larger availability of free nursing posts in the region (OR: 2.70; 95% CI: ). * Corresponding author at: Department of Occupational Health, University of Milan, via San Barnaba 8, Milan, Italy. Tel.: address: donatella.camerino@unimi.it (D. Camerino) /$ see front matter # 2008 Elsevier Ltd. All rights reserved.

2 2 D. Camerino et al. / International Journal of Nursing Studies xxx (2008) xxx xxx Conclusion: Therefore, we conclude that in the relationship between low perceived work ability and intended or actual exit, a significant part is explained by age itself, but also by the age-related differences in occupational and life opportunities. This contribution concludes with some age-related policies aimed at boosting nurses retention. # 2008 Elsevier Ltd. All rights reserved. Keywords: Aging; Personnel turnover; Perceived work ability; Nurses; Well-being What is already known about the topic? Nursing shortage and lengthened working life before retirement exist in Italy as well as in other countries. The work ability index is a predictive measure of early pension or other forms of job change. The relationship between work ability and intent to leave the job or actual leave depends mainly on lifelong career experiences. Besides work ability, job dissatisfaction, low organizational commitment and emotional exhaustion are important mediators between working conditions and turnover among nurses. What this Paper adds? When work ability is low, nurses may have different turnover intentions or behaviours depending on their age. The age effect on turnover may be dependent upon health conditions, experiences and skills acquired, and also upon life and/or work opportunities. The work ability index partly overlap with the contents of well-known indicators of poor well-being such as emotional exhaustion, organizational commitment and job dissatisfaction. 1. Background In Italy, since the early 1970s fertility rates have declined steeply and are today among the lowest in the world. As decreased fertility rates have been accompanied by an increase in longevity, Italy exhibits today the highest dependency ratio among the OECD countries (OECD, 2000). In addition, in Italy older workers tend to exit the labour market relatively early. In 2002, only 56% of males and 27% of females in the age group were still active at the labour market. Large-scale pension reforms (laws 537/93, 335/95, 449/97, 338/2000 and 243/2004) discouraged early retirement and raised age limits for pension at 60 years old for women and 65 for men, granting that an adequate number of contribution years are achieved. In spite of this, the longterm financial sustainability of the Italian system remains under serious risk unless either the minimum age thresholds for retirement will be increased, in line with prospective improvements in life expectancy, or barriers to employment of older workers will be removed. As of February 2001, in Italy 320,657 nurses were workinginthenationalhealthsystemandintheprivate sector (Ministry of Health, 2001). Between 1997 and 2002, on average 12,500 nurses retired each year, while only 3500 new graduates entered the profession. At the end of this period, this imbalance resulted in a shortage of about 100,000 nurses. Only 5.3 nurses per 1000 residents were available, while the adequate proportion was set to 6.9 according to the OECD (OECD, 2000). Especially the youngest (20 25 years) and the oldest (46 55 years) age groups included few nurses, 9% and 3%, respectively. This indicated substantial problems in recruiting new nurses while, at the same time, prior legislation had laid the bases of favourable conditions for early retirement (law no. 503/1992 and subsequent modifications). This shortage of nursing staff boosted a high internal mobility rate among the remaining nurses (Hasselhorn et al., 2003), favouring the search, inside the nursing profession, for better working conditions, and for job posts more suited to personal needs in terms of career perspectives, vocational training opportunities and quality of private life. Therefore, efforts should be spent in protecting nurses work ability and finding ways to attract new nurses and retain older ones in the profession, so as to guarantee adequate coverage of the increasing healthcare needs of a greying and thus more demanding population. The work ability index (WAI), as a measure of the degree of fit of the worker with his/her job, may be considered as a useful tool to better understand relevant antecedents of job turnover. The work ability index proved to be a predictive measure of early pensioning and other conditions such as change of employer and/or profession, long-term absence and also unemployment (Ilmarinen, 2006). Previous results from the NEXT 1 Study (Hasselhorn et al., 2003) showed that the work ability index is a factor associated with intent to leave the nursing profession, and that this relationship may change according to age. 1 The NEXT Study, a multi-focused investigation on the reasons for leaving the nursing profession in Europe, is financed by the European Commission within the Fifth Framework, Project ID: QLK-6-CT , and is academically coordinated by Dr. Hans-Martin Hasselhorn from the University of Wüppertal, Germany. Website:

3 D. Camerino et al. / International Journal of Nursing Studies xxx (2008) xxx xxx 3 2. Introduction In 1976, Stelman maintained that since nursing is physically and emotionally demanding, long-term employment as a nurse may adversely affect work ability, and therefore be conducive to early retirement. Recent results demonstrated that preventing poor work ability, on the one hand, and promoting good work ability, on the other hand, require different working factors to be addressed (Lindberg et al., 2006). The relationship between one s work ability and intent to leave or actual work cessation may depend on the wear and tear of the present and/or the past working experience, and also on health deficiencies especially when the work does not contribute or no longer contributes to the promotion of a worker s health (Molinié, 2005). However, it also derives from work opportunities and/or living conditions (Camerino et al., 2006; De Censo and Robbins, 1996; Lewis and Tomas, 1987). As a whole, personal attitudes towards the end of one s working life develop over the entire individual career history (Hayward et al., 1998). To a certain extent, one s working life course is influenced by rules and opportunities determined by the broad social context (labour market, parental socio-economic status, marital status, group membership, etc.). These define possible role boundaries at any age, boundaries that may vary depending on the particular historical context. Within such personally and socially determined boundaries, there are individuals which in the course of their working life may develop adequate skills and knowledge (McKnight et al., 2005) to successfully fulfil their job requirements. Besides job-related competencies, there are workers also capable of acquiring abilities allowing them to maintain control over their own career across organizations and/or professional domains (Van der Heijden, 2005). These abilities may help workers to recognize themselves as still employable, even at a later stage in their career. Hence, for these people leaving a job may represent a way to improve their own career or living conditions. On the contrary, there are people who may experience recurrent maladjustments during their working life, or situations in which their supervisor and/or management are unable to arrange conditions suitable to improving their work ability. This would probably result in a person realizing at a certain point in his or her career that one s work role is over, and that one s job performance has become inadequate or obsolete. For these workers, leaving the job may be assumed to be a palliative or even a definitive solution. Such explanations of job or professional turnover are based on individual career stages and trajectories. However, when ageing is at issue, we should not disregard the possibility that some older workers do not want to leave their job any longer since they have obtained less demanding jobs and/or found ways to cope. They may have also changed their views and expectations concerning work, or they may regard the short time left until retirement to be bearable. As there may be different reasons leading to the decision to keep or leave a job, it is of utmost importance to better understand what kind of arrangements may sustain employees to carry out their job until they reach retirement age. This holds especially under current labour market circumstances, wherein the age threshold for retirement is raising, and the problem of a nursing shortage requires more nurses to keep working until older ages. Apart from work ability, extensive previous research has shown that also factors such as job satisfaction, organizational commitment and emotional exhaustion (as indicators of psychological well-being) are important antecedents of job turnover (Gaertner, 1999; Gauci-Borda and Norman, 1997; Irvine and Evans, 1995; Tai et al., 1998; Taunton et al., 1997). Accordingly, the present study is aimed at verifying the hypothesis that the relationship between work ability and intention to leave the job/actual leaving is modified by age. In detail, the perspectives of the younger nurses may lead them to seek improvements in their working conditions and/ or job skills, mainly by changing organization, taking further professional education or even change profession, while the perspectives of the older nurses may lead them to the decision to exit the labour market altogether. This relationship is studied while controlling for well-known health antecedents of nursing turnover like job dissatisfaction, organizational commitment and emotional exhaustion. Our research may contribute to shed light into the consequences of a poor work ability on nursing turnover and discuss possible ways to sustain individual careers in order to safeguarding life-long employability in the different age groups. 3. Methodology 3.1. Sample and procedure Between autumn 2002 and spring 2003, data were collected from 5504 out of the 7447 Italian nurses that were recruited for the study (response rate 73.9%). The respondents were employed at 127 different establishments nested within 16 wider healthcare organizations (4673 nurses from nine hospitals, 179 nurses from five nursing homes and 620 nurses from two home care institutions). The institutions were selected by means of a stratified sampling procedure, with strata defined on the basis of territorial area (North, Center and South of Italy), type of institution (hospital, nursing home, and home care institution), and public or private ownership (Hasselhorn et al., 2003, 2005). In all participating institutions, the study was introduced by means of oral presentations. Advertising material (posters and leaflets) was also distributed across the health care institutions in order to improve involvement of all nursing staff employed. A baseline and a follow-up self-administered questionnaires assessing socio-demographic characteristics, work contents, organizational and psychosocial

4 4 D. Camerino et al. / International Journal of Nursing Studies xxx (2008) xxx xxx factors, physical and psychological health and well-being, were distributed to each nurse at the same Institution where they worked. Informed consent was signed by each nurse before participation and then enclosed with the compiled questionnaire. A coding system was used to assure participants anonymity and compiled questionnaires were collected at each institution in a sealed envelope and then delivered at the research institute. The follow-up questionnaire was sent to the respondents 12 months after baseline. In the follow-up questionnaire, the same baseline questions were asked, plus additional information concerning changes occurred between the two assessments (i.e. working hours, work load, job position, etc.) and reasons for remaining. These additional questions were not analyzed in the present study. Data were provided for both measurement occasions by a total of 3329 nurses (response rate at follow-up 63.4%). During this period, 255 nurses had left their job. 2 Accordingly, the samples differed depending on whether the analysis was conducted upon the Italian nurses who left the job in the course of the investigation, or upon those nurses who remained at the same workplace 12 months after the baseline measurement Measures Outcomes Distinguished forms of thinking of quitting have been measured at follow-up by eight single-item questions that were developed by the NEXT Study group specifically for the nursing profession (Hasselhorn et al., 2003). In detail, these address different possibilities a nurse may consider when he or she thinks about changing the job or even the profession. The single-item questions include thinking about : (1) taking a further qualification in nursing, (2) taking a further qualification outside nursing, (3) finding another nursing post in the current institution, (4) leaving the current institution, (5) giving up nursing completely, (6) giving up nursing completely to start a different kind of job, (7) working as a nurse in another country, and (8) working in another country but not in nursing. Responses were rated on a five-point scale: never, sometimes a year, sometimes a month, sometimes a week, every day. At baseline, thinking of quitting was measured by a single item: How often do you think about giving up nursing 2 By means of another measurement instrument, i.e. the exit questionnaire, we could obtain information on 118 out of these 255 subjects about reasons of their job interruption. Although the small sample cannot allow us to generalize the following findings to the all sample of nurses who left, the search for better work and living conditions was reported as the major reason for exit (31.4% and 28%, respectively), while career improvement was reported as the main reason by 11% of the sample. According to another item included in the exit questionnaire, we also found that of the subjects who left their current job, 12.7% retired. completely? For this item, the five response categories (see above) were recoded into two categories: high thinking of quitting (sometimes a month, sometimes a week, everyday), and low thinking of quitting (never, sometimes a year). Actual leaving the job was measured on the basis of information concerning employees turnover gathered at the different participating institutions Predictors Preliminary univariate analyses were conducted in order to determine possibly important predictor variables related to thinking of quitting. Gender, occupational position, type of healthcare institution, location of birth, and availability of nursing posts inside the region of current employment were then included in the main analyses. Age was used as a stratification variable, adopting a cut-off point of more than 45 years to define ageing workers in line with previous literature (Ilmarinen, 2001; Kilbom et al., 1997; Tikkanen, 1999; WHO, 1993). Out of 5504 nurses included in our study, 4250 were aged <45 years old (78.4%) and (21.6%). As indicators of well-being, we used work ability, job satisfaction, organizational commitment, and emotional exhaustion. Perceived work ability was measured by the work ability index. This index has now been translated into 25 languages and it is widely used in epidemiological studies (Ilmarinen and Tuomi, 2004). The WAI comprises seven domains or items (Tuomi et al., 1998): (1) subjective estimation of current work ability compared with optimal life time performance (0 10 points); (2) subjective work ability in relation to physical and mental demand of work (2 10 points; the score is weighted according to whether the work is primarily physical or mental); (3) number of diseases for which a medical diagnosis has been given (1 7 points; set beforehand scores according to the number of diseases diagnosed by a physician); (4) subjective estimation of working impairment due to ill health (1 6 point); (5) sickness absenteeism during the past year (1 5 points); (6) own prognosis of work ability after 2 years (1, 4 or 7 points; set beforehand scores anchored to the possible three answers); (7) psychological resources (enjoying daily tasks, activity and life spirit, optimistic about the future; 1 4 points). The total score is calculated by summing up all item scores (Tuomi et al., 1998) and it ranges 7 49 points, with higher scores indicating a higher perceived work ability. The internal validity of the WAI has been demonstrated (Eskelinen et al., 1991; Nygård et al., 1991), as well as the instrument s stability according to test retest reliability analysis (De Zwart et al., 2002). In the present study, we adopted the short version of the WAI (Nübling et al., 2004) that differs from the original instrument in that item 3 contains only 15 medical conditions instead of 51. The shorter version is more readily completed, and also appears to have a better face validity. Obviously, this was an advantage in a study which required respondents to complete

5 D. Camerino et al. / International Journal of Nursing Studies xxx (2008) xxx xxx 5 Table 1 Risk and reference categories of study variables Variables Risk category Reference category Individual factors associated with different quitting thoughts Gender Female Male Occupational position Staff Managerial nurses Ward sister Type of health care institution Hospital Home care Nursing home Location of birth In another region Same city or area of work In another country External factors associated with different quitting thoughts Perceived availability of nursing posts in the same region of employment Yes No Well-being indicators Job satisfaction (Kristensen, 2000) No Yes Organizational commitment (Allen and Meyer, 1990) No Yes Emotional exhaustion (Borritz and Kristensen, 2001) Yes No Thinking of changing (NEXT Study Group) Time 1 Yes No Work ability index (Tuomi et al., 1998) Poor Excellent Moderate Good a long questionnaire. Nübling et al. (2004) has developed an algorithm to allow for the comparability of the data obtained by the two versions and has found good convergent validity. Job satisfaction was measured using Kristensen s (2000) four-item scale. A sample item is: How pleased are you with your job as a whole? Responses are to be given on a fourpoint scale ranging from (a) very unsatisfied, to (d) very satisfied. Cronbach s alpha was.74. Organizational commitment was assessed by means of a four-item scale referring to the affective component of the commitment to the organization construct (Allen and Meyer, 1990). This component refers to the degree to which the employee identifies with, is involved in, and is emotionally attached to the organization. An example item is: I am proud to belong to this institution. Responses were rated on a five-point scale ranging from: (a) strongly disagree to (e) strongly agree. Cronbach s alpha was.73. Emotional exhaustion was measured using the six-item scale from the Copenhagen Burnout Inventory (CBI) (Borritz and Kristensen, 2001). Participants had to indicate on a five-point scale how often they feel tired, physically exhausted, emotionally exhausted, and how often they think: I can not take it anymore, feel worn out, and feel weak and susceptible to illness. The answering categories are: (a) never/almost never, (b) once or a few times during a month, (c) once or twice a week, (d) three to five times during a week, and (e) (almost) every day. Cronbach s alpha was.89. For the logistic analyses, the continuous variables job satisfaction, organizational commitment and emotional exhaustion were split into quintiles, with the worse quintile used as the risk and the others as the reference category. Table 1 shows risk and reference categories of study variables Drop-outs. Following procedures recommended by Twisk (2003), several logistic regression analyses were conducted to test if drop-outs (i.e. those who did not respond to the 1-year follow-up questionnaire) differed from those who participated in the follow-up measurement. These analyses were conducted excluding those cases (N = 255) who left the job during the 12 months follow-up period. With respect to the study variables, there were significantly more drop-outs among those who, at baseline, declared being a staff nurse (i.e. without managerial duties) compared to a head or a nurse manager, working in a hospital or a nursing home compared to working in a home care service, being born outside the region of employment and having lower perceived availability of alternative nursing posts. All the other predictors and outcome variables were not related to drop-out Data analysis. Due to the discrete nature of the outcome variables (i.e. the single-item measures of thinking of quitting and actual leaving ) and in order to ease interpretation of results in terms of risk, all variables scores have been re-categorized so as to fit into logistic regression models. Multiple logistic regression analyses have been conducted separately for the different outcomes (forms of thinking of quitting and actual leaving ), using a threestep hierarchical procedure. The same analyses have been conducted among the nurses who had remained at the same workplace after 12 months, and among those who had left their job in the course of the investigation, separately for the

6 6 D. Camerino et al. / International Journal of Nursing Studies xxx (2008) xxx xxx two age groups (<45 years versus 45 years). The statistical analyses have been carried out controlling, in the first step, for socio-demographic characteristics (age, gender, occupational position, type of healthcare institution, and national region), and perceived availability of nursing posts in the same region of employment. In the second step, indicators of work-related well-being, and thinking of quitting at baseline measurement were included. Finally, in the third step, work ability was entered so as to estimate its unique effect on the outcome variable. All analyses have been conducted using SPSS version Results The socio-demographic characteristics of the samples are shown in Table 2. Sample distribution remained about the same from the baseline and both the follow-up and the exit samples. Owing to early retirement opportunities, most of our nurses in the baseline sample were younger than 45 years (78.4%). They were mostly female, even if male nurses represented a significant part of the Italian sample (25.7%). In fact, Italy reported the highest prevalence of male nurses among the countries participating in the NEXT Study due to previous unemployment in the South of Italy, and on the islands, and also due to more recent laws that encouraged their entry into the profession (Hasselhorn et al., 2003). Most of the sample was composed by staff nurses (90.8%); the number of head nurses or nurse directors was, as expected, higher among employees 45 years. Finally, most respondents in our sample worked in hospitals and were employed based upon a full-time contract, which points to the delayed process of territorialization of the healthcare system in Italy. Half of the sample consisted of respondents from the North of Italy, and half from the Center and the South of Italy. Table 3 depicts the sample s sociodemographic and work-related characteristics for the two age groups <45 and 45 separately. At baseline (Time 1), the proportion of males, and the proportion of respondents with an occupational position higher than staff nurse was higher among the older than among the younger nurses. Perceived availability of nursing posts in the same region of employment, emotional exhaustion, andthinking of quitting the job were higher among the younger nurses, while older nurses reported a higher organizational commitment. The perceived work ability was only slightly higher among nurses below 45 years of age. At the follow-up measurement (Time 2), the younger nurses reported a significantly higher score for thinking of taking further qualification (both inside and outside nursing; x 2 = (d.f. 1), p <.001; x 2 = (d.f. 1), p <.001, respectively) and a significantly higher score for thinking of changing their job (both inside and outside their current institution; x 2 = (d.f. 1), p <.001; x 2 = 7.20 (d.f. 1), p <.01, respectively), while the other types of thinking of quitting (i.e. giving up nursing completely, giving up nursing completely to start a different kind of job, working as a nurse in another country, and working in another country but not in nursing) were comparable across the two age groups. Less than 10% of nurses with excellent WAI reported having thinking of quitting, while % (according to the type of thinking of quitting ) of nurses with poor WAI declared they think of leaving. Table 2 Socio-demographic characteristics at the baseline and follow-up measurement Baseline sample (N = 5504) Follow-up sample (N = 3329) Exit sample (N = 255) Age < (78.4%) 2554 (77.6%) 208 (81.6%) (21.6%) 737 (22.4%) 47 (18.4%) Gender Female 4072 (74.3%) 2488 (74.7%) 205 (79.8%) Male 1406 (25.7%) 833 (25.1%) 50 (19.5%) Occupational position Staff nurse 4959 (90.8%) 2937 (88.7%) 235 (92.2%) Head nurse 454 (8.3%) 338 (10.2%) 15 (5.9%) Nurse manager 51 (0.9%) 36 (1.1%) 5 (2.0%) Type of healthcare institution Hospital 4673 (85.4%) 2748 (83.0%) 223 (87.1%) Nursing home 179 (3.3%) 111 (3.4%) 15 (5.9%) Home care 620 (11.3%) 452 (13.7%) 18 (7.0%) Location of birth North 2523 (46.5%) 1583 (48.2%) 153 (59.5%) Center 1408 (25.9%) 885 (27.0%) 89 (34.6%) South 1495 (27.6%) 815 (24.8%) 15 (5.8%)

7 D. Camerino et al. / International Journal of Nursing Studies xxx (2008) xxx xxx 7 Table 3 Socio-demographic and work-related characteristics by age <45 years 45 years Baseline assessment Male 22.4% 34.5% Staff nurse 91.9% 77.3% Head nurse 7.5% 19.9% Nurse manager 0.6% 2.8% Hospital 84.0% 80.3% Nursing home 3.5% 3.0% Home care 12.5% 16.7% Availability of nursing posts within the own region 54.7% 44.2% Location of birth 21.9% 19.4% Emotional exhaustion (Time 1) 20.1% 15.9% Job satisfaction (Time 1) 5.6% 8.7% Organizational commitment (Time 1) 14.7% 28.8% Thinking of quitting job (Time 1) 21.3% 15.7% Poor WAI (Time 1) 2.00% 3.3% Moderate WAI (Time 1) 18.9% 22.3% Good WAI (Time 1) 51.7% 47.1% Excellent WAI (Time 1) 27.3% 27.4% Follow-up assessment Taking a further qualification in nursing (Time 2) 11.5% 5.7% Taking a further qualification outside nursing (Time 2) 10.6% 5.3% Finding another nursing job in the current institution (Time 2) 13.2% 7.4% Leaving the current institution (Time 2) 12.6% 8.7% Giving up nursing completely (Time 2) 9.4% 10.7% Giving up nursing completely to start a different kind of job (Time 2) 9.3% 9.5% Working as a nurse in another country (Time 2) 2.3% 1.9% Working in another country but not in nursing (Time 2) 1.1% 1.6% Tables 4 and 5 show the results of the logistic multivariate analysis conducted among the two age groups separately and aimed at testing the associations between perceived work ability and distinguished types of thinking of quitting (ranging from the wish to take further education, both in or outside nursing, to the wish of changing workplace, or even to change profession and/or country), after controlling for socio-demographic characteristics, indicators of wellbeing, and baseline thinking of quitting. Our findings showed that in the fully adjusted model (Step 3) among nurses aged <45 years a poor perceived work ability (and also a moderate work ability, although its effect size was lower) was strongly associated with different types of thinking of quitting, except for thinking about working in another country but not in nursing (OR: ). A higher perceived availability of nursing posts within one s own region was associated with a higher prevalence of thinking of quitting the current institution, and a higher prevalence of thinking of giving up nursing completely. In the final model (Step 3), emotional exhaustion was positively related to thinking of finding another post in the same or in a different healthcare institution. Lower organizational commitment was related to a higher prevalence of thinking of quitting the current institution, orofgiving up nursing completely (either to withdraw from working life completely or to start a different kind of job). For most items pertaining to thinking of quitting, the odds ratios (Step 2) for the relationship between emotional exhaustion and organizational commitment on the one hand, and the different types of thinking of quitting, appeared to be weakened though still significant after perceived work ability was included in the model (see Step 3). Among nurses aged 45 years the work ability index was not associated with any kind of thinking of quitting. A lower perceived availability of nursing posts inside the region of current employment and different location of birth was associated with a higher prevalence of thinking of taking a further qualification in nursing, presumably in order to increase the chance to find employment close to home. Emotional exhaustion was related to thinking about leaving one s current institution, and to thinking about giving up nursing completely (either to withdraw from working life completely or to start a different kind of job). As far as the actual leaving from the job is concerned, our outcomes indicate that among the younger nurses (45), actual leaving from the job was predicted by the perceived availability of nursing posts in the same region of employment, and by the wish to get closer to one s own location of birth, while the work ability index did not play any sig-

8 Table 4 Adjusted multiple logistic regression for the association between baseline social demographic and work-related factors and different types of quitting thoughts among nurses (<45 years old) Taking a further qualification in nursing Taking a further qualification outside nursing Finding another nursing post in your current institution Occupational position Ward sister Staff nurse Type of institution Hospital Nursing home Gender Female Availability of nursing posts Location of birth Emotional exhaustion Job satisfaction Institutional commitment Thinking of quitting job WAI Poor Moderate Good Leaving your current institution Giving up nursing completely Giving up nursing completely to start a different kind of job Occupational position Ward sister Staff nurse Type of institution Hospital Nursing home Gender Female Availability of nursing posts Location of birth Emotional exhaustion Job satisfaction Institutional commitment D. Camerino et al. / International Journal of Nursing Studies xxx (2008) xxx xxx NS-1319; No of Pages 15

9 Table 4 (Continued ) Thinking of quitting job Leaving your current institution Giving up nursing completely Giving up nursing completely to start a different kind of job WAI Poor Moderate Good Working as a nurse in another country Working in another country but not in nursing Actual leaving Occupational position Ward sister Staff nurse Type of institution Hospital Nursing home Gender Female Availability of nursing posts Location of birth Emotional exhaustion Job satisfaction Institutional commitment Thinking of quitting job WAI Poor Moderate Good Empty cells contain infinite values due to denominator being equal to 0. D. Camerino et al. / International Journal of Nursing Studies xxx (2008) xxx xxx 9 NS-1319; No of Pages 15

10 Table 5 Adjusted multiple logistic regression for the association between baseline social demographic and work-related factors and different types of quitting thoughts among nurses (45 years old) Taking a further qualification in nursing Taking a further qualification outside nursing Finding another nursing post in your current institution Occupational position Ward sister Staff nurse Type of institution Hospital Nursing home Gender Female Availability of nursing posts Location of birth Emotional exhaustion Job satisfaction Institutional commitment Thinking of quitting job WAI Poor Moderate Good Leaving your current institution Giving up nursing completely Giving up nursing completely to start a different kind of job Occupational position Ward sister Staff nurse Type of institution Hospital Nursing home Gender Female Availability of nursing posts Location of birth Emotional exhaustion Job satisfaction D. Camerino et al. / International Journal of Nursing Studies xxx (2008) xxx xxx NS-1319; No of Pages 15

11 Table 5 (Continued ) Institutional commitment Thinking of quitting job Leaving your current institution Giving up nursing completely Giving up nursing completely to start a different kind of job WAI Poor Moderate Good Working as a nurse in another country Working in another country but not in nursing Actual leaving Occupational position Ward sister Staff nurse Type of institution Hospital Nursing home Gender Female Availability of nursing posts Location of birth Emotional exhaustion Job satisfaction Institutional commitment Thinking of quitting job WAI Poor Moderate Good Empty cells contain infinite values due to denominator being equal to 0. D. Camerino et al. / International Journal of Nursing Studies xxx (2008) xxx xxx 11 NS-1319; No of Pages 15

12 12 D. Camerino et al. / International Journal of Nursing Studies xxx (2008) xxx xxx nificant role. Among the older nurses (45 years), a higher actual leaving their job was significantly associated with a poor perceived work ability (OR: 7.14; 95% CI: ), and with the perceived availability of nursing posts in the same region of employment. A higher amount of emotional exhaustion was associated with a higher prevalence of actual leaving in Step 2, yet, after the work ability index was entered in the analysis, this effect disappeared. 5. Discussion 5.1. Reflections on the outcomes As a whole, in our study the proportion of Italian registered nurses with different form of thinking of quitting, or that actually left their job, was higher among those perceiving their work ability to be poor or moderate. However, thinking of quitting and actual leaving were also observed, though to a lower extent, among nurses perceiving their work ability to be good or excellent. This may indicate that actual or intended leaving may be also related to factors other than a nurse s perceived health status, in line with recent literature that focused on improving organizational characteristics and working conditions as the main tool in order to retain or attract nurses (Stordeur and D Hoore, 2007; Cohen, 2006). The relationship between work ability and actual leaving was different in the younger (<45 years) and the older nurses age groups (45 years), as already observed in previous NEXT publications (Camerino et al., 2006; Hasselhorn et al., 2003). A poor work ability index was related to more frequent thinking of quitting, yet only among the younger nurses, while it was related to higher actual leaving among the older ones. More specifically, among the younger nurses, a poor perceived work ability was associated with a stronger wish to undertake further education and/or to leave/ change workplace, or the profession altogether. However, for them, a poor work ability index did not predict actual leaving, since several factors prevented younger Italian nurses from having the real opportunity to realize the desired change. Indeed, younger nurses were more likely to actually leave if they were born outside the region of current employment and if they perceived the availability of nursing posts in the same region of employment to be high. Accordingly, younger nurses may be more willing to take the concrete step of leaving their current institution when they feel there are possibilities to find more suitable working conditions somewhere else in the same region of the current employment or closer to their location of birth, perhaps to take advantage of a better supporting social network. Thus, among the younger nurses the reasons for actually leaving their job seem to be related more to life and job opportunities than to their mental or physical health status. Among job opportunities, Bezerra (1997), Cowin and Jacobsson (2003) identified increased autonomy and better professional relationships as critical factors in job retention. Unfortunately, our sample comprised very few nurses older than 50 years old (11.6%). This may be due to the fact that many of them exited the labour force to take advantage of early retirement policies, that during the period of data collection were about to shift toward more restrictive criteria. A nurse decision to retire prematurely may be associated with both the wear and tear of nursing and also the difficulty to keep up with the continuous rearrangements faced by the healthcare system. Our data showed that among the older nurses, actual leaving was more likely to occur among nurses who reported a poor perceived work ability, say those with a lower capacity to exert control over their work conditions. Early retirement may serve as a pathway towards the end of working life for those older employees who feel no longer able to sustain or to improve their own career in nursing. However, beyond poor work ability, actual leaving among older nurses was more likely for these nurses who perceived having higher possibilities to find an alternative job in nursing in the same region of current employment. It may be that, irrespective of their current occupational position (which was controlled for in the analyses), older nurses may more easily take the decision to change workplace when they feel there are opportunities for them to find better alternative working conditions (both in terms of well-being, and in terms of career perspectives). In conclusion, the pattern of relationships between low perceived work ability and thinking of quitting or actual leaving may differ between the younger and the older nurses. Similarly, Cowin and Jacobsson (2003) and Chang et al. (2007) have observed different career needs at different career stages, and suggested that these implications be considered in development career programmes. Differences may be traced back to different possibilities the younger and the older nurses have to translate a poor health status or problems in well-being into actual leaving the job, depending on diverse age-related occupational and life opportunities. When unhealthy, younger nurses may feel unable to change workplace since they lack confidence about the possibility to exert a positive control over their working life situation. Yet, this may lead to poor performance, owing not only to their limited work ability but also to their perceived unsatisfactory working conditions. However, also labour market constraints such as the lack of available job alternatives may prevent them from seeking alternative employment. On the other hand, when unhealthy the older nurses may profit from opportunities (such as early retirement) to withdraw from the working life. However, available job alternatives may support older nurses who feel in control over their working life to continue practicing. For this reason, literature recommended to retain older nurses by means of adequate policies, favouring a better match between their needs and the choice of units and shift schedules, increasing resources and benefits to sustain prolonged participation in the workforce, safeguarding the balance between home and