Introduction

The professional stress syndrome of Burnout in the mental health services is a phenomenon that has been widely studied (e.g. Shirom 2003; Schaufeli 2003; Maslach et al. 2001; Salyers and Bond 2001). The studies that have been carried out so far describe burnout syndrome as a “process rather than an event” and tend to be associated with a condition of daily chronic stress, rather than with exceptional and occasional events. Unsatisfactory work can lead initially to a feeling of emotional exhaustion (EE), next to a defensive reaction such as depersonalization (DP), and finally to a lack of involvement at work and low personal accomplishment (PA). This process has negative implications for job performance and social relationships. Other theoretical models have been presented in which the three dimensions are related differently to each other (e.g. Maslach et al. 2001), and Shirom (2005) supports the need for further theoretical clarification of the burnout construct. Nonetheless, the EE is commonly considered as the central dimension of the burnout syndrome (Moore 2000; Koeske and Koeske 1989; Roelofs et al. 2005). On the other hand, high levels of EE are encountered also in other forms of mental illness not necessarily involved with stress at work, such as depression or anxiety disorders. For that reason, the emergence of both DP and EE high levels seems a more adequate criterion for diagnosing burnout-related disorders (e.g. Brenninkmeijer and Van Yperen 2003; Schaufeli et al. 2001; Schaufeli and Taris 2005).

Despite the aforementioned needs for deeper theoretical work on burnout dimensions and their relationships, the number of burnout interventions has increased in the last 20 years. Shaufeli and Enzmann (1998, pp. 143–183) create a broad classification of interventions along two important axes: (a) the focus (individual, individual-organisational interface, organization), and (b) the purpose (identification, primary or secondary prevention, treatment, rehabilitation). Frequently the interventions are shaped as workshops supporting the participant’s awareness of their work related problems and improving their ability in managing resources by skills training and/or providing social support. According to Shaufeli and Enzmann (2003), the interpersonal approach to burnout seems empirically supported. In particular, he ascribes the appearance of stressful interpersonal relationships at work to the lacking of reciprocity. In other terms, social workers feel that they are entering in time-consuming and highly taxing relationships with others without receiving a comparable benefit. This perceived lack of reciprocity affects their relationships with the organization as well as with team members. Lack of reciprocity in the group of co-workers as well as social comparison and emotional contagion play an important role in the emergence of burnout (Schaufeli 2003; Buunk et al. 2001; Bakker et al. 2000).

The aforementioned increase of the number of interventions did not moderate the need for new research in this area, particularly in Italy. Indeed, a review of Galeazzi and Priebe (2007) on Italian social psychiatry research points out that in the triennium 2004–2006 only 10% of papers published in peer-reviewed journals on Italian samples report findings on intervention trials with pre- and post-measures.

In order to plan a burnout prevention intervention, three relational concerns seem to be most recurring among mental health service workers (e.g. Van der Klink et al. 2001; Murphy 1996; Scarnera 2003). Firstly, Ben-Zur and Michael (2007) support the need for a better control of negative emotions arising from involvement with others’ illness. McCraty et al. (1998) indicate that participants in a training program on managing negative emotions had experienced significant increases in positive affect and significant decreases in negative affect. Secondly, Cummings et al. (2005) state the need for a greater diffusion of specific skills in leading and monitoring employees and personal involvement of the staff members. Gill et al.(2006) examine the effects of different leadership styles on job stress and on burnout. The degree of perceived stress is inversely related to the degree of support exhibited by managers, Finally, Lee and Crockett (1994) call for the promotion of assertiveness in communication between team members. They examine the effectiveness of assertiveness training in reducing levels of stress and improving assertiveness in communication among nurses. At the end of the training and at the follow-up, the experimental group had significantly higher scores on assertiveness and lower scores on stress measures compared to the control group. Shimizu et al. (2003) confirm previous findings on beneficial effects of assertiveness training on self-perceived level of occupational stress. All together, these findings have implications for tailoring the training of employees and managers. Indeed (a) the control of negative thoughts and emotions especially among employees, (b) the promotion of a supporting leadership among managers, and (c) the advancement of assertive communication among all staff members would represent a set of professional skills protecting from the emergence of burnout.

In light of the above, we decide to focus our research effort on an interpersonal level of intervention within a Cognitive-behavioral framework. Indeed, a quantitative meta-analysis determining the effectiveness of interventions on occupational stress (Van der Klink et al. 2001) demonstrated that stress management interventions are effective. In particular, Cognitive-behavioral interventions are more effective than the other intervention types. Recently, another meta-analysis (Richardson and Rothstein 2008) largely confirmed this finding. Surprisingly, it extended previous findings suggesting that the more components are added to a cognitive-behavioural intervention, the less effective it becomes. In tune with this conclusion, the core of the present intervention is an assertiveness training involving all the participants regardless of their specific role or occupation. The expectation is that the assertiveness training would provide the opportunity to harmonize professional aims and personal needs of managers and employees.

Method

Participants

Initially, 40 people were invited to participate to the intervention although only 25 (11 women), accepted. They were employed into two different but interconnected South Italian Organizations: the Mental Health Department of the ASL BA/3 in Altamura (Bari) and the Social Cooperative ‘Questa Città’ in Gravina di Puglia (Bari). The first organization provided psychiatric services in hospitals and wards, the second provided therapeutic/rehabilitative activities in residential communities and day care centres. Even though the number of participants was small, they represented a range of the different skilled workers employed in psychiatric therapeutic/rehabilitative services in Italy. Fourteen people, five males and nine females, were employees directly involved in care and assistance procedures with patients and their relatives (Group 1). Eight of them had a professional qualification specific for their occupations. They were nurses, educators and assistants. The average age was 40.7 years. Eleven people, nine males and two females played a managerial role (Group 2). Seven of them had a professional qualification specific for their occupation. They were psychologists, psychiatrists or managers of rehabilitative services. The average age was 41.9 years. The small sample size precluded the possibility of arranging an experimental and a control group. So, the intervention was evaluated only across time in a pre, post and follow-up data collection.

Materials and Procedure

Preliminarily, participants were asked to complete the following questionnaires:

  1. 1.

    The Italian version of the Maslach Burnout Inventory—MBI (Sirigatti and Stefanile 1993),

  2. 2.

    A tool adopted to modulate training. The Italian version of occupational stress inventory (OSI) based on the work of Cooper, Sloan and Williams (Sirigatti and Stefanile 2002).

After the first evaluation (T0), two skilled counsellors supervised a series of 6 monthly workshops of 3–5 h. The intervention’s aim was to develop techniques able to handle interpersonal relationships in different fields.

The main skill to promote was the communication among staff members involving all the participants regardless of their specific role or occupation. The training provided the opportunity to compare different work concerns emerging among staff members. Two other interventions were also planned: one for employees directly involved in relationship with patients and relatives. They were trained in handling their negative emotions and thoughts. The other intervention was planned for managers, trained to acquire awareness of both their leadership style and the needs of support rising from their assistants. These two programs preliminarily realized and specifically tailored on the needs of managers and employees (e.g. Schaufeli 2003; Van der Klink et al. 2001) were expected to improve the effect of the subsequent core intervention involving all the participants.

The first workshop was preliminarily carried out with each group separately. It was used to deal with specific drawbacks of each group. In particular, the action labelled “Identifying and managing disturbing emotions” (workshop 1a) was carried out with Group 1 members and it was aimed at achieving a cognitive restructuring of interpersonal relationships with severe mental illnesses. This intervention consisted of a cognitive-behavioural session that could limit the onset or re-define dissonant emotions (Zapf 2002; Brotheridge and Grandey 2002; Scorpiniti and Cotugno 2003). The action labeled “Optimal strategies for work task planning and handling of the relationships with staff members” (workshop 1b) was carried out with Group 2 members within a session inspired to Corrigan et al. (1998, 2000) and Corrigan and McCracken (1995) works.

The remaining five workshops, lasting 3 h each, consisted of a didactic and an action part on “Assertiveness” (Naimo 2003). During the didactic part, the following topics were addressed by the counsellors: “Handling different levels of interpersonal relationships” (workshop 2), “Asserting one’s point of view preserving a polite relationship” (workshop 3), “Communicating apprehension and mental preconceptions” (workshop 4), “Avoiding anger due to both inferiority feelings and unproductive self-satisfaction” (workshop 5), “Fostering positive professional relationships” (workshop 6). During the action parts, participants collectively realized and evaluated action strategies to cope with the most important stressors in their work situations. The training program was inspired by Lee and Crockett (1994) and Shimizu et al. (2003) findings.

At the end of each workshop, the participants’ satisfaction was evaluated throughout a questionnaire containing questions on usefulness, learning easiness, personal involvement, and pleasantness of the training situation. Questions were answered through a three-levels-graded scale (i.e. “not enough”, “quite enough”, “very much”). At the end of the intervention (T1), participants were asked to complete the MBI. Eighteen months after the beginning of the intervention, participants were asked to complete the MBI again and an ad hoc questionnaire was used for evaluating the effectiveness of the trained skills (T2). Questions provided information about the perceived frequencies of stress events due to interpersonal relationships (question 1), about the frequency in the use (question 2) and the perceived effectiveness (question 3) of the trained techniques, with clients and relatives (sub question a), with colleagues (sub question b) or with managers and supervisors (sub question c). Questions were answered through a five-levels-graded scale (with levels varying from “never” to “very frequently”). This intervention program was preliminarily approved by the Chairs of ASL BA/3 and of Social Cooperative ‘Questa Città’.

Results and Discussion

The coefficient alpha reliability estimated before (T0) and after (T1 and T2) the intervention for the three dimensions of burnout were 0.84, 0.81 and 0.79 for EE; 0.79, 0.81 and 0.78 for DP; 0.76, 0.76 and 0.78 for PA, respectively. A global measure of the effectiveness of the intervention was obtained by collapsing together questions 2 and 3 of the ad hoc questionnaire. The coefficient alpha reliability of that global measure was 0.79.

To evaluate the effect of the intervention, three-one-way, within factor ANOVAs were performed with time (T0, T1, T2) as within factors and each measure of burnout as dependent variable. The analysis on EE did not show the effect of intervention F (2; 48) = 1.78; P = 0.18 (mean T0 = 15.88; T1 = 16.56; T2 = 15.88). This result suggests that the intervention based on promotion of assertiveness in interpersonal relationships did not operate on the medium-low level of burnout experienced by the participants at T0. The analysis on DP showed the significant main effects of time with F (2; 48) = 5.04; P < 0.05; partial η 2 = 0.18, (mean T0 = 4.52; T1 = 3.76; T2 = 2.80). The main effect of time demonstrated that DP was inclined to linearly decrease from medium-high to medium-low level of burnout after the training on assertiveness and to maintain this trend after 18 months from the beginning of the training (linear trend analysis: F (1, 23) = 6.46; P < 0.05; partial η 2 = 0.22). Finally the analysis on the lower level of PA showed the effect of time with F (2, 48) = 3.98; P < 0.05; partial η 2 = 0.16, (mean T0 = 39.08; T1 = 35.08; T2 = 38.40). The intervention seemed to increase, paradoxically, from low to medium-low the lacking of PA immediately after the intervention. Even though, 18 months after the intervention start, the initial level of this index was restored (quadratic trend analysis: F (1; 23) = 6.93; P < 0.05; partial η 2 = 0.23). Such a variation may be caused by an initial higher awareness of one’s professional ineffectiveness due to the work carried out during the intervention, even though this consideration only partially justified a deterioration of PA level after an intervention planned to prevent it. In conclusion, the intervention seemed to have a durable improvement on DP, an unexpected, even though transitive, detrimental effect on PA and no effect on EE.

It is worth noting that the intervention seemed to have an effect on the burnout index showing the highest level at the beginning of the intervention, namely DP. The intervention seemed not to be effective on medium-low level of burnout showed by EE and PA. The questionnaire on participants’ satisfaction showed that 60% of workers considered the intervention promoting assertiveness very useful, 30% quite useful and the 10% not useful enough. As specified above, an ad hoc questionnaire concerning the application of techniques acquired during the intervention was also submitted to the participants during the follow-up. A global measure of effectiveness was calculated adding together the scores obtained by questions 2 (a, b, c) and 3 (a, b, c) explicitly dedicated to the assertiveness techniques experienced in workshops two to six. The obtained measure was correlated to the burnout indexes evaluated at T2. The global effectiveness of the techniques acquired during the intervention was negatively correlated with DP (r = −0.42, P < 0.05), confirming the relationship between the intervention and this dimension of the burnout. Neither EE nor PA significantly correlated with the global measure of effectiveness obtained by the ad hoc questionnaire.

It is well known that the effects of burnout workshops are successful in reducing levels of EE, even across relatively long periods (e.g. Shaufeli and Enzmann 1998). However, most of these workshops focused on reducing emotional activation. On the contrary, our workshops addressed the issue of managing interpersonal relationships, which have been found to be related largely to DP (e.g. Demerouti et al. 2001; Le Blanc et al. 2007). Our results also showed that the improvement in DP level was stable across time.

Finally, it is essential to illustrate two methodological issues limiting the contribution of the present paper: the absence of a control group restraining the effect of confusing variables, and the small sample size that threatens the precision to provide reliable answers. However, two considerations could restore quite confidence about our results. First, the expected effect of intervention was restricted to DP and it did not involve generically each dimension of burnout. Second, the improvement seemed to be: (a) stable across time, as confirmed by the very low level of DP at T2, and (b) meaningfully related with data obtained by the questionnaire handed out at T2.

Conclusions

In one of the most cited studies on burnout, Ramirez et al. (1996) pointed out that the feeling of being poorly managed and dealing with patients, their relatives and members of staff was considered as an important source of stress at work. In the same paper, burnout was reported as more prevalent among professionals who felt inadequately trained in communication and management skills. Our intervention was specifically thought to provide training on these topics and it can be considered one of the first studies promoting an intervention on burnout at interpersonal level in an Italian sample of mental health workers.

Briefly, the results showed the substantial stability of EE and PA together with a durable reduction of DP. Moreover, the reduction in DP level was associated with the perceived effectiveness of the intervention at the follow-up. Overall, we consider the intervention as partially successful. Indeed, it seemed to work only on that dimension of burnout showing a medium-high level before the intervention—i.e. DP—while the expected effect on EE did not appear. Our explanation is related to the relative level of burnout indices among our non-clinical sample of professionals. Following Le Blanc et al. (2007), if participants are characterized, at the beginning of the intervention, by a relatively low level of a particular index, it is less plausible that the intervention could have a large effect. Given that our intervention already had a reducing effect on DP, one might have expected a larger decrease in (lack of) PA and mainly in EE in samples of participants who severely suffer from burnout.

Nonetheless, results of studies on burnout interventions that failed to modify the level of EE are not new. For instance, Taormina and Law (2000) found that the skills of stress management largely correlated with DP and PA, but not with EE, and interpersonal skills were the strongest predictors of DP. More recently, Cohen and Gagin (2005), promoting training on intervention skills in two groups of social workers, found a larger decrease in DP and PA than in EE. In light of previous findings, we can conclude that an intervention program aimed at encouraging all workers to promote assertive communications, positive feelings and agreement among colleagues may reduce workers’ inclination to depersonalize the relationships with clients (e.g. Zellars et al. 2000). On the other hand, harmony and agreement among colleagues may increase the perception of social support that is associated with lower levels of burnout (e.g., Lee and Ashforth 1993). Further research is needed to better distinguish the interpersonal dynamics and their relationships with the emergence of the burnout syndrome. Replications of the reported intervention are required and they should include a control group. Experimental and control groups should be larger than those used in the present study. Participants with higher levels of EE should also be considered in order to understand if the lack of effects on EE is specific (i.e. our intervention is not effective on EE) or general (i.e. a low level in a particular burnout index prevents to obtain any effect on it by the intervention). Taking into account the aforementioned methodological limitations regarding control group and sample size, the present study might be viewed as an initial guideline for the dissemination of good practices in Italian public and private organizations providing assistance/rehabilitation to people with mental illness.