Staff Turnover in Assertive Community Treatment (Act) Teams: The Role of Team Climate

  • Xi Zhu
  • Douglas R. Wholey
  • Cindy Cain
  • Nabil Natafgi
Original Article

DOI: 10.1007/s10488-016-0740-7

Cite this article as:
Zhu, X., Wholey, D.R., Cain, C. et al. Adm Policy Ment Health (2017) 44: 258. doi:10.1007/s10488-016-0740-7

Abstract

Staff turnover in Assertive Community Treatment (ACT) teams can result in interrupted services and diminished support for clients. This paper examines the effect of team climate, defined as team members’ shared perceptions of their work environment, on turnover and individual outcomes that mediate the climate-turnover relationship. We focus on two climate dimensions: safety and quality climate and constructive conflict climate. Using survey data collected from 26 ACT teams, our analyses highlight the importance of safety and quality climate in reducing turnover, and job satisfaction as the main mediator linking team climate to turnover. The findings offer practical implications for team management.

Keywords

Assertive community treatment Staff turnover Team climate Job satisfaction Burnout 

Introduction

Staffing continuity is a critical program ingredient for Assertive Community Treatment (ACT) because ACT uses a team approach for serving clients with severe mental illness in community settings (Bond et al. 2001; Stein and Santos 1998). Failure to achieve full and continuous staffing can result in interrupted services, reduced quality, and diminished support for clients (Teague et al. 1998). High rates of staff turnover can compromise team operation, impair staff morale, and induce extra costs as the remaining staff must share a heavier workload while new team members are hired and trained (Boyer and Bond 1999; Monroe-DeVita et al. 2012).

Prior research has investigated program characteristics such as fidelity, caseload, service delivery patterns, and team composition associated with staff turnover, burnout, and job satisfaction in ACT (Boyer and Bond 1999; Rollins et al. 2010; Salyers and Bond 2001). This research shows that overall ACT fidelity is negatively associated with turnover. Compared with traditional case management workers, ACT staff experienced lower levels of burnout and higher levels of job satisfaction—two important individual outcomes related to turnover. This paper extends this literature by examining the effects of team climate on turnover and the individual-level factors that mediate the climate-turnover relationship.

We draw on turnover research in mental health services (Aarons and Sawitzky 2006; Glisson et al. 2008; Rollins et al. 2010; Woltmann et al. 2008) as well as general turnover theories (Holtom et al. 2008; Hom and Griffeth 1995) to hypothesize and test relationships between two team climate dimensions, safety and quality climate and constructive conflict climate, and turnover in ACT. We focus on these climate dimensions because they have strong theoretical implications for team members’ job commitment and affective states. Team climate is a manageable aspect of team functioning. Evidence of its impact on staff and program outcomes will contribute to the contemporary ACT literature, which has advanced from program/structural features of ACT to team processes (Monroe-DeVita et al. 2012; Monroe-DeVita et al. 2011; Wholey et al. 2012). Further, we integrate two individual-level predictors, job satisfaction and burnout, in a multi-level analysis to test individual-level mechanisms that may mediate the team climate effects.

In the following sections, we develop hypotheses regarding the cross-level effects that link team climate to staff turnover in ACT and test the hypotheses using data collected from 26 ACT teams. Our results highlight the importance of safety and quality climate in reducing turnover, and job satisfaction as the main mechanism through which safety and quality climate affects actual turnover. The paper concludes with a discussion of practical strategies that ACT administrators and team leaders can use to manage the teams’ work climate to retain high-quality ACT team members.

Theory

Given the strategic importance of retaining high-quality employees, turnover has been studied extensively in various organizational settings (Holtom et al. 2008; Maertz and Campion 1998). It is well recognized that turnover is a multi-level phenomenon influenced by contextual, organizational, and individual factors (Hom and Griffeth 1995), and that employees exhibit a range of motives and processes in leaving their jobs (Maertz and Campion 2004). It has recently been advocated that more research is needed to integrate factors across these different levels to enrich our understanding of turnover behaviors (Holtom et al. 2008). In mental health services, previous research has examined factors associated with turnover at all three levels such as work environment, program characteristics, and individual burnout (Aarons and Sawitzky 2006; Boyer and Bond 1999; Glisson et al. 2008; Rollins et al. 2010; Salyers and Bond 2001). However, the relationships between these factors across different levels are usually not examined together. We explore the cross-level relationships in this study by asking: (1) how does team climate affect individual turnover in ACT teams; and (2) what individual factors mediate the relationship between team climate and turnover?

Organizational climate is a set of enduring characteristics of an organization’s internal environment that distinguish it from other organizations and influence its employees’ behaviors (Pritchard and Karasick 1973). Carr et al. (2003) conducted a meta-analysis showing that climate perceptions affect employees’ job performance, well-being, and intention to leave. In mental health service organizations, Aarons and Sawitzky (2006) reported that demoralizing climate, characterized by high levels of depersonalization and emotional exhaustion, partially mediated the impact of organizational culture on satisfaction and turnover.

Many climate dimensions have been proposed in the organization and team literature (Schneider et al. 2013). Two climate dimensions—safety and quality climate and constructive conflict climate—are salient for turnover behaviors in ACT teams given their potential impacts on employee commitment and affect. Safety and quality climate is a shared perception of the relative importance of safety and quality goals in a given work environment. Pressure to achieve multiple goals such as patient/staff safety, high-quality services, and productivity is commonly experienced by healthcare providers. These goals involve trade-offs as providers need to place one goal over the others as their primary priority in certain circumstances (Gaba et al. 1994). Prior studies suggested that organizational policies and leadership practices (e.g., feedback) that prioritize certain goals can cultivate a climate related to goal orientation, which affects how individuals allocate resources and perform on their jobs (DeShon et al. 2004; Zohar 2000).

Although not specifically studied in the turnover literature, safety and quality climate has several implications for turnover behaviors in ACT teams. First, safety and quality goals are job characteristics that contribute to employees’ intrinsic motivation and satisfaction by enhancing experienced meaningfulness of the work (Hackman and Oldham 1976). Intrinsic reward motivated by a desire to serve those in need is a significant factor influencing work behaviors in healthcare teams (England 2005). Organizations’ commitment to safety and quality goals that align with these core values has been found to alleviate negative reactions to work stressors and reduce turnover intention among healthcare workers (Janssen et al. 1999; Valentine et al. 2011). Second, an emphasis on safety also entails organizational commitment and policies that assure staff safety in the healthcare setting (Zohar et al. 2007). The community environment in which ACT teams operate often raises concerns about the safety of ACT team members in emergency situations (Phillips et al. 2001). Commitment to staff safety will increase team members’ perceptions that they can provide care in a manner that supports their own safety. Thus, we expect that a team climate prioritizing safety and quality goals will reduce staff turnover.

Hypothesis 1

Safety and quality climate will be negatively associated with staff turnover in ACT teams.

Intragroup conflict exists when there are perceived incompatibilities or differences among group members (de Wit et al. 2012). Unmanaged conflict can result in negative affect, which exacerbates the negative effect of conflict on group outcomes including irrational decisions, poor performance, and turnover (Jehn and Bendersky 2003). Research shows that relationship conflict (i.e., interpersonal clashes of personalities, values, beliefs etc.) has negative effects, but task conflict (i.e., differences in ideas and opinions about the tasks being performed) has mixed effects on performance and team member satisfaction (De Dreu and Weingart 2003; de Wit et al. 2012). Jehn and Bendersky (2003) propose a model of constructive conflict suggesting that the conflict-outcome relationship is moderated by the circumstances under which conflict occurs and the processes with which the conflict is managed. Specifically, low relationship conflict and well-managed task conflict allow teams to take advantage of the constructive aspects of conflict including “enhanced information sharing, critical evaluation of divergent opinions, and increased task focus” (Jehn and Bendersky 2003, p. 213).

In ACT teams, constructive conflict climate can be defined as a team environment that has low relationship conflict and supports constructive management of task conflict. Constructive conflict climate encourages team members to express their opinions professionally and explore opposing ideas to understand the problem open-mindedly. Such a climate is expected to reduce conflict-related stress experienced by staff, which was found to increase emotional exhaustion, absenteeism, and turnover intentions among social services workers (Giebels and Janssen 2005).

Hypothesis 2

Constructive conflict climate will be negatively associated with staff turnover in ACT teams.

The cross-level effects that link team climate to individual turnover are likely to be mediated through individual attitudes and emotional reactions toward the work environment (Holtom et al. 2008). Theories of turnover suggest that job satisfaction plays a central role in the processes leading to turnover (Holtom et al. 2008; Lee et al. 1999; Tett and Meyer 1993). Prior research generally finds that job satisfaction is negatively associated with intentions to leave and actual turnover (Harrison et al. 2006), and mediates the impact of work and organizational factors. Another important personal condition that can lead to turnover is burnout (Jackson et al. 1986; Wright and Cropanzano 1998). Mental health workers are particularly prone to burnout (Leiter and Harvie 1996; Pines and Maslach 1978). Aarons and Sawitzky (2006) revealed that emotional exhaustion and depersonalization (two subscales of burnout, Maslach and Jackson 1981) were negatively related to job satisfaction and positively related to turnover among mental health workers.

We speculate that safety and quality climate may reduce ACT team members’ job dissatisfaction and burnout through two mechanisms: (1) supporting goals consistent with team members’ intrinsic values to sustain their enthusiasm toward work and clients (Pines and Maslach 1978); and (2) mitigating accumulated negative attitudes that arise as a result of job demands and unsafe work environments (Nahrgang et al. 2011). Constructive conflict climate, on the other hand, may reduce stress and discomfort associated with task and relationship conflicts, which have been shown to contribute to lower satisfaction (Jehn 1995) and higher burnout (De Dreu et al. 2004). These effects will in turn reduce turnover in ACT teams.

Hypothesis 3

The negative relationship between safety and quality climate and staff turnover will be mediated by individual job satisfaction and burnout in ACT teams.

Hypothesis 4

The negative relationship between constructive conflict climate and staff turnover will be mediated by individual job satisfaction and burnout in ACT teams.

Methods

Data

Survey data were collected using a pooled cross-sectional and time series design. The sample consisted of staff members from 26 ACT teams in Minnesota. Data were collected from three sources. First, team members and team leaders completed a survey at baseline, 6, and 12 months to provide information on team-climate perceptions, work relationships, job satisfaction, burnout, and demographics. Second, team leaders responded to a separate survey that assessed team-level characteristics such as team size, caseload, and organizational support during each survey wave. Team leaders were also asked to report any staff turnover that occurred during the 6 months after each survey wave. Third, the state’s Department of Human Services assessed the program fidelity of each ACT team using a scale closely modeled after the Dartmouth ACT Scale (Teague et al. 1998; Witheridge 2010). We used the most recent fidelity scores assessed before the study period in our analysis.

Both team- and individual-level participation rates were high. We invited all 27 Minnesota ACT teams to participate in the study, and 26 teams agreed (96 %). The total numbers of staff members surveyed were 318, 309, and 304 for the three waves respectively. All respondents gave their informed consent before completing the survey. We received 287 (90 %), 268 (87 %), and 275 (91 %) completed surveys for the three waves. The racial composition of the overall sample was 91 % Caucasian, 5 % Asian, 2 % African American, and 1 % other race and ethnicity. The gender composition was 71 % female and 29 % male.

Measures

Turnover, the dependent variable of this study, was measured as 1 for individuals who had left the ACT team during the 6 months after each survey wave and as 0 for those who had stayed. We used climate and individual measures collected at baseline, 6, and 12 months to predict turnover that occurred between 0–6, 6–12, and 12–18 months.

Safety and quality climate was measured with a 3-item scale that assessed the extent to which the ACT team emphasized safety, quality, and productivity goals (Zohar 2000; Zohar et al. 2007). Items include “I felt that I had the time during visits with consumers to assure safety for me or the consumers, even if it meant visiting fewer consumers,” “I felt that I had the time to assure high quality visits with consumers, even if it meant visiting fewer consumers,” and “I felt that I had to make as many consumer visits as possible, even if it meant lower quality visits or less personal or consumer safety.” This measurement used a 4-point scale ranging from 1 (“Strongly disagree”) to 4 (“Strongly agree”). A factor analysis indicated unidimensionality with the safety and quality items loaded positively and the productivity item loaded negatively. The Cronbach’s alpha for this scale was 0.78. A high safety and quality climate score thus indicates more emphasis on the safety and quality goals and less emphasis on the productivity goal. We aggregated this measure to the team level after testing the within-team agreement (ICC1 ranging from 0.13 to 0.17 over three survey waves).

Constructive conflict climate was measured using a 6-item scale modified from previously validated scales of intragroup conflict (Jehn 1995) and constructive controversy (Shah et al. 2006; Tjosvold et al. 1986). The scale items were adapted to the ACT context based on feedback we received from ACT team leaders and team members on the original scales. Items include “I felt that my ACT team members did NOT get along,” “I felt that personality clashes were evident in my ACT team,” “I felt that conflicts regarding ideas frequently arose in my ACT team,” “I felt that my ACT team members had different opinions about how to organize work,” “I felt that even when we disagreed on my ACT team, we communicated with respect for each other,” and “I felt that on my ACT team we used our opposing views to understand problems.” The measurement used a 4-point scale ranging from 1 (“Strongly disagree”) to 4 (“Strongly agree”). A factor analysis indicated unidimensionality with the first four conflict items loaded negatively and the two constructive controversy items loaded positively. The Cronbach’s alpha for this scale was 0.82. A high constructive conflict climate score indicates low relationship conflict and an emphasis on using divergent views to improve team functioning. We aggregated this measure to the team level (ICC1 ranging from 0.25 to 0.29 over three survey waves).

Job satisfaction was measured with a single item asking team members to assess their overall satisfaction with their job on the ACT team. The satisfaction scale ranged from 1 (“Very dissatisfied”) to 4 (“Very satisfied”).

Burnout was measured using an abbreviated version of the Maslach Burnout Inventory (MBI) to assess overall burnout and three subscales: emotional exhaustion, depersonalization, and personal accomplishment. To minimize respondent burden, we selected 9 items (3 for each subscale) from the 22 original MBI items that had high factor loading onto their respective subscales in the original study (Maslach and Jackson 1981) and were most relevant for ACT. Emotional-exhaustion items include feeling “emotionally drained from work,” “used up at the end of the workday,” and “working too hard on the ACT team.” Depersonalization items include “treating some consumers as if they were impersonal objects,” “become more callous toward people,” and “don’t really care what happens to some consumers.” Personal accomplishment items include “dealt very effectively with consumer problems,” “positively influencing other people’s lives,” and “had accomplished many worthwhile things,” which were reverse-coded to measure burnout. In this study, we created an overall burnout score by averaging across the three subscales, and a sensitivity analysis was performed to verify that using the overall score generated analytical results that were consistent with using separate subscale scores. The Cronbach’s alpha for the overall scale was 0.74.

Individual-level control variables included gender (1 for female and 0 for male), tenure (measured as the number of years that a team member has worked on the ACT team), and team leader (1 for team leader and 0 for other team members), and work interdependence. Work interdependence was measured based on team members’ assessment of how closely they worked with each of the other team members during the past month, using a 4-point scale (1-not closely, 2-a little closely, 3-somewhat closely, and 4-very closely). A work relationship was coded as 1 for a pair of team members if they both responded that they worked somewhat or very closely with each other; and as 0 otherwise. For each individual, work interdependence was measured as the number of work relationships he or she had divided by team size minus one (i.e., proportion of other team members that an individual had close work relationship with). This individual-level control was included because previous research showed that work relationships could lead to commitment, positive emotion, and group cohesion (Lawler et al. 2000; Lawler and Yoon 1996). At the team level, the work interdependence and group cohesion resulted from it are related to team climate; but because our focus was to test team climate effects, we only included work interdependence as a control variable.

Team-level control variables included ACT fidelity, team size, urban team, client-staff ratio, and organizational support. ACT fidelity was measured on a 5-point scale based on the Dartmouth ACT Scale (Teague et al. 1998), assessing the extent to which team structures and procedures adhere to the ACT standards. Team size was measured as the number of team members. Urban team was coded as 1 for teams operating in urban areas and 0 for teams in rural areas. Client-staff ratio was measured as the number of clients that the team served in the 6 months before each survey wave divided by the number of team members. Organizational support was measured using a 7-item scale asking the team leaders to evaluate the extent to which they perceived that the team had control over its work and could obtain support. Sample items include “I felt that my team could obtain support from our parent organization” and “I felt that my team had control over management decisions internal to my team.” The measurement used a 4-point scale ranging from 1 (“Not at all”) to 4 (“A lot”). Perceived organizational support was found to be negatively associated with turnover (Maertz et al. 2007), and thus should be included as a control variable. The Cronbach alpha for this scale was 0.78.

Analysis

Individuals with missing data were excluded from the analysis using list-wise deletion. The analytical sample size was 776. The analytical sample and the deleted observations were not systematically different in terms of turnover, satisfaction, or burnout (F-test = 0.26, 0.63, and 1.90 respectively). We used multivariate regression analyses to test our hypotheses and followed the steps suggested by Baron and Kenny (1986) to estimate the direct and mediated effects of team climate on staff turnover. Because team members were nested within teams and observed repeatedly over time, which resulted in a multi-level data structure and potential dependency among observations (Rabe-Hesketh and Skrondal 2012; Raudenbush and Bryk 2002), we included a random team effect and a fixed wave effect in all models. Stata’s GLLAMM (Generalized Linear Latent and Mixed Models) command was used to estimate the multi-level generalized linear models for job satisfaction, burnout, and turnover (Rabe-Hesketh et al. 2005). Ordered logistic and logistic links were used in GLLAMM for models predicting job satisfaction and turnover respectively because job satisfaction was an ordinal variable and turnover was a binary variable. In addition, we applied methods developed by Imai and colleagues (Imai et al. 2011; Imai et al. 2010) to test the causal mediation effects in the hypothesized relationships using Stata MEDEFF and MEDSENS commands.

Results

Table 1 presents descriptive statistics and correlations among all variables. The six-month turnover rate for ACT team members in our sample is 14 %. We first investigated whether there was a difference in the likelihood of turnover between respondents and non-respondents with a cross-tabulation. A significant difference was observed with non-respondents being more likely to leave the team in the following 6 months (χ2 = 13.35, p < .001). Table 1 shows that, at the bivariate level, safety and quality climate and constructive conflict climate both are positively correlated to job satisfaction and negatively correlated to burnout. Safety and quality climate is also negatively correlated to turnover.
Table 1

Descriptive statistics and correlations

 

Mean

SD

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(1)

Turnover

0.14

0.34

             

(2)

Job satisfaction

3.31

0.73

−0.22***

            

(3)

Burnout

2.08

0.53

0.14***

−0.56***

           

(4)

Safety and quality climate

2.98

0.38

−0.15***

0.24***

−0.25***

          

(5)

Constructive conflict climate

2.35

0.34

−0.06

0.18***

−0.18***

0.39***

         

(6)

Female

0.72

0.45

0.01

0.04

0.13***

−0.05

0.08*

        

(7)

Tenure

2.31

1.44

−0.18***

−0.03

0.03

0.10**

−0.02

−0.10**

       

(8)

Team leader

0.10

0.29

−0.04

0.05

0.00

0.01

−0.01

0.04

0.05

      

(9)

Work interdependence

0.65

0.24

−0.14***

0.14***

−0.01

0.20***

0.14***

0.06

0.15***

0.23***

     

(10)

Fidelity

4.18

0.19

−0.00

0.01

0.03

0.04

−0.09**

−0.02

0.07

−0.02

0.12**

    

(11)

Team size

12.45

2.53

0.05

−0.02

−0.00

−0.19***

0.04

−0.01

0.01

−0.06

−0.18***

0.16***

   

(12)

Urban team

0.67

0.47

0.11**

−0.10**

0.06

−0.34***

−0.10**

0.02

−0.17***

−0.03

−0.13***

−0.23***

0.33***

  

(13)

Client-staff ratio

6.13

1.15

−0.01

−0.04

0.05

−0.38***

−0.03

0.08*

−0.06

0.00

0.02

−0.21***

−0.09**

0.61***

 

(14)

Organizational support

2.77

0.61

−0.06

0.17***

−0.13***

0.43***

0.34***

0.03

0.02

0.01

0.13***

0.24***

0.09**

−0.18***

−0.20***

*** p < 0.001, ** p < 0.01, * p < 0.05,  p < 0.10

Table 2 presents results from multi-level generalized linear regression models with Model 1 predicting job satisfaction, Model 2 predicting burnout, and Models 3 (unmediated) and 4 (mediated) predicting turnover. Models 1–3 show that safety and quality climate has a significant positive effect on job satisfaction, a significant negative effect on burnout, and a significant negative effect on turnover. In Model 4, after controlling for the mediators, the negative effect of safety and quality climate on turnover remains significant and the coefficient is reduced, which suggests a partial mediation (Sobel test, t = −3.08, p < .01). These results support hypothesis 1 and partially support hypothesis 3; that is, safety and quality climate is negatively associated with turnover, which is mediated by job satisfaction, but not by burnout.
Table 2

Regression analyses predicting job satisfaction, burnout, and turnover

Variable

Model 1—satisfaction

Model 2—burnout

Model 3—turnover-unmediated

Model 4—turnover-mediated

Effect

SE

Sig.

Effect

SE

Sig.

Effect

SE

Sig.

Effect

SE

Sig.

Job satisfaction

         

−0.76

0.15

***

Burnout

         

0.16

0.22

 

Safety and quality climate

0.99

0.25

***

−0.36

0.07

***

−1.61

0.40

***

−1.16

0.42

**

Constructive conflict climate

0.27

0.24

 

−0.11

0.07

0.23

0.35

 

0.32

0.38

 

Female

0.08

0.16

 

0.16

0.04

***

−0.37

0.24

***

−0.31

0.24

 

Tenure

−0.11

0.05

*

0.03

0.01

*

−0.29

0.08

 

−0.38

0.09

***

Team leader

0.44

0.26

−0.03

0.06

 

−0.14

0.37

 

−0.05

0.36

 

Work interdependence

0.77

0.33

*

0.08

0.09

 

−0.35

0.48

 

−0.47

0.50

 

Fidelity

−0.28

0.42

 

0.09

0.12

 

0.45

0.59

 

0.17

0.67

 

Team size

0.06

0.04

 

−0.01

0.01

 

−0.10

0.06

−0.08

0.06

 

Urban team

−0.53

0.24

*

0.04

0.07

 

1.19

0.44

**

0.88

0.47

Client–staff ratio

0.19

0.10

−0.02

0.03

 

−0.37

0.14

*

−0.33

0.16

*

Organizational support

0.25

0.14

−0.01

0.04

 

0.00

0.21

 

0.04

0.24

 

Wave 2

0.21

0.19

 

−0.10

0.05

*

−0.30

0.27

 

−0.33

0.28

 

Wave 3

0.09

0.21

 

−0.08

0.05

 

−0.36

0.28

 

−0.35

0.30

 

Log likelihood

−760.68

−569.06

−235.17

−212.28

Likelihood ratio R-squared

0.09

0.10

0.17

0.22

*** p < 0.001, ** p < 0.01, * p < 0.05,  p < 0.10

Regression results do not support hypotheses 2 or 4 regarding the total and mediated effects of constructive conflict climate on turnover. As shown in Table 2, constructive conflict climate predicts neither turnover nor job satisfaction. Constructive conflict climate has a marginally significant and negative effect on burnout.

To further explore the role of job satisfaction and burnout in leading to turnover, we conducted a post hoc analysis to investigate burnout and job satisfaction as consecutive mediators. We first tested burnout as an independent mediator without job satisfaction, and found that burnout significantly predicted turnover (b = 0.77, p < .001) and mediated the relationship between safety and quality climate and turnover (Sobel test, t = −3.23, p < .01). Because such effects diminished in the concurrent model, we explored the possibility that the effect of burnout on turnover is fully mediated by job satisfaction (a consecutive mediation model). The results suggest that job satisfaction and burnout are correlated, but there is no evidence supporting that they affect turnover consecutively. Overall, job satisfaction is a stronger predictor of turnover than burnout.

Among the control variables, tenure is negatively associated with job satisfaction and turnover, and positively associated with burnout. Members of urban teams are more likely to leave their teams and have lower job satisfaction. These results are consistent with prior research findings that turnover is more likely to occur among employees early in their tenure with their organizations (Hom and Griffeth 1995; Hom et al. 2008), and with more alternative job opportunities such as in the urban settings (Griffeth et al. 2005). Client-staff ratio is marginally and positively associated with job satisfaction and negatively associated with turnover.

Table 3 presents results from the causal mediation tests. When using job satisfaction as a mediator, one unit increase in a team’s safety and quality climate is associated with a significant decrease in the odds of staff turnover (OR 0.81); and 18 % of the total effect is mediated by job satisfaction. When considering burnout as a mediator, safety and quality climate yields the same total effect of which only 4 % is mediated by burnout. Constructive conflict climate does not markedly change the odds of staff turnover in both tests using job satisfaction and burnout as the mediator. We evaluated the robustness of our results to potential violations of distribution and independence (i.e., sequential ignorability in causal mediation analysis; Imai et al. 2010) assumptions. For safety and quality climate, the analysis indicated that the conclusion about the mediation effect of job satisfaction was robust unless a significant correlation existed between error terms due to omitted confounders related to both the mediator and the outcome (sensitivity parameter ρ = −0.30). In contrast, the conclusion about the mediation effect of burnout was less robust and sensitive to potential violations of the independence assumption (ρ = 0.1).
Table 3

Causal mediation analyses

Predictor

Mediator

(1) Direct effect

(2) Total effect

(3)  % of total effect mediated

OR (95 % CI)

OR (95 % CI)

% (95% CI)

Safety & quality climate

Job satisfaction

0.84 (0.82 to 0.89)

0.81 (0.79 to 0.86)

18 (17 to 27)

Burnout

0.82 (0.79 to 0.88)

0.81 (0.79 to 0.86)

4 (4 to 6)

Constructive conflict climate

Job satisfaction

1.03 (1.01 to 1.04)

1.03 (1.00 to 1.04)

−16 (−52 to −13)

Burnout

1.03 (1.01 to 1.04)

1.03 (1.01 to 1.04)

−3 (−7 to −3)

Discussion

Previous research has documented that mental health organizations in general and ACT teams in particular commonly experience high staff turnover rates (Rollins et al. 2010; Woltmann et al. 2008). Staff turnover results in interrupted services and diminishing quality. It is costly to replace employees and recover from productivity loss. Turnover of high-performing employees is especially detrimental to organizations. Given such implications, ACT researchers and practitioners have been strongly interested in strategies for reducing turnover. We extend the existing research on turnover by examining some manageable aspects of the team environment, namely team climate, that affect turnover to develop research-supported strategies.

Our main findings suggest that a team climate that emphasizes client/staff safety and service quality goals as priorities is associated with lower staff turnover in ACT teams. This relationship is partially mediated by individual job satisfaction. Safety and quality climate is also associated with lower burnout, another important job outcome, among ACT team members. However, we did not find evidence for the hypothesized mediating effect of burnout. We found that a team’s constructive conflict climate only marginally predicted burnout, but not job satisfaction or actual turnover. This finding suggests that conflict management may reduce job burnout, but is less essential for staff retention. These findings together give a more systematic answer to the questions of which team climate dimension is associated with staff turnover and what individual factors may mediate the relationship between team climate and turnover in ACT. Although job satisfaction and burnout were both found to be associated with turnover among healthcare workers in previous research (Estryn-Béhar et al. 2007; Jayaratne and Chess 1984), we found job satisfaction to be a stronger predictor of reduced turnover. Burnout may be associated with various forms of job withdrawal—absenteeism, intention to leave, and actual turnover. But for people who stay on the job, burnout can lead to an array of outcomes such as lower productivity and effectiveness, decreased job satisfaction, and reduced commitment to the job or the organization (Maslach et al. 2001). It is possible that burnout “affects” turnover because it is associated with job satisfaction and job/organization commitment.

Although our primary interest was the main effects of team climate on staff outcomes, recent research has raised the question: what are the implications of variability in individual perceptions of climate within the teams? Researchers have explored this question by examining whether climate strength (i.e., low variability and high consensus in climate perceptions) moderates the climate-outcome relationship such that the relationship will be stronger when climate strength is high (Colquitt et al. 2002; González-Romá et al. 2002); however, such studies have produced mixed, inconclusive results (Schneider et al. 2013). Given that our hypotheses related to constructive conflict climate were disconfirmed, we tested the possibility that this might be due to variability in climate perceptions. We did not find a significant moderating effect of climate strength for either climate dimension.

Among other team factors, we found that program fidelity was not associated with either team climate or individual outcomes (i.e., burnout, satisfaction, and turnover) in our sample. This finding was likely due to the fact that our sample teams had high levels of fidelity to ACT standards with very little variation (mean fidelity score 4.18 and standard deviation 0.19 on a 5-point scale). In addition, the fidelity measure used was based on the Dartmouth ACT Scale, which focused primarily on structural features rather than process aspects of ACT teams. Studies using the Tool for Measurement of ACT (Monroe-DeVita et al. 2011), a more contemporary fidelity measure that includes important team process items, may find fidelity to be significantly associated with staff outcomes. We found that high client-staff ratio was associated with higher job satisfaction and lower turnover. This could have resulted from the fact that client-staff ratio in the sample teams ranged from 3.7 to 9.8 where the upper limit was close to the expected ratio of 10 clients per provider as specified in the state standards. Additionally, several factors associated with potential negative effects of excessive caseloads such as fidelity and safety and quality climate have been controlled for in the analysis. This finding suggests that holding other factors constant, having the full caseload may give staff the opportunity to do what they took the job to do, which contributes to positive perceptions of the work and lower turnover. At the bivariate level, we found organizational support to be correlated to the two team climate measures (Pearson’s r = 0.43 and 0.34, p < .001) and job satisfaction and burnout (Pearson’s r = 0.17 and −0.13, p < .001). This suggests that organizational support, which includes items related to team autonomy and parent organization’s endorsement, may influence team climate and team member outcomes.

Implications for Practice and Policy

Our findings offer several practical implications for ACT administrators and team leaders. First, a team’s safety and quality climate is strongly related to staff turnover, and thus should be the focus of managerial interventions. Research suggests that safety and quality climate is a social-cognitive mediator between organizational events and employee behaviors (Zohar and Luria 2004). Team leaders plays a critical role in creating a safety and quality climate, and the most efficient way is to assess policies and procedures as a gestalt, with goal priorities as the grouping principle (Zohar 2003). In other words, ACT team leaders need to prioritize safety and quality goals by continuously rewarding and supporting behaviors that exemplify such priorities, especially when they are achieved under the pressure of competing goals. Leadership is also critical in shaping team members’ perceptions of their work as meaningful, an important psychological state preceding job commitment and reduced turnover (Aarons 2006; Hackman and Oldham 1976).

Second, parent organizations and other external factors may affect teams’ practices and climate perceptions derived from such practices. From a policy perspective, ACT administrators and parent organizations need to grant sufficient autonomy to team operation and restrain from imposing over-demanding productivity goals (e.g., in the form of a higher client-staff ratio), especially when such goals contradict ACT principles. In an event that external demands are imposed, team leaders should, but may have limited ability to, buffer the impact of external policies and demands (Morgeson and DeRue 2006).

Third, job satisfaction is the main mediating factor, compared with burnout, linking team safety and quality climate to turnover. In practical terms, although job stress and burnout are inherent aspects of mental health work, ACT administrators could, for retention purposes, focus on team conditions that contribute to fulfilling employees’ intrinsic motivations and satisfaction when certain levels of stress and burnout exist. Fostering a safety and quality climate fits with this retention strategy as it focuses on intrinsic motivations (Hackman and Oldham 1976). Unrectified and persistent burnout, however, can lead to negative job outcomes including ineffectiveness at work, decreased satisfaction and commitment, and eventually turnover (Maslach et al. 2001). Thus, burnout should also be the target of managerial interventions.

Contributions and Limitations

This study contributes to the ACT literature by: (1) focusing on specific and manageable dimensions of team climate and investigating their impact on staff turnover; and (2) integrating team- and individual-level predictors and explanations to develop more explicit implications for policy and practice.

Despite these contributions, the study has several limitations that suggest new avenues of research. First, this research focused on two selective climate dimensions that were deemed important for staff turnover. By doing so, we may have overlooked other climate dimensions such as service climate and psychological safety that can potentially influence turnover behaviors in ACT teams. Monroe-DeVita et al. (2012) suggest that no single strategy is sufficient for ensuring operational excellence of ACT teams over time and it is useful to implement a blend of different strategies. Future research should explore effects of other climate dimensions to develop a more comprehensive understanding of turnover in ACT teams. Second, although goal orientation is an important aspect of team climate, how different goals align, compete, or obliquely relate with one another is variable across different work settings. It makes defining a climate dimension related to goal orientation challenging. In this study, we found that the safety and quality goals were unified in ACT teams and the productivity goal was at the opposite end of the same dimension. Thus, we defined safety and quality climate based on the empirical evidence. Further research could be done to validate and refine this measure using other samples of ACT teams. Third, to reduce respondent burden, we used selected items from the Maslach burnout inventory, which reduced reliability of this measure. We selected the items that had the high correlations with the specific burnout dimensions in order to retain much of the reliability from the full scale. Nevertheless, with the abbreviated scale, the absence of statistically significant finding should be interpreted with caution as it may be a false negative. Fourth, we tested the hypotheses in ACT teams implemented in a single state, which reduced the generalizability of our findings. We tried to enhance the validity of our findings with theoretical articulations and rigorous methodology. Future research is needed to test the hypothesized relationships in other ACT contexts to help gauge the generalizability of our findings. Fifth, employees leave their jobs with different motivations (Maertz and Griffeth 2004). Some turnover may be due to better job opportunities rather than poor conditions in the current work environment. In this study, we could not differentiate these two types of turnover. Future research is needed to explore different types of turnover and their respective contributors. Lastly, although generally being considered as a negative outcome, turnover can have positive impacts on teams. For instance, exit of low-performing members provides an opportunity for teams to reinforce performance standards and transform team climate. Low to moderate levels of turnover may facilitate team learning as new members bring new skills and ideas and/or cause rethinking of team practices. Recent literature has recognized staff continuity to be less of a defining feature of ACT, but an important aspect of team functioning (Monroe-DeVita et al. 2011). Future research is needed to develop an understanding of different types of turnover and their respective effects on team and individual outcomes.

Acknowledgments

Funding

This study was funded by the National Science Foundation (Grant Number SES 0719257).

Compliance with Ethical Standards

Conflict of interest

Xi Zhu and Douglas R. Wholey received the above-mentioned research grant from the National Science Foundation that supported the data collection for this study. The authors declare that they have no conflict of interest.

Ethical Approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The study procedures were approved by the University of Minnesota’s Institutional Review Board. This article does not contain any studies with animals performed by any of the authors.

Informed Consent

Informed consent was obtained from all individual participants included in the study. The paper has not been presented at a meeting, and is not under consideration in any other peer-reviewed journals for publication.

Copyright information

© Springer Science+Business Media New York 2016

Authors and Affiliations

  • Xi Zhu
    • 1
  • Douglas R. Wholey
    • 2
  • Cindy Cain
    • 3
  • Nabil Natafgi
    • 4
  1. 1.Department of Health Management and Policy, College of Public HealthUniversity of IowaIowa CityUSA
  2. 2.Division of Health Policy and Management, School of Public HealthUniversity of MinnesotaMinneapolisUSA
  3. 3.Department of Health Policy and Management, Fielding School of Public HealthUniversity of California Los AngelesLos AngelesUSA
  4. 4.Department of Health Management and Policy, College of Public HealthUniversity of IowaIowa CityUSA

Personalised recommendations