Changes in VA Psychiatrists’ Attitudes About Work Environment and Turnover During Mental Health Service Enhancement

When organizations introduce new policies, expand the range of services offered, and recruit or train staff for new roles, employees may experience stress and uncertainty (1,2). Because of increased stress and uncertainty, employees often react negatively to organizational change efforts (3,4). Organizational changes can lead to adverse changes in employee perceptions of the organizational environment and job satisfaction (5).

The Veterans Health Administration (VHA) of the U.S. Department of Veterans Affairs (VA) underwent significant changes starting in 2004 to augment mental health resources while increasing direction and accountability for their use. The emphasis was partially due to the increase since 2003 in the number of patients seeking treatment upon discharge from overseas military service operations. Compared with civilians, returning veterans are diagnosed at higher rates of major depression, posttraumatic stress disorder, and traumatic brain injury (6). VHA expanded treatment programs for these and other conditions. In 2004, VHA introduced the Comprehensive Mental Health Strategic Plan, which required greater integration and coordination of mental health services into overall health care (7). Key recommendations included improving equity of access to care, developing and implementing a full continuum of recovery-oriented services, increasing substance abuse treatment services, and addressing long-term care needs (8).

In subsequent years, additional funds were devoted to the strategic plan beyond the historical funding amount (9). In 2008, VA introduced an update to the Uniform Mental Health Services Handbook, offering more than 200 initiatives and 400 mental health services. A key component was greater focus and initiative for primary care and mental health integration based on collaborative care models (10). Subsequently, more programs, such as suicide prevention and readjustment counseling, were created for returning veterans with mental health issues. Expectations for implementation of the strategic plan evolved at a rapid pace and required substantial changes in resource allocations and decision making. Facilities experienced budgetary challenges in responding quickly to the changes (11). An evaluation of the changes noted difficulties in fully assessing the extent of progress made in implementing changes and in precisely tracking the effect of any one change (12).

Because the strategic reforms occurred quickly, psychiatrists could be expected to be directly affected. They experienced new reporting relationships, reassignment to new or different clinical units, more complex work-required contacts, and increased productivity and reporting requirements. Changes in work roles and responsibilities like these have been associated with greater stress, dissatisfaction, and turnover among non–mental health staff (13,14). Moreover, individual reactions to organizational change and the specific targets of the change can have complex effects on attitudes and behaviors (15,16). Changes may enhance or inhibit employee attitudes, job satisfaction, and burnout (17,18). Employees may be more satisfied when change offers the opportunity to learn new skills and work methods (19). Poor change management and stress from the change process itself, however, can lead to less employee trust in the organization and management, lower job satisfaction, and higher resignation rates (20,21). In addition, incomplete implementation of program elements may be seen as unfulfilled promises, protracted change processes, or both, further compromising levels of trust.

On the basis of research surrounding organizational change and its influence on employees, we hypothesized that work satisfaction and workplace perceptions of psychiatrists would significantly decrease over time as complex and varied changes were implemented. We examined how workplace perceptions and satisfaction changed since the strategic plan was introduced in 2004. To our knowledge, this is one of the largest psychiatrist samples used over time within a single health care organization.

Methods

This was a cross-sectional study with data collected from 139 VHA medical centers from 2004 to 2010, with the exception of 2005. We obtained data on work environment and satisfaction from the annual All Employee Survey (AES), which is a census survey of eligible full-time employees (N>212,000 in 2004 and N>254,000 in 2010 for VHA).

VHA administered the AES in the spring of 2004 and annually starting in 2006. The AES contains sections designed to assess perceptions at three levels: individual, work group, and organization. We focused on individual and work group perceptions. The same survey questions were included each year of the survey. Respondents were not directly compensated for participation, but some facilities provided incentives based on response rates at the work group or facility level. Surveys were completed via one of three modalities: Internet, interactive voice response by telephone, or paper and pencil. More than 90% of respondents used the Internet-based survey. We identified psychiatrists on the basis of self-identified occupation responses. Survey responses were anonymous; thus, respondents were not linked between successive survey administrations.

Measures

Individual satisfaction and job characteristics.

We examined ratings of overall job satisfaction using a single item from the AES, “Compared to what you think it should be, what is your current overall level of satisfaction with your job?” Level of satisfaction was indicated on a 5-point Likert scale ranging from 1, not at all satisfied, to 5, very satisfied. Assessment with single-item measures has been shown to be valid (22). Research has found positive associations between provider satisfaction levels and patient satisfaction (23,24) and greater care adherence (25). Recent pay satisfaction research shows considerably lower compensation rates for staff in mental health in both public and private behavioral health care organizations (26). We also assessed satisfaction with amount of work and senior management. Because turnover and lower morale are concerns during organizational change, we examined intention to leave the workplace. Because organizational change may require developing new skills and changes in job control, we also examined perceptions of developing new job skills and job autonomy with two separate AES items. Respondents rated all three items using a 5-point Likert scale ranging from 1, strongly disagree, to 5, strongly agree. [Item content is reported in an online data supplement to this article.]

Work group perceptions.

We selected measures from prior research on the AES (27) that focused on respondents’ evaluation of factors related to their work group: workplace civility (k=8, α=.93) and management for achievement (k=5, α=.94). Respondents rated scale items on a 5-point Likert scale ranging from 1, strongly disagree, to 5, strongly agree.

Measures were also supported and identified from prior research examining well-being of psychiatrists within the work environment (28,29). Workplace civility represents courteous and respectful workplace behaviors, such as showing personal interest in coworkers, cooperation, fair resolutions to conflict, and valuing differences among individuals (30). Individuals experiencing job strain may be less likely to cooperate with one another (31). In addition, positive working relationships among staff may support successful quality improvement initiatives (32).

Management plays an essential role in making quality improvement efforts (33). Management actions, such as providing a vision, supporting employees, and modeling behavior are important during organizational change (34). Managers also are seen as principally responsible for acquiring and managing resources in order to ensure competitive advantage (35).

Turnover.

Psychiatrist turnover was obtained from fiscal year (October to September) 2004 to 2010 from a VA human resources database. The overall turnover rate was computed as the number of employee departures divided by the average number of full-time–equivalent employees for the year. Similar to the AES data set, this database excluded contractors and fee-basis employees.

Covariates.

We modeled individual-level factors that could influence survey ratings: gender, race, ethnicity, age, tenure, and supervisory level (34). Using work group codes assigned by the survey coordinator, we classified respondents into either an inpatient or an outpatient setting, which may be sensitive to organizational changes in different ways. Respondents were also classified as being an employee of either a medical center or a community-based outpatient clinic to capture any structural and procedural differences (36,37).

At the facility level, we modeled for medical center complexity on the basis of 2008 values. VHA classifies facilities into five unique complexity levels based on patient risk, patient volume, teaching, and research activity. Complexity level has been used in studies involving safety climate (38), team training programs (39), and readiness to change (40). We also modeled the ratio of psychiatric beds to total hospital beds. Because VHA is organized into 21 regional networks, we treated each network as a fixed effect to control for variation in clinical or administrative practice guidelines. The network variable also served as a measure for geographical region.

Statistical analysis

Multilevel analyses performed with SAS 9.2 PROC GLIMMIX nested data for individual psychiatrists within facilities. For the turnover model, facilities were nested within regional networks, and we controlled for facility-level characteristics. We obtained adjusted mean score values for each measure for each survey year, accounting for individual and organizational characteristics and study year. We regressed satisfaction measures on time (year) using ordinary least-squares (OLS) models in order to measure the significance of linear trends in satisfaction over the study period. We applied a Tukey-Kramer post hoc adjustment to assess statistically significant differences between individual years (2004–2006, 2006–2007, and 2007–2008). The study was approved by the VA Boston Healthcare System Institutional Review Board.

Results

Descriptive statistics for the respondents are reported in Table 1 for the aggregate study period. We obtained 7,218 responses over all years. The number of respondents per survey year ranged from 739 to 1,596. The response rate among full-time psychiatrists, identified from the VA database, ranged from 48% to 74% and averaged 65% across years.

Adjusted mean values for survey measures from the multilevel regression models and significance tests of the OLS national linear trend (that is, satisfaction regressed on time) are reported in Table 2. Five of the satisfaction measures exhibited a significant upward national linear trend between 2004 and 2010 (Figure 1) and four of the measures remained flat (Figure 2). Those with upward trends tended to have the greatest significant change between 2004 and 2006 and remained relatively flat during 2006–2010; the increase in adjusted mean scores across all AES measures ranged from .13 to .25 in 2004–2006 and from .02 to .12 for 2006–2010.

Satisfaction with senior management was significantly higher in 2006 and 2008 than in 2004. Pay satisfaction was significantly higher in all years subsequent to 2004. Both skill development and workplace civility were significantly higher than in 2004 in all years except 2007. Management for achievement was rated higher in all years after 2004. Intention to leave, job autonomy, and turnover rates did not significantly change. Satisfaction with amount of work was significantly lower in 2007 and 2010 than in 2006. The pattern for overall job satisfaction differed; values in 2010 were significantly lower than in 2006.

Discussion

Satisfaction ratings and work environment perceptions generally increased over time among the cohort of psychiatrists responding to the AES between 2004 and 2010. Most of the increase in survey measures occurred between 2004 and 2006, with gains maintained thereafter. Although even positive changes for an organization and customers can have unintentional negative consequences, our study suggests that VHA psychiatrists were not negatively affected by the change transformation and may have been positively affected despite the rapid pace of change, increased expectations, and implementation of new resources and services. Measures assessing management practices—satisfaction with senior management and management for achievement—showed the largest trend increase, suggesting that beneficial changes in management practices occurred during the transformation, which could explain more positive ratings on other aspects of the job. Findings generally support research emphasizing the role of management in the transformation process and invite closer study of management behaviors during transformations.

Job autonomy did not differ over time, a somewhat surprising finding considering that organizational changes often involve greater standardization and centralization of clinical and administrative practices. The strongest trend increase was for pay satisfaction. The increase may have been a result of changes to physician compensation to include performance pay (41) or changes in the non-VA labor market. Satisfaction with skill development also increased, which may suggest recognition and appreciation of the need for skill acquisition in response to a dynamic, changing organizational environment.

Psychiatrists tend to be among the more satisfied physician specialties (42,43). We were unable to find information on whether non-VHA psychiatrists were becoming more or less satisfied during the study period. A Norwegian study reported that physicians who changed job positions over a five-year period, such as through a promotion, were more satisfied than physicians who remained in their current position (44). Although we do not know whether respondents changed job positions, this may have partially influenced the findings. This study also has implications for VHA plans to continue expansion of services and mental health staffing (45). Generally positive incumbent attitudes toward the workplace may help in the recruitment of new employees (46).

Dyrbye and Shanafelt (47) noted that reform activities may lead to increased levels of work and family conflict, loss of job control, and burnout among physicians. Further research into management practices associated with sustained positive perceptions during the implementation and maintenance phases reported here may be informative in light of ongoing national reform challenges. Innovations such as the patient-centered medical home, accountable care organization, and electronic health record implementation require substantial modifications in infrastructure, job requirements, and staffing. Management activities that minimize the negative impacts of change efforts on employee morale or even improve morale are matters of great importance, particularly because employee satisfaction can influence patient experiences of care (48,49) and technical quality of care (50,51).

It is unclear, however, which specific mental health organizational changes led to more positive perceptions among psychiatrists. Improved satisfaction may be due to some level of standardization of clinical practice across regional networks (due to leadership practices), performance measures, or greater attention to developing databases, measuring outcomes, and information technology generally. Alternatively, a greater focus and emphasis on collaborative care may have led to more satisfaction with working conditions, relationships, and care plans among psychiatrists (52,53).

On the other hand, it is also possible that psychiatrists in this study were able to cope with the stress and changes (54) more effectively. VHA physicians may also be more tolerant of oversight and constraints on clinical practice. Glassman and colleagues (55) reported that 73% of 3,682 VHA physicians agreed that drug formulary restrictions were important for containing costs. Similar views reported by physicians working in private managed care settings suggest that physicians self-select into practice settings that most closely reflect their preferences (56). Salaried physicians of health maintenance organizations (HMOs), compared with those with contractual HMO relationships, reported higher satisfaction and were less likely to report a lack of clinical freedom, income pressure, and absence of continuity of care compared with those with contractual relationships (57,58).

We note several limitations. First, our analysis was limited by lack of information regarding independence of observations across years. Because the surveys were conducted anonymously, we did not know whether individuals were repeat respondents over multiple years and could not control for repeated measurements by individual psychiatrists. This increased the probability of making a type I error. Second, we did not assess how changes, such as in job roles or responsibilities, directly affected staff or the extent of provider involvement or openness to the change. A more direct assessment of these changes would have been useful in explaining the pattern of results at a provider level. Third, we did not directly assess the extent of the changes occurring within facilities or whether some changes may have been more important than others. Activities to provide collaborative care may explain some of the positive change, but because of the large number of changes made over this period, including heterogeneous changes at the facility level, we were not able to test this hypothesis directly.

Some facilities may have been more adept at allocating resources, recruiting staff, and starting new programs. There was wide variation in how the funds were utilized in the field, which created complications in knowing the extent of facility-level changes. For example, some medical centers had a performance goal focused on timely hiring of new staff. Some facilities may have already been in compliance or only needed minor adjustments to meet new requirements. Also, it is unclear whether workplace perceptions among other mental health providers, such as psychologists or social workers, were affected in a similar way.

Our study focused on a large health care system, but results may be less likely to generalize to smaller or solo practices of providers. Smaller practices may differ in unique and important ways and may be more negatively affected by reforms if they lack resources to adapt. We were unable to determine how quickly changes in satisfaction and work environment occurred because the AES was not fielded in 2005. It is possible that satisfaction dropped significantly in the year immediately after the comprehensive changes and that clinicians had adapted to changes by 2006. A second possibility is that high-level changes were customized to local work environments and these efforts (the process of customization) were reflected in higher satisfaction scores by 2006.

Further qualitative research is needed to better understand the influence of system changes in mental health care on clinician satisfaction. VHA has already identified several areas of future expansion, and understanding how these changes are implemented across settings would be valuable, considering our findings and study limitations. Interviews with mental health care leadership about their experiences with the change process as well as the perceived facilitators and barriers of implementation would help expand and clarify findings. Further, a number of other factors may be worth examining, including employee engagement (59,60) and job embeddedness (61), which have been known to influence turnover and other organizational outcomes.

Conclusions

VHA reorganization increased system accountability for providing mental health care, which was likely to influence providers through changes in work roles and reprioritization of job tasks. Contrary to expectations, we found that the reorganization did not adversely affect provider satisfaction ratings. VHA psychiatrists’ satisfaction and perceptions of the work environment were generally stable or increased slightly after the 2004 organizational transformation. Factors relating to management showed the strongest trend increase, and turnover rates and intention to leave were most stable after implementation. Although challenges and frustrations unquestionably were experienced during the organizational change, job satisfaction—a concern for organizations implementing changes—did not decrease. Further evaluation of the factors associated with effective change management in the mental health workforce is warranted.