The influence of work stress and work support on burnout in public hospital nurses.

Lazarus and Folkman's (1984) transactional stress-strain-coping theory provides the theoretical background for the present thesis. This theory proposes that strain (i.e., burnout) occurs when demands (i.e., work stressors) exceed coping resources (e.g., social support). The current thesis explores the influence of social support on the stress-burnout relationship in nurses. A sample of Australian nurses working across three public hospitals in Brisbane's metropolitan district were recruited to investigate the nature and level of burnout experienced by nurses. Burnout is characterised by emotional exhaustion, depersonalisation and reduced personal accomplishment. The present research addresses gaps in the empirical literature by investigating the key work stressors experienced by Australian nurses and by establishing nurses' referent levels of work stress, social support, and burnout. In addition, the research explores the complex relationships between work stress, social support and burnout. The majority of nursing studies have failed to consider how support from within the nurses' work environment mitigates burnout. The present research builds upon previous nursing literature by examining the and effect hypotheses. Studies have consistently found support for the main effect model, however the hypothesis that social support buffers the negative effects of stress has resulted in highly conflicting findings. Some theorists (Cohen & Wills, 1985; Cutrona & Russell, 1990) propose that the buffering effects of social support will only be found if there is an adequate match between the needs elicited by the stressful event and the type of support an individual receives. The present study extends the stress or support matching theory by exploring the matching between specific types of stressors to specific types (i.e., emotional and instrumental) and sources (i.e.,supervisor and coworkers) of support. Cutrona (1990) suggests that the controllability of a stressor is the primary dimension in determining an appropriate match between stressors and types of support. Cutrona proposes that controllable stressful events elicit needs for instrumental support and uncontrollable events elicit needs for emotional support. Heeding Cutrona's advice, independent raters classified nurses' work stressors as controllable or uncontrollable stressful events prior to investigating the stressor-support matching theory. Three sequential studies were undertaken to explore the variables of interest to this research program. In Study 1, focus groups were conducted with 68 nurses (11 males, 34 females) from two public hospitals. The qualitative data was subjected to content analysis. The findings revealed that Australian nurses are exposed to a range of job-specific stressors (i.e., Job Conditions, Job Uncertainty, Interpersonal Conflict and a Lack of Professional Recognition and Support) and generic role stressors (i.e., Role Overload, Role Conflict and Role Ambiguity). The findings prompted the research to utilise Wolfgang's (1988) Health Professions Stress Inventory to measure nurses' job-specific stressors and Osipow and Spokane's (1987) Occupational Roles Questionnaire to measure nurses' role stressors in Study 2. The findings from Study 1 also confirmed that the way nurses perceive work support is consistent with current social support literature. Nurses indicated that their two main sources of support were their coworkers and their supervisor. Furthermore, nurses discussed social support from a multidimensional perspective, recognising different types of support that were broadly classified as emotional and instrumental support. Based on these findings, the researcher developed a work support measure specifically for the purpose of this research. Items were taken from established social support scales and were slightly modified to ensure that they were contextually relevant to nurses. In Study 2, 273 nurses (38 males, 235 females) completed a multi-measure questionnaire. While there was sufficient research evidence to indicate that the Occupational Roles Questionnaire (Osipow & Spokane, 1987) and the Maslach Burnout Inventory (Maslach, Jackson, & Leiter, 1996) possess adequate levels of reliability and validity, less was known about Wolfgang's Health Professions Stress Inventory and the work support scales designed for this research program. Factor analysis of the Health Professions Stress Inventory revealed a four-factor solution: Lack of Professional Recognition and Support, Patient Care Uncertainty, Job Conditions, and Interpersonal Conflict. Cronbach's coefficient alphas ranged from .62 to .83. Factor analysis of the Coworker Support Scale revealed a two-factor solution, representing emotional and instrumental support. Cronbach's coefficient alphas for the Emotional Coworker Support and Instrumental Coworker Support were .92 and .88 respectively. Contrary to expectations, factor analysis of the Supervisor Support Scale revealed a one-factor solution. It was therefore deemed appropriate to examine Supervisor Support as a unidimensional construct in subsequent analyses. Cronbach's coefficient alpha for the Supervisor Support scale was .96. Overall, the results from Study 2 provided supporting evidence to suggest that the measures used in the current research program were psychometrically sound. In Study 3, the data collected in Study 2 was subjected to further statistical analysis. The findings from Study 3 indicated that overall, the sample of Australian nurses reported low to moderate levels of work stress, moderate levels of work support and moderately high levels of burnout. For Emotional Exhaustion, predictor variables accounted for 42.2% of the total variance. Sociodemographic factors explained a small but significant proportion of the variance (2.7%). Work stressors however, were the main predictors of Emotional Exhaustion, explaining 41.5% of the total variance. Role Overload, Job Conditions and Role Conflict were the main determinants of Emotional Exhaustion, with Role Overload explaining most of the variance. For Depersonalisation, the predictor variables accounted for 34.2% of the total variance. Sociodemographic factors (11.5%) and work stressors (33.6%) both explained a significant proportion of the variance. Role Conflict and Patient Care Uncertainty were the main determinants of Depersonalisation, with Role Conflict explaining most of the variance. For Personal Accomplishment, Role Conflict and Role Ambiguity explained 20.5% of the total variance, with Role Conflict explaining most of the variance. Sociodemographic factors and job-specific stressors were not significant Predictors of Personal Accomplishment. Evidence for main effects of work support on burnout were limited. There was no evidence to suggest that work support had significant main effects on Emotional Exhaustion. Supervisor Support had a small, but significant main effect on Depersonalisation (β = -.15, p < .05) and Personal Accomplishment (β = -.24, p < .01). There was no evidence of main effects for Emotional and Instrumental Coworker Support. Furthermore, the present research found no significant evidence to support the buffering effect of work support on burnout. Theoretical and practical implications of these findings are discussed.