Untangling the relationship between HRM and hospital performance: the mediating role of attitudinal and behavioural HR outcomes †

AbstractThis study examines the complex linkages between employees' perceptions of HR systems and hospital performance. In particular, it explores the role of attitudinal and behavioural HR outcomes as potential mediating mechanisms of the HRM–performance link. Structural equation models using multi-informant (e.g. medical and non-medical staff, patients) and multi-source data from 167 acute hospital trusts in the English National Health Service reveal a positive relationship between employees' HR system perceptions and patient satisfaction. Moreover, results suggest that this relationship is mediated by employees' civility towards patients, that is, by the extent to which employees treat patients with courtesy, dignity and respect. This behavioural HR outcome, in turn, is negatively affected by employees' intention to leave, an essential attitudinal HR outcome that mediates the link between HR system perceptions and civility towards patients. This study, therefore, contributes to a better understanding of the mechanisms through which HRM can enhance patient satisfaction in health care organisations.Keywords:: employees' attitudes and behaviourshospitalsHR systemspatient satisfaction AcknowledgementsThe authors would like to thank special issue editors Timothy Bartram and Peter J. Dowling as well as the two anonymous reviewers for their helpful comments on earlier versions of this manuscript. All remaining errors and omissions are those of the authors. The authors are also indebted to the Department of Health, the NHS and the individual NHS acute trusts for their time and effort invested in supporting this research project. The authors would also like to acknowledge the excellent research assistance provided by Patrick Cichy and Sebastian Schäfer.Notes1. It should be noted that the NHS Staff Survey did not contain information about HR practices such as selection and pay, which have been commonly included in previous studies on HR systems. As a possible explanation, West et al. (Citation2006) argue that NHS hospitals reveal little variation in their pay and selection practices due to clear directives from the Department of Health.2. The partial model including HR system perceptions, civility towards patients and patient complaint rates is not displayed in this paper. However, these results can be made available by the authors upon request. Similar to the overall model depicted in Figure 1, in this model, the direct effect and hypothesised indirect effect were fitted simultaneously so as to estimate each effect while controlling for the other (Iacobucci et al. Citation2007).3. To shed further light on the underlying mechanisms of the HRM–performance chain in hospitals, we examined whether civility towards patients as a behavioural HR outcome mediates the relationship between attitudinal HR outcomes and patient complaint rates. As indicated above, no such relationship was found for employees' job efficacy that is neither significantly related to the mediator of civility towards patients nor to the dependent variable of patient complaint rates. In contrast, the relationship between employees' intention to leave and patient complaint rates is mediated by employees' civility towards patients. The significant path between intention to leave and patient complaint rates becomes insignificant when the mediator variable is included. A Sobel test supported the finding of mediation (z = 2.70, p < 0.01).4. Given that the Sobel test relies on strong distributional assumptions, we also followed the approach recommended by Preacher and Hayes (Citation2004) and computed bootstrap results for each of the hypotheses using 1000 replications. The results of this procedure were highly consistent with the conventional Sobel test results. As an additional robustness check, we performed regression analyses to test our hypotheses. Baron and Kenny's (Citation1986) procedure, which is the most prevalent mediation test in organisational and management research, was employed. According to Baron and Kenny (Citation1986, p. 1176), the variables (civility towards patients, job efficacy and intention to leave) in our proposed model can be considered as mediators when the following conditions are met: (a) variations in levels of the independent variable significantly account for variations in the presumed mediator, (b) variations in the mediator significantly account for variations in the dependent variable and (c) when the presumed mediator is included as a control variable, a previously significant relationship between the independent and dependent variables is no longer significant. Similar to the reported PLS results, all the above conditions are met for the mediators of civility towards patients and intention to leave; our analyses reveal, however, that the second condition is not met for job efficacy (the effect of job efficacy on civility towards patients is negative and not significant). As such, the Baron and Kenny (Citation1986) procedure provides additional support for our mediation Hypotheses 1 and 3. These results are available from the authors upon request.5. See Note 2.