Data-based nurse staffing indicators with Bayesian networks explain nurse job satisfaction: a pilot study.

Aim. This paper is a report of a pilot study to examine the relationship of nursing intensity, work environment intensity and nursing resources to nurse job satisfaction. Background. There is an ever increasing amount of information in hospital information systems; however, still very little of it is actually used in nursing management and leadership. Methods. The combination of a retrospective time series and cross-sectional survey data was used. The time series patient data of 9704 in/outpatients and nurse data of 110 nurses were collected from six inpatient units in a medical clinic of a university hospital in Finland in 2006. A unit-level measure of nurse job satisfaction was collected with a survey (n = 98 nurses) in the autumn of 2006. Bayesian networks were applied to examine a model that explains nurse job satisfaction. Results. In a hospital data system, 18 usable nurse staffing indicators were identified. There were four nurse staffing indicators: patient acuity from nursing intensity subgroup, diagnosis-related group volume from work environment subgroup, and skill mix and nurse turnover from nursing resources subgroup that explained the likelihood of nurse job satisfaction in the final model. The Bayesian networks also revealed the elusive non-linear relationship between nurse job satisfaction and patient acuity. Conclusion. Survey-based information on nurse job satisfaction can be modelled with data-based nurse staffing indicators. Nurse researchers could use the Bayesian approach to obtain information about the effects of nurse staffing on nursing outcomes. [ABSTRACT FROM AUTHOR]