After IRB approval, data collection occurred in six NJ hospitals over two quarters. A total of 2086 patients, and 717 full and part-time RNs, from 45 adult inpatient units volunteered to complete the survey. Results using Structural Equation Modeling (SEM) to test unit level data obtained from the Press Ganey Patient Satisfaction Survey and the Employee Perspective Survey found no significant relationships between RN job satisfaction and patient satisfaction with nursing care. As predicted, a positive relationship was found between the subscale for direct supervisor and RN job satisfaction. The mediating effect of the direct supervisor on the relationship between RN job satisfaction and patient satisfaction was not supported.

SEM evaluates two models concurrently: a measurement model and theoretical model. It was selected to test the patient satisfaction-RN satisfaction model as it measures all variables simultaneously and examines the reliability and validity of the measurement tools. Differences found from previous studies may be due to types of instruments, data collection processes, timing, and/or statistical testing as this was the first study to use SEM to test this relationship.

A benefit of this research lies in the discovery of the “best practice” processes used to collect patient satisfaction data at the six hospitals. For instance, mailing patient satisfaction surveys within two days of discharge and to more than 86% of eligible discharge patients has been found to increase sample response. Further study using the Employee Perspective Survey and controlling the timing of patient data collection and synchronizing it with employee data collection should be conducted.