Adults' Perception of Empathy When Interacting with a Nursing Robot or a Physically Present Nurse: A Randomized Non-Inferiority Comparison

• Main Author: Crain, Dennis Raymond
• Other Authors: Shea, Kimberly D.
• Language: en_US
• Published: The University of Arizona. 2015
• Subjects: Non-inferiority; Nursing; Presence; Robotics; Technology; Empathy
• Online Access: http://hdl.handle.net/10150/593624

Background Nursing presence is an intersubjective connection between the nurse and patient that results in improved patient outcomes. Present day task-oriented healthcare robots possess an evolving capacity to address task-based attributes of nursing care but are far less capable of addressing attributes of nursing presence. The purpose of this study was to explore adults' perception of nurse-expressed empathy, an attribute of nursing presence, as enacted by a semi-autonomous robot nurse compared to a human nurse following a discussion of the adults' health concerns or issues. Methods The design for this study employed a non-inferiority randomized comparison of two groups. The overall hypothesis was that adults' perception of nurse-expressed empathy during human-robot interactions was not inferior to the perception of nurse-expressed empathy during human-human interactions. From a broad geographic community 102 adults, age 21 to 80, were recruited and assigned to an active control or reference treatment group using stratified and blocked randomization. In each group, participants discussed the impact of health issues or concerns on their daily life. Participants in the reference treatment group interacted with a semi-autonomous robot. Participants in the control group interacted with the researcher face-to-face. Participants' perception of nurse-expressed empathy was measured using the Empathic Understanding Scale of the Barrett-Lennard Relationship Inventory. A confidence interval approach using 95%-95% method was used to assess non-inferiority. The first confidence interval was obtained from analysis of seven historical studies that measured empathy using the Empathic Understanding Scale. The second confidence interval was obtained from analyses of the difference in mean perceived empathy between the two study groups. Results Three normalized statistical methods used to evaluate non-inferiority were significant (p<.025) and contained confidence intervals less than the non-inferiority margin (δ= 3.33). This resulted in the rejection of the null hypothesis that empathy communicated by a robot was inferior to empathy communicated by a human nurse. Conclusions This study provided evidence that nurses operating semi-autonomous robots can communicate empathy to adults. Innovation and collaboration among nurses, computer scientists and engineers will ensure that successive generations of robots maintain a nursing perspective while operating at their optimal capacity.