Advances in technology have traditionally been a means for positive change in healthcare. Developments of new devices for assessment, diagnosis, and treatment are commonly associated with improved patient outcomes (White & Cocchi, 2014). One such innovation in healthcare technology is the recent development of virtual nurse avatars generated by computer software. Duties presently delegated to this form of technology include such activities as: collecting patient information, issuing hospital discharge instructions, and coaching patients as they make lifestyle changes (Relational Agents Group, 2015). As applications for virtual nurse avatars expand, the responsibilities of human nurses will change. The arc of this evolution will be determined by the limits of technology, concepts of care, and the multi-faceted roles of the professional nurse.

In this article, we will share examples of avatar applications in nursing, discuss concerns about virtual nurse avatars, reinforce nursing as a caring profession, present avatars as caring agents, and consider the future of nursing avatar technology. We will conclude by noting that although virtual nurse avatars can acceptably perform some nursing tasks, they are limited in their ability to make complex judgments and engage in collaboration.

Avatars and Avatar Applications

Virtual nurse avatars, also referred to as 'relational agents,' are computer-generated tools that are used to provide care for patients. Virtual nurse avatars, also referred to as 'relational agents,' are computer-generated tools that are used to provide care for patients. As animated figures on a screen, the tools are capable of replicating nursing behaviors and performing various nursing activities, such as providing patients with information, assessing patient health status, and exhibiting behaviors associated with compassion, for example nodding in agreement (Relational Agents Group, 2015).

Literature suggests that nurse-avatar inventions can be successful as tools for health interventions. In treating patients with chronic pain and depression, Mccue et al. (2015) found that subjects who interacted with an avatar self-reported 100% compliance with the avatar’s suggestions to reduce stress and 89% compliance with the avatar’s healthy eating suggestions; 44% of the subjects indicated that they would prefer interaction with an avatar over interaction with a clinician. However, because compliance levels were self reported, critics may question whether accountability dynamics may affect patient preferences in care providers; it is easier to lie in computer-mediated environments (Ducklow & Mortenson, 2009).

In a study that tested the use of avatars as a means to reduce feelings of loneliness and isolation in older adults, Ring, Barry, Totske, and Bickmore (2013) reported mixed results. On a scale of satisfaction that ranged from one (low satisfaction) to seven (high satisfaction), the average subject rating for the avatar was 4.4. Study participant comments regarding the avatars indicated reactions ranging from contempt to affection towards these tools.

Zhou, Bickmore, Paasche-Orlow, and Jack (2014) reported that virtual nurse avatars can successfully perform discharge duties in hospitals, such as reviewing care plans and medications. In this study, the researchers used two different avatar models: one avatar appeared as a Caucasian female, the other avatar appeared as an African-American female. Researchers found a correlation between perceived patient similarities to the avatar and patient satisfaction. Moreover, among all patients who answered satisfaction questions, 36.3% indicated a preference for the avatar over professional staff, 24% indicated a preference for a doctor or a nurse, and 39.3% indicated no preference

When testing the capacity of a virtual nurse avatar to assist patients with low health literacy on hospital discharge, Bickmore, Pfeifer and Jack (2009), found that the avatars elicited positive reviews from patients for the quality time they offered. For example, one patient commented, “It’s more helpful than talking to a person; it’s just like a nurse, but she explained everything to the T.” Another said, "It was just like a nurse, only better, because sometimes a nurse just gives you a paper and says ‘here you go.'" And finally, one response concluded with “She cared about me, you know?” (Bickmore et al., 2009, pp. 6-7).

Concerns Regarding Virtual Nurse Avatars

Development and research regarding virtual nursing avatars invites speculation about the future roles of human nursing professionals. Research indicates that 47% of total employment in the United States is at risk of being replaced by technological innovations (Frey & Osbourne, 2013). Extreme models suggest that human nurses might one day be replaced fully by a combination of virtual nursing avatars and robots powered by computer technology. On more conservative models, however, virtual nurse avatars will evolve as a means to bridge the gap between healthcare supply and demand (Charova, Schafer, & Garron, 2011).

As an occupation, nursing has traditionally been viewed as a profession that requires interventions, judgments, and caring. While a nurse avatar can be programmed to generate clinical judgments and suggest or coordinate interventions, avatars are not typically understood as agents of care. The inevitable question is whether the duties of a caring profession require a human professional.

Nursing as a Caring Profession

The American Association of Colleges of Nursing (AACN) has identified caring as a fundamental component of nursing (2008). In one sense of the term, care can be understood as a physical action without a socio-emotional component. One can 'take care' of an object, while forming only a physical relationship with the object: for example, one might take care of an empty bottle by placing it in the recycling bin. In a nursing context, a particular socio-emotional relationship is not required to report vital signs of an unconscious patient; a particular skill-set is required. As such, the presence of a human is not required for the delivery of care. Such care could be equally well provided by a sophisticated robot, powered by computer technology.

In another sense, however, nursing requires a sense of care that encompasses more than physical actions towards objects. Selman (2011) has suggested that descriptions of the nursing profession should include a sense of care that encompasses compassion and elements of altruism. Similarly, the AACN (2008) indicated that altruism, integrity, compassion, and respect for human dignity characterize a caring nurse professional.

Terminology within the realm of ethics, including terms such as 'compassion' or 'respect' are exclusive to the domain of human morality. Inanimate objects, such as virtual nursing avatars, are not expected to have moral characteristics; hence, they are not traditionally viewed as experiencing compassion or respect. Therefore, virtual nurse avatars do not possess characteristics traditionally viewed as important in the nursing profession.

Avatars as Caring Agents

Although non-sentient beings, such as avatars, are incapable of showing the human emotions of compassion and caring, virtual nursing avatars can be successful at demonstrating behaviors that are interpreted as emotions. Although non-sentient beings, such as avatars, are incapable of showing the human emotions of compassion and caring, virtual nursing avatars can be successful at demonstrating behaviors that are interpreted as emotions. Technology development companies characterize virtual nursing avatars as being designed to appear empathetic by asking questions about a patient’s pain, feelings, or the day’s activities. Moreover, a virtual nurse avatar’s human appearance is a feature that prompts patients to see it as being human, as opposed to a being a software program (Drell, 2014).

The effectiveness of an empathetic design can be observed in the patient responses during the Bickmore et al. study (2009). Researchers programmed a virtual nurse avatar to address patients by name, engage in social conversation with elements of human tone, and use humor while sharing hospital discharge information. When asked, “How much do you feel that (avatar name) cares about you?” (p. 8), the mean patient rating in two trials was 5.4 and 5.8 on a seven-point scale, wherein the maximal score (seven) indicated that the patient perceived that the avatar cared very much.

It is possible that the conventional meanings of the term 'care' is changing to become inclusive of virtual nursing avatars; words do change meanings over time. Language historian, Anne Cruzan, suggests that value terms, such as awful and nice have had very different meanings in human history. For example, the term, awful, was historically a positive descriptor, meaning something worthy of awe. The term, nice, was once used to refer to things that were silly and foolish (Curzan, 2014). The 'care' terms could evolve to refer to something entirely non-human, and in its evolution, ascribe related nursing characteristics such as compassion, respect, and integrity to virtual nurse avatars.

Future of Nursing Avatar Technology

Applications for virtual nurse avatars are expanding. Relational Agents Group, a team of researchers from Northeastern University, is presently working with Boston Medical Center on a five-year project that uses virtual nurse avatars to assist with hospital discharge. The avatars make small talk, go over care plans, review medication protocols, and answer questions in a consistent and thorough manner. The goal of the project is to reduce re-admissions due to poor communication and better understand factors that lead to patient re-admission. It is supported by grants that include funding from the National Institutes of Health and the Blue Cross Blue Shield Foundation (Relational Agents Group, 2015).

A virtual-nurse, avatar-based service called Sense.ly has launched its own project with a tool named Molly. Molly has been delegated the tasks of facilitating patient check-ins with care providers. Also available as a smart-phone app, Molly has three-dimensional, animated features and a behavioral repertoire that includes, for example, subtle behaviors such as blinking and nodding. Sense.ly’s avatar-based, virtual nurse system notifies an off-site clinician if a patient’s check-in data indicates that health risks have increased. The customers listed on Sense.ly’s webpage include MetLife, Microsoft, and San Mateo Medical Center in San Mateo, California (Sense.ly, 2015).

Human nurses are uniquely situated to navigate the ambiguities of human ethical reasoning in hospital settings, and uniquely situated to provide the therapeutic benefits of human contact and touch. There are, however, limits to the roles computer-generated tools can fulfill in nursing care. As much as nursing is defined by its caring, the profession also requires an ability to make complex judgments and collaborate (AACN, 2008). Charova, Schaeffer, and Garron (2011) indicate human moral reasoning and judgment does not follow the logical patterns of reasoning that govern computer software. Human nurses are uniquely situated to navigate the ambiguities of human ethical reasoning in hospital settings, and uniquely situated to provide the therapeutic benefits of human contact and touch.

Elements of collaboration and teamwork are also limiting factors when it comes to applications for virtual nursing avatars. While it is true that computers can interact with each other and share directives, there is an integrative style that is unique to human teamwork and collaboration. Robert Miller is a computer scientist at the Massachusetts Institute of Technology where he specializes in crowdsourcing. In an interview, Miller illustrated the uniquely human asset of collaboration using the example of deciphering illegible handwriting. Absent a programmed algorithm, computer tools would not be able to translate the markings. A team of individual human beings, however, would be more successful, as these individuals can integrate their experiences and understandings of contexts to decipher the handwriting (Copeland, 2013). Additionally, a professional nurse or nursing team can quickly assess a patient’s limitations in terms of cognitive ability and make adjustments accordingly.

Conclusions

The literature indicates that there are tasks that virtual nurse avatars can successfully complete in a way that is acceptable to patients—avatars can collect patient information and successfully provide discharge instructions. Virtual nurse avatars, however, are limited in their abilities to make complex moral judgements and engage in collaboration. The ability to collaborate, using experience and education to understand the conditions and best practices for each patient, will continue to be a uniquely human asset in nursing.

Nursing shortages are projected to continue. Such shortages are associated with increased medical errors, decreased quality of care, increased waiting times, negative patient outcomes, and occupational stress for nurses (AACN, 2014). While virtual nurse avatars are unsuitable for many of the roles filled by professional nurses, avatars can, in collaboration with nursing professionals, take on tasks that will help bridge the gap between healthcare supply and demand.

Authors

Miriam Bowers Abbott, MA, BA
Email: mabbott@mccn.edu

Ms. Abbott currently serves as an Instructor of Ethics and English Composition in the Online RN to BSN program at Mount Carmel College of Nursing in Columbus, OH. Her educational background includes an MA and a BA in philosophy with an emphasis on moral reasoning and logic. She continues to develop her professional focus on the moral requirements of a caring profession.

Peggy Shaw, MSN, RN
Email: pshaw@mccn.edu

Ms. Shaw is an Assistant Professor of Nursing in the Online RN to BSN program at Mount Carmel College of Nursing. Ms. Shaw recently has received the Innovation in Nursing Education Award from the Ohio League of Nursing and the Mount Carmel Excellence in Nursing Award. In addition to her experience as a nurse educator, she has worked in out-patient oncology for ten year and hospice for three years.

Both authors share an interest in nursing ethics and nursing care. They have both developed online instructional materials that replace face-to-face classroom contact, and believe that as nursing education uses more technological mediums for communication, nursing practice might follow suit. They have been intrigued by recent innovations in virtual nurse avatar technology and the questions these innovations have raised for the future of nursing. One of their goals is to better understand how ethical concepts of ‘care’ are being described. They are also seeking to understand the extent to which virtual nursing technology can replace human nurses by looking for uniquely human assets in nursing as a caring profession. This article is a result of their review of research that seeks to provide answers to those questions.