Artificial intelligence in health care: Implications for nurse managers

Abstract

The systemic effect of digitalisation, technological developments, improved information processing infrastructures, and the use of data is rapidly transforming health care systems internationally. An increasing pace of development and adoption of artificial intelligence-based technologies have raised expectations, discussions, and concerns. These technologies offer great potential to support different stakeholders in the health care setting; however, all direct and indirect impacts of using these technologies are not always clear. Assessment frameworks have been introduced for generating evidence for decision-makers regarding, for example, health, economic, organisational, social, legal, and ethical implications of these technologies based on a systematic evaluation targeted at safety, efficacy, quality, appropriateness, cost-effectiveness, and efficiency aspects (World Health Assembly, 2014). Research evidence is crucial for assessing these technologies and their impact for appropriate adoption of artificial intelligence in nursing and health care. This special issue of the Journal of Nursing Management is dedicated to artificial intelligence in nursing and health care and it explores implications for nursing management. Artificial intelligence may be defined as human-designed software (potentially also including hardware) systems, which "act in the physical or digital dimension by perceiving their environment through data acquisition, interpreting the collected structured or unstructured data, reasoning on the knowledge, or processing the information, derived from this data and deciding the best action(s) to take to achieve the given goal" (AI HLEG, 2019, p. 6). Several themes emerged from examining the articles included in this special issue. The first subset of studies mainly considered what artificial intelligence-based technologies are relevant for nursing and explored the ethical issues related to artificial intelligence and its applications. Specifically, Chew and colleagues conducted a scoping review to identify artificial intelligence-based technologies that can improve nursing care. The results of 37 studies included in this review provide a useful overview that suggests that these technologies help with the following nursing tasks: documentation, formulating nursing diagnoses and care plans, patient monitoring, and patient care prediction. Further, Su and colleagues conducted a specific bibliometric analysis to understand the impact of artificial intelligence-based technologies on nursing management. The results provide a thorough overview of the studies published on the topic. Finally, Zhu and colleagues examined ethical issues related to a subset of artificial intelligence-based technologies used in elderly care via a scoping review. The results highlight several ethical issues when applying artificial intelligence-based technologies in health care and provide helpful guidance on the next steps needed to evaluate whether these technologies are applied ethically. The second subset of articles focused on understanding the role of artificial intelligence in assessing and improving the quality of care. Specifically, a meta-analysis by Chen and colleagues showed that artificial intelligence-based technologies can be used for improving health care workers' compliance with hand hygiene protocols in four ways, including automated training, electronic counting devices and remote monitoring, real-time hand hygiene reminders and feedback, and automated monitoring. The results of this meta-analysis show that these technologies effectively increase compliance and encourage nurse managers' adoption. A rapid review by Lobo and colleagues describes the literature on technology-based interventions for caregivers of patients suffering from a stroke. The study results show that technology can be used to educate and support caregivers, minimising uncertainty and ensuring a better quality of care for patients after stroke. Zhou and colleagues implemented an e-handover system and assessed its effect on paediatric nursing handover quality and efficiency. Study results show that an e-handover system improved nurse handover quality, optimised workflow, and increased work efficiency. An additional study by Zeyu and colleagues showed that an artificial intelligence-based video surveillance system used in a nursing home could help nurses to spend less time on patients' daily assessments. In addition, research by von Gerich and colleagues explored the opportunities of using artificial intelligence for semi-automated evaluation of provided care through specific indicators extracted from structured and unstructured information in electronic health records. They concluded that electronic health records may function as one source of information in tools that support the assessment of the quality of care. The third subset of articles regarded the use of artificial intelligence for risk prediction. For example, Flaks-Manov and colleagues explored the potential benefit of combining an automated 30-day readmission prediction model with clinicians' risk assessments. The authors concluded that this combination improves risk identification concerning readmission of elderly patients. Another example is a study by Ladios-Martin and colleagues, who developed a model for detecting individuals at risk of falls. The results inform the selection of variables to input in such a model, and their evaluation showed that a model including a fall prevention variable outperformed a model without it in detecting the risk of falls. The authors suggest that such models may support clinical practice when embedded into electronic health record systems. The fourth subset of studies focused on the experiences and beliefs of robotics in nursing. On the one hand, Ergin and colleagues surveyed nurse managers' opinions on artificial intelligence and robot nurses in Turkey, where two thirds of respondents were familiar with the concepts of artificial intelligence and robots. A similar number of respondents stated that robot nurses would benefit the nursing profession, while three-quarters voiced disbelief about robots replacing nurses. On the other hand, Teng and colleagues described how robots influenced nurses' time pressure and turnover intention and concluded that while robots in nursing may help reduce nurses' workload, robots also require nurses' resources to function properly. Finally, the use of robots was also studied in the context of nursing education, where Ropero-Padilla and colleagues explored the experiences and perceptions of last-year students on the acceptability and feasibility of using a chatbot for clinical decision-making and patient safety issues. Although the findings gave direction for improvement needs regarding the developed chatbot, they also supported the acceptability and feasibility of adopting a chatbot in clinical decision-making and patient safety education. The final subset of articles focused on nurses' experiences of artificial intelligence-based applications. Here, a study by Li and colleagues surveyed the association of a leader's innovation expectation with a nurse's innovation behaviour, considering a mediating effect of job control and creative self-efficacy. Based on the results, the leader's expectation was positively associated with the nurse's behaviour. Creative self-efficacy and job control mediated this relationship. The authors concluded that the leader's expectations enhance the nurse's self-efficacy and job control, contributing to enthusiasm for innovation. Finally, in a qualitative study by Laukka and colleagues, the future role of artificial intelligence in specialised care was described as experienced by nurse leaders and digital service developers. The findings were categorised under transformed work, care and services, and organisations. The respondents thought that the role artificial intelligence-based technologies in nursing would be significant in the future, likely reinforcing rather than replacing professionals or traditional care. Positive consequences for patients, professionals, and leaders were expected to be seen. Overall, the studies included in this special issue regard the relevance, effectiveness, impact, and experiences related to artificial intelligence in health care. Research findings show promising potential of the technologies studied in supporting nursing and evaluating and improving quality of care. As only one study included was a systematic review and meta-analysis, it seems a higher level of evidence of the effectiveness of these technologies in the clinical setting still seems to be scarce. It is also important to acknowledge that the adoption of artificial intelligence varies by setting and level of digital maturity (Flott et al., 2016). In the acute care setting, for example, this is seen through areas related to strategy, information technology capability, interoperability, governance and management, patient-centeredness, competence, and data analytics (Duncan et al., 2022). In addition to the characteristics and impact of the technology itself and the competences needed to use it, artificial intelligence implications need to be extended to the environment. The environmental impact of health care is significant. The health care sector has, for example, been estimated to cause 10% of annual greenhouse gas emissions (Eckelman & Sherman, 2016). Action to reduce the environmental impact of health care is urgently needed for environmental sustainability, which may be defined as "a condition of balance, resilience, and interconnectedness that allows human society to satisfy its needs while neither exceeding the capacity of its supporting ecosystems to continue to regenerate the services necessary to meet those needs nor by our actions diminishing biological diversity" (Morelli, 2011, p. 5). Although autonomous artificial intelligence has the potential to reduce the environmental burden of health care (Wolf et al., 2022), strong leadership is needed to guide technological developments, building on a value-based approach with practical relevance. Awareness and measures are needed to avoid a vicious circle on health care pollution and its health impacts, which in turn increase the need for more health care services that again cause further pollution (Sapuan et al., 2022). Safe and appropriate adoption of any technology in the health care setting requires capabilities and competence. Nurses are key actors in providing knowledge regarding the content and features of relevant technologies that need to be prioritised based on the needs of clinical practice. Successful and sustainable application of artificial intelligence in health care requires nursing leadership and management. The requirements regarding the level and depth of expertise related to artificial intelligence in nursing and health care vary depending on role, responsibility, and setting. In general, digital literacy has become a requirement for all nurses in everyday practice. Artificial intelligence-based technologies are taking an increasing role in this perspective. Structures are needed to ensure that all nurses understand the opportunities, benefits, challenges, and threats related to artificial intelligence in nursing and health care. This increases the expectations for the inclusion of topics related to informatics and artificial intelligence on all educational levels and in professional development in nursing, from bedside clinical competence to leadership and management. None.

Ähnliche Arbeiten