Artificial intelligence (AI) is a relatively new discipline, beginning in the 1950s but gaining much attention in the last decade.
Developed by the interpenetration of many disciplines such as mathematical logic, fuzzy mathematics, computer science, cybernetics, information theory, management theory, physiology, psychology, linguistics and philosophy, AI is an emerging discipline in which new ideas, concepts, theories and technologies are constantly emerging, as well as a frontier discipline in development (Jones et al., 2018). Machine learning (ML) is a specific classification of AI that studies and develops computational algorithms that transform empirical data into complex relationships or patterns and help humans make accurate decisions (Wang & Summers, 2012). In other words, AI enables a machine to simulate human thinking capability and behaviour, whereas ML allows a machine to automatically learn from past data to make more accurate decisions without being programmed.
Currently, the advancement of AI is in full swing, and it has permeated almost every corner of the human economy, life and work. Research in AI has continued to grow at a rapid rate from 2010 to 2021, with an average annual growth rate of 33% in the number of papers on the subject (Zhang et al., 2022). This shows that even after more than two years of the epidemic, there has been no decline in AI advancement.
There have been many public discussions and media predictions about what professions would be replaced by AI, and nurses and midwives are usually considered the most difficult ones. This is mainly because both nurses and midwives frequently interact with people both physically and emotionally, and constantly make judgements based on actual situations. AI relies on known knowledge and experience and often lacks flexibility. When faced with unexpected clinical conditions and unknown problems, AI cannot analyse and deal with them as quickly as humans. It may also lack empathy when dealing with humans. Therefore, in the field of nursing and midwifery (NaM), the development of AI emphasises ‘machine-human’ synergy, rather than being seen as a disruption or replacement. With the increasing development of AI, we are already seeing positive outcomes of such synergy, including virtual care and chatbots, robotics, and data analytics and clinical decision support.
Let’s take a closer look at these streams, their limitations, and potential evolvements in the future.
Virtual care and chatbots
Staffing shortage has been a prominent issue in NaM, especially since the COVID-19 pandemic. Virtual care leverages AI technology to create visualisation and auralisation of care services and enrich the interactions with patients. This type of technology also forms part of the virtually integrated care team consisting of nurses, midwives and other health professionals.
Chatbot like Molly is an example of such virtualised experiences. It helps patients schedule medical appointments, handle insurance inquiries, and collect patient data, without intervention from a real clinician. In NaM, AI-powered chatbots can handle time-consuming and repetitive responsibilities traditionally managed by nurses and midwives and provide virtualised clinical services such as giving clinical advice and triaging patients. Its scalability ensures the accessibility is around the clock and expands the coverage to remote locations.
Right now, chatbots are still considered a specialised clinical service, used exclusively in healthcare settings. Although tech giants are also investing in adapting these services into their assistant voice assistant services e.g. Apple Siri, Amazon Alexa, Google Assistant and Microsoft Cortana, their success will still rely on the collaboration between healthcare experts, technology developers, regulators and policymakers (Sezgin et al., 2020).
A common problem in the development of conversational assistants is that they are limited to keyword and word processing, single-sentence questions and answers, and often challenged by not understanding the context and emotional expression. Developing a super-emotional chatbot is not an easy task, but scientists and researchers in the field of Natural Language Processing are continuing to develop stronger algorithms for sentiment analysis interpretation of semantics (Zhang et al., 2022).
Robotic technology is another good effort to address issues such as the shortage of skilled nursing staff and the rising costs of medical and elderly care. In some hospitals and aged care facilities, logistic robots like Moxi, GoCart and Tug navigate autonomously and are designed to take care of deliveries of meals and other supplies, which free up caregivers’ time so that they can spend more time with care recipients. At the same time, these solutions have become more and more affordable and will reduce the overall operating costs of the facility.
Apart from the logistic robots that complete non-patient-facing tasks, there are also nursing robots such as Hobbit, Care-O-bot now vibrant in Homecare settings, designed to resemble caregivers and nurses and to deal with medical emergencies. The design aims to develop and promote client attachment, but this can lead to certain ethical issues. These robots are primarily concerned with happy and pleasurable emotional states, but human emotions are so variable that the robot may not be equipped with artificial empathy to attend to the patient’s frustration, anger and defiant behaviour. In contrast, patients, usually elderly or suffering from illness, may become emotionally dependent on the robot and any attempt to withdraw may lead to their distress and sadness. In return, the robot does not ‘feel’ anything for the patient. This again highlights the need for robots to work alongside human caregivers to maintain genuine social contact and empathy (Weidemann & Rußwinkel, 2021). A good use case of empathy in AI is to deploy AI-powered robots to support the care of people with dementia and help clinicians to make appropriate clinical decisions and refine treatments and interventions while maintaining a consistent empathetic approach, which in turn improves clinical outcomes.
The more empathetic social robots have been trialled in healthcare facilities, such as Pepper at Townsville Hospital in Queensland. Pepper acted as a concierge in the Emergency Unit, answering questions that were generally posed to nurses, such as where can family access a coffee, how is parking paid for, where are the exits, and so on (Griffith, 2019). If an individual asked Pepper where they can smoke a cigarette, not only would they be told that smoking was not allowed on the campus, but they would also be given information on how to quit (Griffith, 2019). Researchers at the Queensland University of Technology also trialled Pepper and considered social robots as an innovation that can relieve some of the pressures that health centres experience (Griffith, 2019).
The potential benefit of robotics is that it will speed up the digitalisation of operational management in care settings, once the robotic population reaches certain points. These robots constantly collect data related to service demands, delivery time, client feedback, and regularly scan the surrounding physical environment. If enough robots are deployed in a healthcare facility, these mobile data collection points will contribute a vast amount of data to help management in each NaM unit to improve the operational efficiency and quality of services (Griffith, 2019).
Data analytics and clinical decision support
The growth of healthcare institutions has brought rapid growth in healthcare data, proliferating the amount of professional information on NaM practice. Using a big data information integration platform and ML, nurses and midwives can access the complex and complicated data in real-time, and gain meaningful and valuable insights to support clinical management and decision making. A continuously evolving informatics system will help NaM managers to monitor staffing and allocation, coordinate care, streamline operations and processes, and deliver better care. This can in turn transform and further guide the construction and development of the efficient and effective NaM information management system.
Nurses and midwives using AI and EHR data have been able to read and translate signals into precise patient monitoring. For example, using ML to recognise patterns of co-occurring alarms (such as arrythmia alerts and hemodynamic monitoring), the SuperAlarm tool predicts when patients in critical care may need resuscitation. Not only does it improve care outcomes but it also reduces alert fatigue by consolidating multiple alerts into more meaningful information to support clinical decision making (Xiao et al., 2020).
As more AI applications are adopted in NaM fields of practice, the future of AI in data analytics may lead to building prototypes of NaM services based on big data. Effective and efficient prototypes will potentially be developed into a new model of care. The promising future also comes with shortcomings. The development of healthcare AI and the subsequent maintenance requires a significant investment in time and money. There is also the concern of energy consumption with AI applications, because incremental performance increases may come at the cost of huge amount of ML training, which corresponds to surging energy consumption. Healthcare AI product developers and practitioners should consider how AI’s ecological footprints may affect the environmental and economical sustainability. Moreover, although there are rigorous ethics reviews, existing AI algorithms have a track record of absorbing biases from training datasets related to ethnicity, gender, sexuality, age and religion (Rosales & Fernández-Ardèvol, 2019). AI practitioners including nurses and midwives should ensure that the benefits of AI are shared fairly among all people (Robert, 2019).
Overall, AI in nursing and midwifery is developing its unique outlook. To optimise the associated benefits, clinicians in the field of NaM should have a growth mindset to adapt to AI technology and be aware of its drawbacks such as lack of ability to contextualise, various ethical issues and potential threat to the environmental and financial sustainability of the world.
Nurses and midwives are best positioned to measure the impact of AI and how AI affects patient outcomes. Surprisingly, NaM are often just the users of AI technology, and the development lifecycle of AI technology in NaM is lacking in engagement with nurses and midwives (Ronquillo et al., 2021). NaM should be proactive in contributing to the AI system development as subject matter experts and patient advocates.</h4?robotics<>
Cedar Yin CHIA, June 2022
Griffith, C. (2019). Next time you go to hospital or the doctor, look out for a robot named Pepper helping out.
Jones, L. D., Golan, D., Hanna, S. A., & Ramachandran, M. (2018). Artificial intelligence, machine learning and the evolution of healthcare: A bright future or cause for concern? Bone & joint research, 7(3), 223-225. https://doi.org/10.1302/2046-3758.73.BJR-2017-0147.R1
Robert, N. (2019). How artificial intelligence is changing nursing. Nursing Management, 50(9), 30-39. https://doi.org/10.1097/01.Numa.0000578988.56622.21
Ronquillo, C. E., Peltonen, L.-M., Pruinelli, L., Chu, C. H., Bakken, S., Beduschi, A., Cato, K., Hardiker, N., Junger, A., Michalowski, M., Nyrup, R., Rahimi, S., Reed, D. N., Salakoski, T., Salanterä, S., Walton, N., Weber, P., Wiegand, T., & Topaz, M. (2021). Artificial intelligence in nursing: Priorities and opportunities from an international invitational think-tank of the Nursing and Artificial Intelligence Leadership Collaborative. Journal of advanced nursing, 77(9), 3707-3717. https://doi.org/https://doi.org/10.1111/jan.14855
Rosales, A., & Fernández-Ardèvol, M. (2019). Structural Ageism in Big Data Approaches. Nordicom Review, 40(s1), 51-64. https://doi.org/doi:10.2478/nor-2019-0013
Sezgin, E., Huang, Y., Ramtekkar, U., & Lin, S. (2020). Readiness for voice assistants to support healthcare delivery during a health crisis and pandemic. npj Digital Medicine, 3(1), 122. https://doi.org/10.1038/s41746-020-00332-0
Wang, S., & Summers, R. M. (2012). Machine learning and radiology. Med Image Anal, 16(5), 933-951. https://doi.org/10.1016/j.media.2012.02.005
Weidemann, A., & Rußwinkel, N. (2021). The Role of Frustration in Human–Robot Interaction – What Is Needed for a Successful Collaboration? [Original Research]. Frontiers in Psychology, 12. https://doi.org/10.3389/fpsyg.2021.640186
Xiao, R., Do, D., Ding, C., Meisel, K., Lee, R., & Hu, X. (2020). Generalizability of SuperAlarm via Cross-Institutional Performance Evaluation. IEEE access : practical innovations, open solutions, 8, 132404-132412. https://doi.org/10.1109/access.2020.3009667
Zhang, D., Maslej, N., Brynjolfsson, E., J., E., Lyons, T., Manyika, J., Ngo, H., Niebles, J. C., Sellitto, M., E., S., Shoham, Y., J., C., & Perrault, R. (2022). The AI Index 2022 Annual Report. https://aiindex.stanford.edu/wp-content/uploads/2022/03/2022-AI-Index-Report_Master.pdf