Patient-Centered Care and Health Information Technology

Bull World Health Organ 2020;98:245–250 | doi: http://dx.doi.org/10.2471/BLT.19.237198

Policy & practice

245

Introduction Empathy, compassion and trust are fundamental values of a patient-centred, relational model of health care. In recent years, the pursuit of greater efficiency in health care, including economic efficiency, has often resulted in these values being side-lined, making it difficult or even impossible for health-care professionals to incorporate them in practice. Artificial intel- ligence is increasingly being used in health care and promises greater efficiency, and effectiveness and a level of personalization not possible before. Artificial intelligence could help improve di- agnosis and treatment accuracy, streamline workflow processes, and speed up the operation of clinics and hospital departments. The hope is that by improving efficiency, time will be freed for health-care professionals to focus more fully on the human side of care, which involves fostering trust relationships and engag- ing with patients, with empathy and compassion. However, the transformative force of artificial intelligence has the potential to disrupt the relationship between health-care professionals and patients as it is currently understood, and challenge both the role and nature of empathy, compassion and trust in this context. In a time of increasing use of artificial intelligence in health care, it is important to re-evaluate whether and how these values could be incorporated and exercised, but most importantly, society needs to re-examine what kind of health care it ought to promote.

Empathy, compassion and trust Over the past decades, the rise of patient-centred care has shifted the culture of clinical medicine away from paternalism, in which the therapeutic relationship, the relationship between the health-care professional and the patient, is led by medical expertise, towards a more active engagement of patients in shared medical decision-making. This model of engagement requires the health-care professional to understand the pa- tient’s perspective and guide the patient in making the right decision; a decision which reflects the patient’s needs, desires

and ideals, and also promotes health-related values.1 The central point of the patient-centred model of doctor–patient relationship is that medical competency should not be reduced to technical expertise, but must include relational moral com- petency, particularly empathy, compassion and trust.2

Empathy, compassion and trust are broadly recognized as fundamental values of good health-care practice.3–5 Empathy allows health-care professionals to understand and share the patient’s feelings and perspective.6 Compassion is the desire to help, instigated by the empathetic engagement with the patient.7,8 Patients seek out and prefer to engage with health professionals who are competent, but also have the right inter- personal and emotional skills. The belief and confidence in the professional’s competency, understanding and desire to help is what underpins patient trust.9–13 Research has demonstrated the benefits of patient trust and empathetic care, including improved patient satisfaction, increased treatment adherence and improved health outcomes.14,15

Despite their importance, empathy and compassion in health care are often side-lined. In recent years, for example, socioeconomic factors, including an ageing population and austerity policies in Europe that followed the 2008 economic collapse, have led to the marginalization of these values.16 As health-care systems struggle with resourcing, the space for empathy and compassion has shrunk while the need for ef- ficiency has grown.17 In the United Kingdom of Great Britain and Northern Ireland, high-profile cases and reports, such as the Francis report, which followed the Mid Staffordshire scandal,18 the report by the Health Service Ombudsman entitled Dying without dignity,19 and the Leadership Alliance for the Care of Dying People report,20 all pointed at the lack of empathy as a major problem in clinical care. What these cases also showed was a conflicting relationship between the need for empathy and the pursuit of greater economic efficiency and of meeting operational targets. In 2017, Sir Robert Francis, who chaired the inquiry into the Mid Staffordshire scandal, mentioned in an interview that “at the time at Mid Staffordshire there was huge pressure on organizations to balance their books, to make pro-

Abstract Empathy, compassion and trust are fundamental values of a patient-centred, relational model of health care. In recent years, the quest for greater efficiency in health care, including economic efficiency, has often resulted in the side-lining of these values, making it difficult for health-care professionals to incorporate them in practice. Artificial intelligence is increasingly being used in health care. This technology promises greater efficiency and more free time for health-care professionals to focus on the human side of care, including fostering trust relationships and engaging with patients with empathy and compassion. This article considers the vision of efficient, empathetic and trustworthy health care put forward by the proponents of artificial intelligence. The paper suggests that artificial intelligence has the potential to fundamentally alter the way in which empathy, compassion and trust are currently regarded and practised in health care. Moving forward, it is important to re-evaluate whether and how these values could be incorporated and practised within a health-care system where artificial intelligence is increasingly used. Most importantly, society needs to re-examine what kind of health care it ought to promote.

a The Ethox Centre, Nuffield Department of Population Health, University of Oxford, Old Road Campus, Oxford, OX3 7LF, England. Correspondence to Angeliki Kerasidou (email: [email protected]). (Submitted: 11 June 2019 – Revised version received: 16 December 2019 – Accepted: 17 December 2019 – Published online: 27 January 2020 )

Artificial intelligence and the ongoing need for empathy, compassion and trust in healthcare Angeliki Kerasidoua

246 Bull World Health Organ 2020;98:245–250| doi: http://dx.doi.org/10.2471/BLT.19.237198

Policy & practice Empathy and artificial intelligence in health care Angeliki Kerasidou

ductivity improvements and matters of that nature. It all became about figures in the books, rather than outcomes for the patient. And I do believe there’s a danger of that happening again.”21 Research in 2017 in accident and emergency depart- ments in England on the effect of auster- ity policies on the everyday experiences of health-care professionals found that the pressure to meet targets negatively af- fected the doctors’ and nurses’ ability and opportunity to practise empathetic and holistic care,22 which led to moral distress and burnout among these professionals.23

Against this backdrop, artificial intelligence has been heralded as a way to save struggling national health-care systems24 and transform the future of health care by providing greater effi- ciency, effectiveness and high levels of personalized care.25

Artificial intelligence in health care

Artificial intelligence is broadly de- fined as “computing technologies that resemble processes associated with human intelligence, such as reasoning, learning and adaptation, sensory under- standing, and interaction.”26 The hope is that these technologies will transform health-care delivery by“ streamlining workflow processes […] improving the accuracy of diagnosis and personal- izing treatment, as well as helping staff work more efficiently and effectively.”25 Artificial intelligence could help health- care systems achieve greater efficiency, including economic efficiency, in two ways: (i) by improving time to and ac- curacy of diagnosis and treatment for patients, and where possible assisting with early prevention; and, (ii) by using health-care staff more efficiently.

A report published in 2018 in the United Kingdom suggested that the national health system could save up to 10% of its running costs by outsourc- ing repetitive and administrative tasks to artificial intelligence technologies.24 The same report also envisaged bedside robots performing social-care tasks such helping patients to eat, wash and dress, thus reducing the workload on care staff by 30%. But it is not only nursing and administrative tasks that artificial intelligence can help with. With regard to effectiveness, artificial intelligence systems could be used to deliver better clinical services both by

assisting with the diagnosis and manage- ment of patients, and by providing the diagnosis and prescribing treatments. Research conducted so far has shown that machines can perform as well as, or even better than, humans in detect- ing skin cancer,27 heart arrhythmia28 and Alzheimer disease.29 Furthermore, hu man – machi ne p ar t nerships c an provide far better results than either humans or machines alone.30 In these examples, the principal benefits of ar- tificial intelligence stem from its ability to improve efficiency and effectiveness by guiding diagnoses, delivering more accurate results and thus eliminating human error. With regard to greater efficiency through prevention, artificial intelligence technologies that track and analyse the movement of individuals could be used to detect people at risk of stroke and eliminate that risk through early intervention.31

Health care is already using tech- nology to improve its efficiency and effectiveness. From scalpels and syringes to stethoscopes and X-ray machines, the list of technologies used in medicine to facilitate and improve patient care is long. However, artificial intelligence differs from previous medical techno- logical advances. Whereas previous technologies were used to increase the senses and physical capacities of health- care professionals, consider, for example, how the stethoscope enhanced the hear- ing of doctors and X-rays their vision, the main role of artificial intelligence is to increase their reasoning and decision- making capacities. In this way, artificial intelligence is entering the health-care arena as another morally relevant actor that assists, guides or makes indepen- dent decisions regarding the treatment and management of patients.

Proponents of artificial intelligence technolog y in health care maintain that outsourcing tasks and decisions to rational machines will free up time for health-care professionals to engage in empathetic care and foster trust rela- tionships with patients.4,25,32,33 A review, outlining recommendations for National Health Service to be the world leader in using technology to benefit patients, notes that while artificial intelligence cannot deliver indispensable human skills, such as compassion and empathy, “the gift of time delivered by the intro- duction of these technologies […] will bring a new emphasis on the nurturing of the precious inter-human bond, based

on trust, clinical presence, empathy and communication.”25

The hope is that more free time for health-care professionals would not only lead to more trustworthy and empathetic care for patients, but also to less stress for and burnout of doc- tors and nurses.34 In addition, despite concerns that artificial intelligence will lead to job losses in health care, a report by the British Academy on the impact of artificial intelligence on work pointed out that professions that require the application of expertise and interaction with people will be less affected by auto- mation through artificial intelligence.35 According to these aforementioned publications, the introduction of artifi- cial intelligence technologies in health care offers the possibility of a win–win situation: patients benefit from more accurate diagnosis, better treatment outcomes, and increased empathy and compassion from medical staff, who in turn experience greater job satisfaction and less burnout.

The reimagination of health care, where artificial intelligence takes over specific, and even specialist, tasks while freeing time for health-care profession- als to communicate and empathize with patients, assumes that the value attached to empathy, compassion and trust will remain high. However, patients and the health-care system might value accuracy and efficiency more than empathy and judgement, which could shift the focus in medicine away from human-specific skills.36 In which direction health-care delivery will evolve is an important theoretical and practical question that requires examination. Currently, it is still unclear whether and how health- care practice will be transformed by artificial intelligence, and what effect it may have, particularly on the role of health-care professionals and on the therapeutic relationship.

Potential implications of artificial intelligence

Clinical competency is a fundamental aspect of the identity of health-care professionals and underpins the trust relationship between doctors and pa- tients. Patient trust is based on the belief that doctors and nurses have the right skills and expertise required to help the patient and also the right motivation to do so. This combination of clinical skill

247Bull World Health Organ 2020;98:245–250| doi: http://dx.doi.org/10.2471/BLT.19.237198

Policy & practice Empathy and artificial intelligence in health careAngeliki Kerasidou

with empathy and compassion is what justifies patients assuming a position of vulnerability towards the health- care professionals. Vulnerability is a fundamental characteristic of a trust relationship.37 The person placing trust in another knows and accepts that this trusted person can decisively influence the outcome of the entrusted action. Trust relationships involve a degree of uncertainty that cannot be mitigated; it is only the belief in the trusted person’s abilities and good will that justifies tak- ing on the risk of this uncertainty. In the clinical context, the patient knows that things can go wrong, but believes and hopes that this wrong would not be intentional, but rather because of bad luck or unforeseeable circumstances. Rules and regulations are put in place to protect patients from negligence and preventable mistakes. The constant quest to improve care highlights the fundamental moral obligations of non- maleficence and of acting in the best interests of patients. However, the fact remains that, in some cases, preventable harm could be the outcome of a medi- cal action.

The use of artificial intelligence to optimize accuracy of diagnosis and treatment could raise issues of account- ability when things go wrong, not only in cases where doctors follow the recom- mendations of artificial intelligence, but also when they decide to override these recommendations.38 In such situations, it is unclear who should be held account- able, whether responsibility should lie with the algorithm developer, the data provider, the health system that adopted the artificial intelligence tool, or the health-care professional who used it. In addition, even in situations where the role of artificial intelligence is assistive, health-care professionals might not feel confident to override its recom- mendation. If machines are brought into health care because they are better than humans at making certain rational decisions, how could humans rationally argue against them? Yet, the question of accountability is not the only issue raised here. The role and nature of trust in the therapeutic relationship is also at stake. Would and should patients still trust health-care professionals? If the introduction of artificial intelligence tools results in outsourcing clinical and technical skills to machines, would a belief in the good will of the doctor be enough to sustain a therapeutic trust

relationship as currently understood? One of the great promises of artificial intelligence is that by increasing effec- tiveness, accuracy and levels of person- alization in clinical care, it will succeed in replacing trust with certainty.39 In this case, patients might stop considering health-care professionals as experts in whose skills and knowledge they need to trust. This change might lead to a different relationship between health- care professionals and patients, one not characterized by vulnerability, but one of an assistive partnership.2 However, even in this more positive scenario, the transformation of society’s expecta- tions of care provision and the role of health-care professionals are unclear. It is important therefore to consider how the introduction of artificial intelligence will alter the public’s perception and understanding of trust in the clinical encounter as well as the way in which trust relationships will be formed in this context.

Similarly, artificial intelligence calls into question the role and value of empathy and compassion in health care. As mentioned earlier, in patient- centred care, empathy allows health-care professionals to understand the patients’ perspective, and thus helps health pro- fessionals tailor care to promote the patients’ values and address their indi- vidual needs. Empathy and compassion therefore play a very important role in an interpersonal model of care that rejects medical paternalism and brings the doc- tor and the patient together to discuss options and find appropriate solutions.40 To preserve this ideal of patient-centred care, ar tificial intelligence systems should be built in a way that allows for value-plurality, meaning the possibility that different patients might hold differ- ent values and have different priorities related to their care.41 In this way, the ethical ideal of shared decision-making can be maintained and not be replaced by another form of paternalism, one practised not by doctors, but by artificial intelligence algorithms.

Even if artificial intelligence tools are able to operate in a care context characterized by value-plurality, the role of empathy remains unclear. If what patient-centred care needs to survive in a future of artificial intelligence health care is machines programmed to incor- porate more than one value, what does this mean about the nature and role of empathy in care provision? Is empathy

still a professional value, or should it be now understood as another technology to be written into code and optimized? Indeed, research in the field of artificial intelligence suggests that it is possible to create empathetic machines42,43 as a way of relieving doctors and nurses from the substantial emotional work their profes- sions require.44 The likely effects of such complete optimization and operational- ization of health care are unclear. This optimization could improve health-care outcomes and personalized care; alterna- tively, it could lead to the reinstitution of a reductionist approach to medicine.45,46 Beyond these practical concerns, one should also consider whether something intangible, yet morally important will be lost if the therapeutic relationship is reduced to a set of functions performed by a machine, however intelligent. On the other hand, will our current understand- ing of empathy, compassion and trust change to fit the new context where some parts of care are provided by intelligent machines?

Conclusion The potential impact of artificial intel- ligence on health care, in general, and on the therapeutic relationship between health-care providers and patients, in particular, is widely acknowledged,38,47,48 as is the fact that society needs to learn how to deal “with new forms of agents, patients and environments.”49 Artifi- cial intelligence has great potential to improve efficiency and effectiveness in health care. However, whether artificial intelligence can support other values central to the delivery of a patient-cen- tred care, such as empathy, compassion and trust, requires careful examination. Moving forward, and as artificial intel- ligence is increasingly entering health care, it is important to consider whether these values should be incorporated and promoted within the new type of health care that is emerging and, if yes, how. More importantly, it is crucial to reflect on what kind of health care society should promote and how new technologies, including artificial intel- ligence, could help achieve it. ■

Acknowledgements AK is also affiliated with the Wellcome Centre for Ethics and Humanities, Uni- versity of Oxford.

Competing interests: None declared.

248 Bull World Health Organ 2020;98:245–250| doi: http://dx.doi.org/10.2471/BLT.19.237198

Policy & practice Empathy and artificial intelligence in health care Angeliki Kerasidou

摘要 医疗保健领域内的人工智能和对理解、同情和信任的持续需求 理解、同情和信任是以患者为中心的、关系型医疗保 健模式的基本价值。近年来，为了提高医疗保健的效 率，包括经济效率，往往导致对这些价值观的背离， 使医疗保健专业人员难以将其纳入实践。人工智能正 越来越多地被应用于医疗保健。这项技术为医疗保健 专业人员提供了更高的效率和更多的空闲时间，使他 们能够专注于人性化的护理，包括培养信任关系，理 解并同情患者。本文探讨了人工智能倡导者提出的高

效、理解、可信赖的医疗保健理念。本文表明，人工 智能有可能从根本上改变目前人们在医疗保健中看待 和实践理解、同情和信任的方式。展望未来，重新评 估这些价值观是否以及如何能够在越来越多地使用人 工智能的医疗保健系统中纳入和实施是一件十分重要 的事。最重要的是，社会需要重新审视什么类型的医 疗保健值得推广。

Résumé

L'intelligence artificielle et le besoin constant d’empathie, de compassion et de confiance dans le secteur de la santé L'empathie, la compassion et la confiance sont des valeurs fondamentales d'un modèle de soins de santé centré sur les relations avec le patient. Mais ces dernières années, la quête d'efficacité dans le secteur, y compris au niveau économique, a souvent relégué ces valeurs au second plan et les professionnels de la santé ont donc eu du mal à les intégrer à leur pratique. De son côté, l'intelligence artificielle gagne en importance. Cette technologie devrait accroître l'efficacité tout en libérant du temps pour les professionnels de la santé, qui pourront ainsi se concentrer sur l'aspect humain des soins, notamment en établissant une relation de confiance et en faisant preuve d'empathie et de compassion envers les patients. Le présent article s'intéresse à l'idée d'un système de

soins de santé efficace, qui repose sur l'empathie et la confiance, et à laquelle adhèrent les adeptes de l'intelligence artificielle. Il suggère que l'intelligence artificielle a le potentiel nécessaire pour transformer radicalement la manière dont l'empathie, la compassion et la confiance sont considérées et appliquées aujourd'hui dans le secteur de la santé. À l'avenir, il est essentiel de réexaminer l'importance de ces valeurs et la façon dont elles pourraient être incorporées et mises en œuvre dans un système de santé où l'intelligence artificielle devient peu à peu incontournable. Et surtout, la société a besoin de se demander quel modèle de soins de santé elle souhaite promouvoir.

Резюме

Искусственный интеллект и постоянная потребность в эмпатии, сочувствии и доверии в сфере здравоохранения Эмпатия, сочувствие и доверие — это основополагающие ценности ориентированной на пациента реляционной модели здравоохранения. В последнее время стремление повысить эффективность систем здравоохранения, в том числе их рентабельность, приводит к тому, что этим ценностям часто не уделяется должного внимания, что в свою очередь значительно осложняет их использование на практике работниками сферы здравоохранения. Применение искусственного интеллекта в сфере здравоохранения неуклонно растет. Эта технология привлекательна перспективой повышенной эффективности и тем, что она оставляет медицинским работникам больше свободного времени для непосредственной работы с пациентами, в

том числе для налаживания доверительных отношений и применения эмпатии и сочувствия в профессиональном общении с пациентами. В этой с татье рассматривается представление об эффективной системе здравоохранения, построенной на основе эмпатии и доверия, которое предлагается специалистами, продвигающими внедрение технологий ИИ в сфере здравоохранения. В статье выдвигается предположение о том, что искусственный интеллект потенциально способен коренным образом изменить сегодняшнее представление о применении эмпатии, сочувс твия и доверия в сфере здравоохранения и внедрении соответствующих практик. В дальнейшем важно заново оценить возможность включения этих

ملخص الذكاء االصطناعي واحلاجة املستمرة للتعاطف والشفقة والثقة يف الرعاية الصحية

الرعاية لنموذج األساسية القيم هي والثقة والشفقة التعاطف السعي أدى األخرية، السنوات يف املريض. عىل املرتكزة الصحية لتحقيق املزيد من الفعالية يف الرعاية الصحية، بام يف ذلك الفعالية االقتصادية، يف الغالب إىل تباعد هذه القيم، مما جعل من الصعب الذكاء ُيستخدم املامرسة. يف دجمها الصحية الرعاية أخصائيي عىل هذه وتقدم الصحية. الرعاية يف متزايد بشكل االصطناعي ألخصائيي احلر والوقت الفعالية من بمزيد وعودًا التكنولوجيا يف بام الرعاية، من اإلنساين اجلانب عىل للرتكيز الصحية الرعاية ذلك تعزيز عالقات الثقة واالندماج مع املرىض من خالل التعاطف

والشفقة. يناقش هذا املقال رؤية تتميز بالفعالية والتعاطف لرعاية تشري االصطناعي. الذكاء مؤيدي يطرحها بالثقة، جديرة صحية التي الطريقة تغيري إمكانية لديه االصطناعي الذكاء أن إىل الورقة منها، كل ممارسة وكيفية والثقة، والشفقة التعاطف إىل هبا ينظر بشكل جذري يف جمال الرعاية الصحية. ومع امليض قدما، من اهلام إعادة تقييم ما إذا كان يمكن دمج وممارسة هذه القيم، داخل نظام الرعاية الصحية، حيث يستخدم الذكاء االصطناعي بشكل متزايد، وكيفية القيام بذلك. واألهم من ذلك، حيتاج املجتمع إىل التحقق

من نوع الرعاية الصحية الذي يمكن هلذه القيم أن ترتقي به.

249Bull World Health Organ 2020;98:245–250| doi: http://dx.doi.org/10.2471/BLT.19.237198

Policy & practice Empathy and artificial intelligence in health careAngeliki Kerasidou

ценностей в систему здравоохранения, все чаще использующую технологию искусственного интеллекта, и их применения на практике. Что наиболее важно, общество нуждается в пересмотре

того, развитие какого типа системы здравоохранения следует поощрять.

Resumen

La inteligencia artificial y la continua necesidad de empatía, compasión y confianza en la atención sanitaria La empatía, la compasión y la confianza son valores fundamentales de un modelo relacional de atención sanitaria centrado en el paciente. En los últimos años, la búsqueda de una mayor eficiencia en la atención sanitaria, incluida la eficiencia económica, ha dado lugar con frecuencia a que estos valores se vean relegados a un segundo plano, lo que dificulta que los profesionales sanitarios los incorporen en la práctica. La inteligencia artificial se utiliza cada vez más en la atención sanitaria. Esta tecnología promete una mayor eficiencia y más tiempo libre para que los profesionales sanitarios se centren en el lado humano de la atención, lo que incluye el fomento de las relaciones de confianza y el trato a los pacientes con empatía y compasión. En este artículo

se examina la visión de una atención sanitaria eficiente, empática y confiable que proponen los defensores de la inteligencia artificial. El artículo sugiere que la inteligencia artificial tiene el potencial de alterar fundamentalmente la forma en que la empatía, la compasión y la confianza se consideran y practican actualmente en la atención sanitaria. Para avanzar, es importante volver a evaluar si dichos valores se podrían incorporar y practicar en un sistema de atención sanitaria en el que se utiliza cada vez más la inteligencia artificial, y de qué manera. Lo más importante es que la sociedad debe reconsiderar qué tipo de atención sanitaria debe promover.

References 1. Emanuel EJ, Emanuel LL. Four models of the physician–patient relationship.

JAMA. 1992 Apr 22–29;267(16):2221–6. doi: http://dx.doi.org/10.1001/ jama.1992.03480160079038 PMID: 1556799

2. Bauchat JR, Seropian M, Jeffries PR. Communication and empathy in the patient-centered care model: why simulation-based training is not optional. Clin Simul Nurs. 2016;12(8):356–9. doi: http://dx.doi.org/10.1016/j. ecns.2016.04.003

3. Serrant L. Compassion NHS: evidencing the impact. London: NHS England; 2016. Available from: https://www.england.nhs.uk/wp-content/ uploads/2016/05/cip-yr-3.pdf [cited 2019 Apr 17].

4. Tweedie J, Hordern J, Dacre J. Advancing medical professionalism. London: Royal College of Physicians; 2018.

5. Spiro H. Commentary: the practice of empathy. Acad Med. 2009 Sep;84(9):1177–9. doi: http://dx.doi.org/10.1097/ACM.0b013e3181b18934 PMID: 19707052

6. Halpern J. From detached concern to empathy: humanizing medical practice. New York: Oxford University Press; 2001. doi: http://dx.doi. org/10.1093/acprof:osobl/9780195111194.001.0001

7. Nussbaum M. Compassion: the basic social emotion. Soc Philos Policy. 1996;13(01):27–58. doi: http://dx.doi.org/10.1017/S0265052500001515

8. Goetz JL, Keltner D, Simon-Thomas E. Compassion: an evolutionary analysis and empirical review. Psychol Bull. 2010 May;136(3):351–74. doi: http:// dx.doi.org/10.1037/a0018807 PMID: 20438142

9. Charon R. The patient–physician relationship. Narrative medicine: a model for empathy, reflection, profession, and trust. JAMA. 2001 Oct 17;286(15):1897–902. doi: http://dx.doi.org/10.1001/jama.286.15.1897 PMID: 11597295

10. Chin JJ. Doctor–patient relationship: a covenant of trust. Singapore Med J. 2001 Dec;42(12):579–81. PMID: 11989581

11. O’Neill O. Autonomy and trust in bioethics. Cambridge: Cambridge University Press; 2002. doi: http://dx.doi.org/10.1017/CBO9780511606250

12. Mechanic D. Changing medical organization and the erosion of trust. Milbank Q. 1996;74(2):171–89. doi: http://dx.doi.org/10.2307/3350245 PMID: 8632733

13. Halpern J. What is clinical empathy? J Gen Intern Med. 2003 Aug;18(8):670– 4. doi: http://dx.doi.org/10.1046/j.1525-1497.2003.21017.x PMID: 12911651

14. Kelley JM, Kraft-Todd G, Schapira L, Kossowsky J, Riess H. The influence of the patient–clinician relationship on healthcare outcomes: a systematic review and meta-analysis of randomized controlled trials. PLoS One. 2014 Apr 9;9(4):e94207. doi: http://dx.doi.org/10.1371/journal.pone.0094207 PMID: 24718585

15. Joffe S, Manocchia M, Weeks JC, Cleary PD. What do patients value in their hospital care? An empirical perspective on autonomy centred bioethics. J Med Ethics. 2003 Apr;29(2):103–8. doi: http://dx.doi.org/10.1136/ jme.29.2.103 PMID: 12672891

16. Kerasidou A, Kingori P, Legido-Quigley H. “You have to keep fighting”: maintaining healthcare services and professionalism on the frontline of austerity in Greece. Int J Equity Health. 2016 Jul 26;15(1):118. doi: http:// dx.doi.org/10.1186/s12939-016-0407-8 PMID: 27457098

17. Kerasidou A, Horn R. Making space for empathy: supporting doctors in the emotional labour of clinical care. BMC Med Ethics. 2016 Jan 27;17(1):8. doi: http://dx.doi.org/10.1186/s12910-016-0091-7 PMID: 26818248

18. Francis R. Report of the Mid Staffordshire NHS Foundation Trust Public Inquiry. London: The Stationary Office; 2013.

19. Dying without dignity: investigations by the Parliamentary and Health Service Ombudsman into complaints about end of life care. London: Parliamentary and Health Service Ombudsman; 2015.

20. One chance to get it right. London: Leadership Alliance for the Care of Dying People; 2014. Available from: https://wales.pallcare.info/files/ One_chance_to_get_it_right.pdf [cited 2019 Apr 27].

21. Matthews-King A. Government says NHS hospitals can wring out another £300m in efficiency savings. Independent. 2017 Nov 8.

22. Kerasidou A. Empathy and efficiency in healthcare at times of austerity. Health Care Anal. 2019 Sep;27(3):171–84. doi: http://dx.doi.org/10.1007/ s10728-019-00373-x PMID: 31152291

23. Kerasidou A, Kingori P. Austerity measures and the transforming role of A&E professionals in a weakening welfare system. PLoS One. 2019 Feb 13;14(2):e0212314. doi: http://dx.doi.org/10.1371/journal.pone.0212314 PMID: 30759147

24. Darzi L. The Lord Darzi review of health and care. London: Institute for Public Policy Research, 2018.

25. Topol E. The Topol review: preparing the healthcare workforce to deliver the digital future. London: National Health Service; 2019.

26. Artificial intelligence (AI) in healthcare and research. London: Nuffield Council on Bioethics; 2018.

27. Esteva A, Kuprel B, Novoa RA, Ko J, Swetter SM, Blau HM, et al. Dermatologist-level classification of skin cancer with deep neural networks. Nature. 2017 Feb 2;542(7639):115–8. doi: http://dx.doi.org/10.1038/ nature21056 PMID: 28117445

28. Hannun AY, Rajpurkar P, Haghpanahi M, Tison GH, Bourn C, Turakhia MP, et al. Cardiologist-level arrhythmia detection and classification in ambulatory electrocardiograms using a deep neural network. Nat Med. 2019 Jan;25(1):65–9. doi: http://dx.doi.org/10.1038/s41591-018-0268-3 PMID: 30617320

29. Fraser KC, Meltzer JA, Rudzicz F. Linguistic features identify Alzheimer’s disease in narrative speech. J Alzheimers Dis. 2015;49(2):407–22. doi: http:// dx.doi.org/10.3233/JAD-150520 PMID: 26484921

30. Patel BN, Rosenberg L, Willcox G, Baltaxe D, Lyons M, Irvin J, et al. Human– machine partnership with artificial intelligence for chest radiograph diagnosis. NPJ Digit Med. 2019 Nov 18;2(1):111. doi: http://dx.doi. org/10.1038/s41746-019-0189-7 PMID: 31754637

250 Bull World Health Organ 2020;98:245–250| doi: http://dx.doi.org/10.2471/BLT.19.237198

Policy & practice Empathy and artificial intelligence in health care Angeliki Kerasidou

31. Villar JR, González S, Sedano J, Chira C, Trejo-Gabriel-Galan JM. Improving human activity recognition and its application in early stroke diagnosis. Int J Neural Syst. 2015 Jun;25(4):1450036. doi: http://dx.doi.org/10.1142/ S0129065714500361 PMID: 25684369

32. Davenport TH, Glover W, Artificial J. Intelligence and the augmentation of health care decision-making. Massachusetts: NEJM Catalyst; 2018.

33. Topol E. Deep medicine: how artificial intelligence can make healthcare human again. New York: Basic Books; 2019.

34. Nundy S, Hodgkins ML. The application of AI to augment physicians and reduce burnout [internet]. Health Affairs Blog. 2018 Sep 18. Available from: https://www.healthaffairs.org/do/10.1377/hblog20180914.711688/full/ [cited 2019 Apr 27].

35. The impact of artificial intelligence on work: an evidence synthesis on implications for individuals, communities, and societies. London: British Academy and Royal Society; 2018.

36. Susskind R, Susskind D. The future of the professions: how technology will transform the work of human experts. Oxford: Oxford University Press; 2015.

37. Wright S. Trust and trustworthiness. Philosophia. 2010;38(3):615–27. doi: http://dx.doi.org/10.1007/s11406-009-9218-0

38. Ross J, Webb C, Rahman F. Artificial intelligence in healthcare. London: Academy of Medical Royal Colleges; 2019.

39. Zuboff S. The age of surveillance capitalism: the fight for a human future at the new frontier of power. London: Profile Books; 2019.

40. Ahmad N, Ellins J, Krelle H, Lawrie M. Person-centred care: from ideas to action. London: The Health Foundation; 2014.

41. McDougall RJ. Computer knows best? The need for value-flexibility in medical AI. J Med Ethics. 2019 03;45(3):156–60. doi: http://dx.doi. org/10.1136/medethics-2018-105118 PMID: 30467198

42. el Kaliouby R. We need computers with empathy. Cambridge: MIT Technology Review; 2017.

43. Johnson K. Google empathy lab founder: AI will upend storytelling and human–machine interaction: Venturebeat. 2018 Mar 11. Available from: https://venturebeat.com/2018/03/11/google-empathy-lab-founder-ai-will- upend-storytelling-and-human-machine-interaction/ [cited 2019 Jan 20].

44. Srivastava K, Das RC. Empathy: process of adaptation and change, is it trainable? Ind Psychiatry J. 2016 Jan-Jun;25(1):1–3. doi: http://dx.doi. org/10.4103/0972-6748.196055 PMID: 28163400

45. Goldhahn J, Rampton V, Spinas GA. Could artificial intelligence make doctors obsolete? BMJ. 2018 Nov 7;363:k4563. doi: http://dx.doi. org/10.1136/bmj.k4563 PMID: 30404897

46. Beresford MJ. Medical reductionism: lessons from the great philosophers. QJM. 2010 Sep;103(9):721–4. doi: http://dx.doi.org/10.1093/qjmed/hcq057 PMID: 20395240

47. Fenech M, Strukelj N, Buston O. Ethical, social and political challenges of artificial intelligence in health. London: Wellcome Trust Future Advocacy; 2018.

48. Loh E. Medicine and the rise of the robots: a qualitative review of recent advances of artificial intelligence in health. BMJ Leader. 2018;2(2):59–63. doi: http://dx.doi.org/10.1136/leader-2018-000071

49. Floridi L, Cowls J, Beltrametti M, Chatila R, Chazerand P, Dignum V, et al. AI4people – an ethical framework for a good AI society: opportunities, risks, principles, and recommendations. Minds Mach (Dordr). 2018;28(4):689– 707. doi: http://dx.doi.org/10.1007/s11023-018-9482-5 PMID: 30930541

Copyright of Bulletin of the World Health Organization is the property of World Health Organization and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.