5.4 Interface with patients

Patient attitudes, knowledge levels and responses to the use of artificial intelligence (AI) technologies.

5.4.1 Patient confidence in artificial intelligence (AI)-assisted clinical decisions

Patient attitudes, knowledge levels and responses to the use of AI technologies in their care can vary widely, from apprehension to enthusiasm.¹²⁷

Interviewees for this research perceived that most patients have little to no understanding of the implication of AI use in healthcare, although it should not be assumed that they are not interested and will not have opinions about it. Interested patients should be able to know how the technology they rely on for their health has been developed and what data it relies on. They should also be able to get an explanation from their clinician of how a particular decision has been made, even if the AI itself operates as a ‘black-box’. This suggests that different patients will require explanations at various levels of detail.  

Interviewees perceived that clinicians should be confident in being able to meet these expectations, including explaining to patients the scope of AI use in their care and the ‘checks and balances in place’. They also cautioned that if AI is introduced into clinical reasoning and decision making (CRDM) in an opaque way, or if clinicians are not confident in explaining its scope, limitations, benefits and risks, there is a real risk of undermining patient involvement in shared clinical decision making.

5.4.2 The changing relationship between clinicians and patients 

The patient-clinician relationship is at the heart of CRDM. It is a key tenet of modern medicine that the patient should be involved in any decision about their care.¹²⁸

As such, the involvement of AI in CRDM can present a ‘third-wheel’ effect,¹²⁹ in the sense that it is as important for the patient to have confidence in the AI as it is for the clinician, and that the AI may interrupt the patient-clinician relationship. For this reason, it is important that clinicians can take on the role of communicator and educator to their patients, and explain the role and limitations of AI being used in their care.

Interviewees for this research noted that the adoption of AI has the potential to reshape the relationships between clinicians and their patients by introducing further transparency and opportunity for collaboration in shared decision making.

Patients will increasingly have enhanced access to medical knowledge and subsequent decision making for their health, and will require support with assessing probabilities and risks.¹³⁰ Healthcare workers in certain specialties will need to move from being an ‘oracle’ of clinical information to a health ‘counsellor’, enabling high-quality, data-driven, shared clinical decision making.⁹

These new dynamics will dictate the required skills for managing the clinician-patient relationship. Clinicians will need to manage the interaction between patients and increasingly complex AI systems, including knowing the limits of AI and communicating this to patients.¹⁰ This will require an ability to guide the patient through uncertainty around potentially complex diagnostic decisions, and empower patients to take joint responsibility for their healthcare where appropriate.⁹

As healthcare workers interact with ever smarter machines, the demand for soft skills will rise. Social and emotional skills are already becoming more important as technologies take over more physical, repetitive and basic cognitive tasks.⁹

Interviewees concluded, in agreement with previous research, that focusing on the core human skills that AI and computers cannot achieve, such as collaboration, reflection, compassion and empathy will be essential. ^5,130

Certain specialities such as oncology have great experience of using these human skills to support CRDM, for example, due to the complexity and uncertainty of cancer care. Learning to translate these skills, through education, to other areas where AI will impact CRDM will be beneficial.

References

¹²⁷ Young AT, Amara D, Bhattacharya A, Wei ML. Patient and general public attitudes towards clinical artificial intelligence: a mixed methods systematic review. Lancet Digit Heal. 2021;3(9):e599-e611. doi:10.1016/S2589-7500(21)00132-1

¹²⁸ De Silva D. Helping people share decision making | The Health Foundation. The Health Foundation. https://www.health.org.uk/publications/helping-people-share-decision-making. Published 2012. Accessed March 7, 2022.

¹²⁹ Triberti S, Durosini I, Pravettoni G. A “Third Wheel” Effect in Health Decision Making Involving Artificial Entities: A Psychological Perspective. Front Public Heal. 2020;8(April):1-9. doi:10.3389/fpubh.2020.00117

¹³⁰ Building a Smarter Health Care Workforce Using AI. AHA Cent Heal Innov. 2019. https://www.aha.org/system/files/media/file/2019/09/Market_Insights_AI_Workforce_2.pdf. Accessed March 7, 2022.

⁹ Sinha S, Al Huraimel K. Transforming Healthcare with AI. In: Reimagining Businesses with AI; 2020:33-54. doi:10.1002/9781119709183.ch3

¹⁰ Liu X, Keane PA, Denniston AK. Time to regenerate: the doctor in the age of artificial intelligence. J R Soc Med. 2018;111(4):113-116. doi:10.1177/0141076818762648

⁵ Topol E. The Topol Review: Preparing the Healthcare Workforce to Deliver the Digital Future. 2019. https://topol.hee.nhs.uk/the-topol-review/. Accessed February 28, 2022.