Artificial Intelligence and the Doctor-Patient Relationship

The doctor-patient relationship describes the interaction of the patient with the medical provider in the professional setting with the intent of improving patient care. It is built on several principles including mutual trust and knowledge- the patient's faith in the doctor`s competence, the doctor's knowledge of the patient`s hopes, beliefs and trust in the patient`s accurate reporting of symptoms; physician empathy- understanding the patient`s experiences, concerns, outlook etc. and being able to express this to the patient compassionately; and loyalty- the patient to the physician and the doctor`s commitment to the patient never to abandon or disregard their wishes1Hoff T Collinson GE. How Do We Talk About the Physician-Patient Relationship? What the Nonempirical Literature Tells Us.Med Care Res Rev. 2017; 74 (Jun): 251-285https://doi.org/10.1177/1077558716646685Crossref Scopus (26) Google Scholar. The provider must act ethically supporting a professional demeanor, maintaining confidentiality, appropriately informing the patient of all diagnoses, providing optimal care and informed consent for medical or surgical therapies. The provider should be a good communicator listening to the patient`s concerns and answering all their questions. A favorable doctor-patient relationship has the ability to improve certain aspects of patient care. Symptom recovery has been shown to depend on the physician`s knowledge of the patient`s ailments and emotional state2Stewart MA McWhinney IR Buck CW. The doctor/patient relationship and its effect upon outcome.J R Coll Gen Pract. 1979; 29: 77-81PubMed Google Scholar. A patient who is satisfied with the physician is more likely to keep appointments and follow the suggestion of the provider regarding compliance with medicines and treatment regimens. Medical errors, patient complaints and negligence claims are also reduced3Chipidza FE Wallwork RS Stern TA. Impact of the Doctor-Patient Relationship.Prim Care Companion CNS Disord. 2015; 17 (Oct 2210.4088/PCC.15f01840)https://doi.org/10.4088/PCC.15f01840Crossref Scopus (132) Google Scholar. However, modern medicine is changing in ways which might adversely impact this relationship. As physician compensation has decreased, physicians often must see more patients to cover expenses etc. This will restrict the time spent with each individual patient in history taking, examination, explanation etc. Medicine is a business and remuneration is driven largely by procedures which are better compensated than just a simple follow up visit. These procedures, however, are often essential for accurate diagnosis. In cardiology, this means noninvasive testing which is usually performed by technicians without direct physician contact during the procedure. Diagnostic imaging performed in radiology is another common example as a physician only interprets the X rays, CT scans, MRIs etc. after the patient exits (all of which will likely be analyzed in the future by AI alone). Modern medicine appears to minimize direct physician-patient interaction in other ways. Telemedicine has been popularized over the last several years and particularly so during the COVID pandemic. Although generally demonstrating high levels of satisfaction from patient and provider, 4Haleem A Javaid M Singh RP Suman R. Telemedicine for healthcare: Capabilities, features, barriers, and applications.Sens Int. 2021; 2100117https://doi.org/10.1016/j.sintl.2021.100117Crossref Scopus (177) Google Scholar the absence of direct physical evaluation could lead to errors in diagnosis. Even with in person visits, we and others have observed that the physical exam has become less important5Hyman P. The Disappearance of the Primary Care Physical Examination-Losing Touch.JAMA Intern Med. 2020; 180 (Nov 1): 1417-1418https://doi.org/10.1001/jamainternmed.2020.3546Crossref PubMed Scopus (32) Google Scholar and among our colleagues, diagnosis is, as mentioned above, heavily dependent on non-invasive testing, imaging and laboratory studies with little emphasis in a majority of cases on direct examination. One might even consider that the routine use of automatic blood pressure measurements, not without benefit however, represents an early example of reducing patient/physician contact. In this evolving medical environment, what could be the effect of artificial intelligence (AI) on this relationship? With the popularization of AI tools such as Chat GPT, there has been much interest in its potential application in medicine. As physicians must obtain and interpret a variety of information, improvements in AI stand to help streamline and improve disease pattern recognition. While implementation of AI might assist in the objective aspects of care, little has been written about its potential impact on the doctor-patient relationship. In medicine, one incorporates data from multiple sources such as patient history, physical examination if performed, non-invasive and invasive imaging, blood testing etc. In a review of the existing literature, Kodera et al6Kodera S Akazawa H Morita H Komuro I. Prospects for cardiovascular medicine using artificial intelligence.J Cardiol. 2022; 79 (Mar): 319-325https://doi.org/10.1016/j.jjcc.2021.10.016Abstract Full Text Full Text PDF Scopus (5) Google Scholar discuss the ability for AI to improve clinical care in cardiology. For example, implementation of AI in ECG interpretation could prove to be helpful with identifying difficult rhythms while also making precise predictions for a reduction in left ventricle ejection fraction and amyloidosis. AI can also serve to improve interpretation of echocardiograms by providing a more accurate measurement of heart and valve function. AI has been implemented in advanced imaging where it has been used to automatically interpret fractional flow reserve from cardiac CT. AI may also facilitate myocardial diagnoses from MRI. There is nearly no area in medicine and care delivery that AI is not evaluating. Furthermore, AI-driven applications might even capture dictation of medical notes; with applications attempting to synthesize patient interviews and laboratory test results to write notes without clinician intervention7Beam AL Drazen JM Kohane IS Leong TY Manrai AK Rubin EJ. Artificial Intelligence in Medicine.N Engl J Med. 2023; 388 (Mar 30): 1220-1221https://doi.org/10.1056/nejme2206291Crossref Google Scholar. Given the interest in the field, more and more articles are being published and the New England Journal of Medicine has decided to launch a new journal in 2024 solely devoted to medical AI. How might the doctor-patient relationship evolve with AI-driven medical care? Medicine presents a complex challenge as patient care requires a melding of being analytical, non-analytical and compassionate in conveying the assessment and its significance to the patient. One cannot simply follow an algorithm and produce an answer. One must be able to interact and treat the patient and not just the disease process. With improvements in clinical care as eventually one hopes will develop with the increased utilization of AI, it is possible that more time might be available to enhance doctor-patient relationships. From freeing up time by assisting with transcription and documentation, added time might also reduce physician burnout and allow for a greater connection with the humanistic side of medicine. It can also serve to further democratize patient care by more rapidly providing patients with access to their information in a way a layperson could comprehend8Love R. Council post: The role of AI in Democratizing Healthcare: From diagnosis to access.Forbes. 2023; https://www.forbes.com/sites/forbesbusinesscouncil/2023/11/02/the-role-of-ai-in-democratizing-healthcare-from-diagnosis-to-access/?sh=1992bf1c1010Google Scholar. Thus, if AI was utilized solely as an adjunct to the provider for improving medical diagnoses and outcomes, the overall effect in the foreseeable future, seems positive. At the present time, however, general acceptance of an AI-based medical care requires the fine tuning of AI algorithms that must be rigorously evaluated prior to acceptance. More work is also necessary to assess any potential negative aspects of AI. Liability for adverse outcomes would be complicated as potential sources could stem from incorrect implementation or bias incorporated into underlying datasets. Additionally, the rapid access to information could prove to be a detriment. Patients may mistakenly interpret the potential for disease as a proven diagnosis which in turn would come with other psychological stresses9Dalton-Brown S. The Ethics of Medical AI and the Physician-Patient Relationship.Camb Q Healthc Ethics. 2020; 29 (Jan): 115-121https://doi.org/10.1017/s0963180119000847Crossref Google Scholar. We express here our viewpoint for the foreseeable future based on our concept of optimum medical practice with an AI assisted, practitioner-driven format. Future generations, however, may have a different opinion and welcome an impersonal AI android or humanoid robot (similar to the television character, Data on Star Trek: The Next Generation) as their super intelligent practitioner without the need for any human intervention. With what we perceive as a gradual separation of the patient from direct physician/provider contact, eventually patients may not appreciate what they are missing, that is the benefits of a positive doctor-patient relationship. Furthermore, in medically underserved areas, all knowing androids in the distant future might be a reasonable solution to broaden appropriate medical care to these populations. In conclusion, we contend that the potential advantages of AI in medicine, while evolving, can provide amazing benefits not presently available. When combined with an intact doctor-patient relationship, we believe the end result will be even more advantageous! Patients will always be best served with an empathetic practitoner who embodies all the positive attributes contained in this relationship. None