Resumo

LOBO, Luiz Carlos. Artificial Intelligence and Medicine. Rev. Bras. Educ. Med. [online]. 2017, vol.41, n.2, pp.185-193. ISSN 1981-5271.  https://doi.org/10.1590/1981-52712015v41n2esp.

While discussions develop regarding problems in the doctor-patient relationship and the deficiency of the clinical examination in medical practice, which leaves diagnoses more dependent of complementary tests, the importance of the computer in medicine and public health is highlighted. This is happening, either through the adoption of clinical decision support systems, the use of new technologies, such as wearable devices, or the storage and processing of large volumes of patient and population data. Data storage and processing capacity has increased exponentially over recent years, creating the concept of “big data”. Artificial Intelligence processes such data using algorithms that continually improve through intrinsic self-learning, thus proposing increasingly precise diagnostic hypotheses. Computerized clinical decision support systems, analyzing patient data, have achieved a high degree of accuracy in their diagnoses. IBM’s supercomputer, named “Watson”, has stored an extraordinary volume of health information, creating a neural network of data processing in several fields, such as oncology and genetics. Watson has assimilated dozens of medical textbooks, all the information from PubMed and Medline, and thousands of medical records from the Sloan Kettering Cancer Memorial Hospital. Its oncology network is now consulted by numerous specialists from all over the world. The English supercomputer Deep-Mind, by Google, has stored data from 1.6 million National Health Service patients, enabling the development of new clinical decision support systems, analysis of these patient data and generating alerts on their evolution in order to avoid contraindicated or conflicting medications, whilst also sending timely updates to the physicians about the health of their patients. Analyzing a set of dermatological images in a melanoma study, Deep-Mind showed a higher level of performance than that of specialists (76% versus 70.5%), with a specificity of 62% versus 59% and a sensitivity of 82%. Nevertheless, whereas the computer provides the know-what, it is the physician that will discuss the medical problem and the possible solutions with the patient, indicating the know-why of his or her case. This area requires continuous focus on the quality of medical training, emphasizing knowledge of the physiopathology of the organic processes and the development of the abilities to listen to, examine and advise a patient and, consequently, propose a diagnosis and treatment, accompanying his or her evolution.

Palavras-chave : Doctor-Patient Relationship; Clinical Examination; Decision Support Systems; Artificial Intelligence; Wearable Devices; Medical Education.

        · resumo em Português     · texto em Português     · Português ( pdf )