ABC | Volume 113, Nº5, November 2019

Guidelines Guideline of the Brazilian Society of Cardiology on Telemedicine in Cardiology – 2019 Arq Bras Cardiol. 2019; 113(5):1006-1056 heterogeneous patterns of ventricular repolarization on electrocardiogram, 57 prediction of cardiovascular risk in large cohorts, 58 and prediction of urgent revascularization in emergency patients with chest pain, 59 among others. However, AI studies are generally based on observational data from administrative databases or clinical records, which potentially have different types of biases and confounding factors. 54 AI applications in telemedicine are promising but still very limited. 60 In the area of telediagnosis, efforts for automated classification and diagnosis in electrocardiography and cardiovascular imaging 61 are promising but still incipient. As for cardiovascular interventions, a recent review 62 found 8 studies incorporating machine learning in a real-life research setting, of which only three were evaluated in a randomized controlled trial. Of the 8 interventions, 6 (75%) showed statistical significance (at a p level of 0.05) in health outcomes. Some of these interventions are directly related to telecardiology and assessed interventions for weight loss, stress control, smoking cessation, and personalized nutrition based on glycemic response. Most studies had small sample sizes and short duration, reflecting a need for investments and further studies exploring the potentialities in the area. In a recent review, Topol 63 highlighted the presuppositions that will guide the future of AI in medicine: the patient must be considered the center for the implementation of any new technology, the incorporation of these new technologies for diagnosis and treatment should occur after robust validation of their clinical effectiveness, the use of digital tools and decision algorithms by patients should be an option for those who feel empowered to do so, and interdisciplinary training must involve health care professionals, engineers, computer scientists, and bioinformaticians. These minimum conditions Figure 1.4 – Distribution of cell phones and cardiologists, Brazil. a) Ratio cardiologists/1,000 inhabitants (2017), b) Density of cell phone density/100 inhabitants (2019). 51 Source: Scheffer M, Cassenote A, Guilloux AGA, Mioto BA, Mainardi GM. Medical Demographics in Brazil 2018. São Paulo: FMUSP, CFM, Cremesp; 2018. 45 Cell phones/100 inhab. Cardiologists/1000 inhab. Cardiologists/1000 inhab. Cellphones/100inhab.Cardiologists/1000 inhab. Cell phones/100 inhab. 1021