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 Table 1.2 – Distribution of physicians by region of the country, disaggregated by specialization and region, grouped as generalists or specialists Region Physicians Generalists Specialists Population Cardiology Cardiology/1,000 inhabs. Physician/1,000 inhabs. North 20,884 10,128 10,766 17,936,201 441 0.025 1.16 Northeast 80,623 34,461 46,162 57,254,159 2,534 0.044 1.41 Midwest 37,536 12,828 24,708 15,875,907 1,464 0.092 2.36 Southeast 244,304 91,124 153,180 86,949,714 8,383 0.096 2.81 South 68,430 20,948 47,482 29,644,948 2,694 0.091 2.31 Source: Scheffer M, Cassenote A, Guilloux AGA, Mioto BA, Mainardi GM. Medical Demography in Brazil 2018. São Paulo: FMUSP, CFM, Cremesp; 2018. 45 *Population estimated by IBGE in 2017. Data from the National Register of Health Establishments ( Cadastro Nacional de Estabelecimentos de Saúde, CNES), provided by the Ministry of Health, 46 show the same trend of concentration of medical professionals in the South and Southeast regions in February 2019, as seen in Figure 1.2. 46 1.7.3. eHealth Strategy The International Telecommunication Union (ITU), 47 an agency of the United Nations (UN), has been working in collaboration with the WHO to create a global environment for eHealth strategy implementation, especially in telemedicine. 47,48 The eHealth strategy is particularly important in the control of chronic noncommunicable diseases like hypertension, diabetes, heart diseases, and age-related diseases. The implementation of eHealth and telemedicine has progressed substantially in recent years, 49 but a recent systematic review on the cost effectiveness of eHealth implementation found shortage of studies and could not assess the impact of the strategy on health systems or social aspects, although most studies showed the strategy to be efficacious and cost effective. 49 1.7.4. Telecommunications and Data Infrastructure Up to 95% of the world’s population is estimated to have access to mobile telephony; in Brazil, this coverage may exceed 98%. Access to mobile phone services has progressed remarkably in Brazil, and the use of mobile phone equipment per inhabitant has increased from 2009 to 2019, 50,51 followed by a downward trend since then (Figure 1.3). Figure 1.4 shows the distribution of cell phones per 100 inhabitants and the ratio between cardiologists and cell phones per 1,000 inhabitants in Brazil in 2018. In terms of optical fiber coverage, the concentration is also greater in the Brazilian South and Southeast regions. Figure 1.5 shows the distribution of optical fiber backhaul in Brazilian municipalities. Backhaul is the portion of a hierarchical network (like cellular mobile communication networks) that is responsible for connecting the main network and the subnets. As shown in the map in figure 1.5, the concentration of optical fiber networks is lower in municipalities of the North region, which also concentrates the largest proportion of isolated municipalities. Data from figures 1.4 and 1.5 show a trend of concentration of cardiologists in areas with a higher concentration of enabled mobile devices. The correlation coefficient of this relationship is 0.67, which indicates that the availability of cardiologists correlates highly with access to mobile phones. These data indicate a greater challenge to the implementation of telemedicine in remote areas, considering that the shortage of physicians follows the same distribution of the deficient telecommunications infrastructure in Brazil. A detailed analysis of the costs and benefits of this expansion should direct incentives to this area. 1.8. Artificial Intelligence Artificial intelligence (AI) is a complex framework of sophisticated mathematical-computational models that allows the construction of algorithms to emulate various human tasks. AI encompasses an increasing number of subareas translating into different combined or complementary methodologies and approaches. Some examples include artificial neural networks (particularly deep learning models and convolutional networks), support vector machines, evolutionary algorithms, and natural language processing. The elaboration of analytical algorithms derived from large databases allows for interactive interpretation and apprehension, recognition of hidden patterns of combined information not obtained with traditional statistical methods, and assistance in more accurate decision making. The availability of this huge amount of data and new analytical techniques – big data analytics – opens up new scientific possibilities and AI applications, such as machine learning and data mining, which are already widely applied in telecardiology to diagnose combinations of multiple modalities of images, biobanks, electronic cohorts, on-site and distance clinical monitoring sensors, electronic health records, genomes and other molecular techniques, among others. 52-54 The implementation of these applications in clinical cardiology has grown exponentially 55 and has prognostic features, like the use of an algorithm derived from magnetic resonance based on three-dimensional patterns of right ventricular systolic function to assess with high accuracy the outcomes in pulmonary arterial hypertension, 56 identification of phenotypic patterns in heart failure with preserved ejection fraction and unfavorable prognosis confirmed by 1019