Two recent major events characterize the object of study on this axis: the disaster of Vale dam in Brumadinho and the COVID-19 pandemic. For various reasons, acute events of great intensity cause enormous stress on people and health systems. There is an immediate need to plan actions to assist victims and mitigate risks. Too often, disasters and epidemics cannot be planned and are unprecedented in nature, meaning that data that can help with planning to deal with them is either nonexistent or not very robust.
Using the lessons learned from previous disasters and epidemics can be useful to guide health actions to deal with new diseases, such as COVID-19. Health planning actions include fast, adequate and equitable care of health systems to significantly increase the demand of sick individuals, understand the distribution of the disease and definition of risk groups (who may benefit most from vaccines and new drugs when available). In the context of COVID-19, an important use of AI has been to observe and predict the evolution of the pandemic, dealing with an enormous and unprecedented amount of data derived from public health surveillance, real-time monitoring of epidemic outbreaks, projection of trends, information and regular updates on the situation of government institutions and bodies and information on the use of health services.99 In a recent study, we assessed the impact of non-pharmaceutical interventions on virus transmission and spread of SARS-CoV-2 in Brazil. The genomic sequencing of 427 new genomes and the analysis of a set of genomic data geographically representative of 21 of the 27 Brazilian states identified more than 100 international introductions of SARS-CoV-2 in Brazil. National air travel during this period coincided with the spread of SARS-CoV-2 from large urban centers to the rest of the country. Our data suggest that current interventions remain insufficient to keep the transmission of the virus under control in Brazil.100 The experience with COVID-19 has been instrumental in understanding how to use AI to predict the behavior of the epidemic in an attempt to better inform health managers to organize services and define non-pharmaceutical interventions, such as social isolation. A huge challenge is to place large social differences in the prediction equation, especially in large cities, and within a territory that is larger than the whole of Europe.
Unlike COVID-19, there are epidemics that are repeated frequently, following standards that are only partly expected and predictable. Dengue, for example, happens only in times of rain and heat, since it requires the growth of the vector mosquito Aedes aegypti, and follows a cycle every 2 or 4 years, depending on the city or country under study. Applying a Bayesian inference model in geo-localized cases of dengue, we showed that the transmission clusters expanded by linearly increasing their diameter over time. The cluster diameter, duration and size of the epidemic were proportionally smaller when control interventions were more timely and intense.101 In fact, using a stochastic transmission model based on detailed entomological, epidemiological and population data, we estimated that the use of ultra-low volume insecticide prevented about 24% of symptomatic cases that occurred in the area during the 2015-2016 epidemic season.102 However, although we know a lot about the expansion of the vector mosquito, the need for rain and adequate temperatures for its growth, since disease cycles occur every 2-4 years, it is very difficult to accurately predict when dengue will enter a given location.103 The last two major dengue epidemics in Belo Horizonte had 160 thousand and 115 thousand cases in 2016 and 2019 , respectively (data from the Belo Horizonte Health Department). The first starting at the end of December of the previous year and the other only in March 2019. Knowing that there are no adequate ways to control dengue or other arboviruses (vaccines or adequate vector control measures), preparing the health system to deal with this increase in demand in a coordinated and equitable way is not simple. In a recent study, we quantitatively assessed the usefulness of data acquired by Twitter for the early detection and monitoring of dengue epidemics, 18 weekly, both at the country and city levels. We demonstrate that tweets can estimate the incidence of dengue in the current week (nowcasting), but can also predict dengue in up to 4 weeks in the future, both in the country and in the city levels with high estimation capacity. The data obtained from Twitter were better than the data available on the web, on Google Trends and the access logs to Wikipedia to predict the risk of dengue18. In fact, the use of Tweets with other epidemiological data performs well in predicting entry and spread of dengue in an urban center.103 On the other hand, to obtain data that allow a better geographical definition of the distribution of the disease in large urban centers (regions or neighborhoods) and that allow a good separation between dengue and other arboviruses, such as zika and chikungunya, is an enormous challenge. In our previous experience with the Dengue Observatory, we have noticed that the municipal manager, who makes the decisions that matter most to the population, has difficulty understanding the potential of using AI to monitor epidemics and plan health actions to mitigate its effects.
Environmental disasters are increasingly common worldwide, often causing mental disorders and various diseases in the affected populations. The cities of Mariana and Brumadinho were the scene of serious disasters caused by the rupture of dams associated with iron mining, with enormous social, environmental and health repercussions. Studies show that the post-disaster crisis initially affects the typical daily structure, which mainly impacts vulnerable groups, such as children and the elderly. This public is in stages of life more sensitive to mental disorders and are even more vulnerable in face of disorganized social structure and family nucleus. The sudden change in reality and losses can cause physiological and psychological damage, leading to worrying conditions, such as impaired cognitive development, anxiety, depression and suicide. Thus, in addition to the strong emphasis on mitigation, more research at regional and local levels is needed to properly understand, prepare and adapt to the impacts of environmental disasters 104–106. A disciplined, long-term approach to studying these problems, involving public health agencies and clinical and academic communities, is necessary to comprehensively understand the significance of the impact of these events on the health of the affected populations.
In common, disasters and epidemic events bring the enormous challenge of planning health actions in a coordinated and equitable way, so that the impact of events on populations is as small as possible. Furthermore, it is essential that the “managers at the end of the line” (municipal level) understand their value, their use, limits and advantages, so that actions can be taken quickly and effectively.
18. Marques-Toledo C de A, Degener CM, Vinhal L, Coelho G, Meira W, Codeço CT, et al. Dengue prediction by the web: Tweets are a useful tool for estimating and forecasting Dengue at country and city level. PLoS Negl Trop Dis. 2017 Jul;11(7):e0005729.
99. Bragazzi NL, Dai H, Damiani G, Behzadifar M, Martini M, Wu J. How Big Data and Artificial Intelligence Can Help Better Manage the COVID-19 Pandemic. Int J Environ Res Public Health [Internet]. 2020 May 2;17(9). Available from: http://dx.doi.org/10.3390/ijerph17093176
100. Candido D da S, Claro IM, de Jesus JG, Souza WM de, Moreira FRR, Dellicour S, et al. Evolution and epidemic spread of SARS-CoV-2 in Brazil [Internet]. Infectious Diseases (except HIV/AIDS). medRxiv; 2020. Available from: https://www.medrxiv.org/content/10.1101/2020.06.11.20128249v1
101. Guzzetta G, Marques-Toledo CA, Rosà R, Teixeira M, Merler S. Quantifying the spatial spread of dengue in a non-endemic Brazilian metropolis via transmission chain reconstruction. Nat Commun. 2018 Jul 19;9(1):2837.
102. Marini G, Guzzetta G, Marques Toledo CA, Teixeira M, Rosà R, Merler S. Effectiveness of Ultra-Low Volume insecticide spraying to prevent dengue in a non-endemic metropolitan area of Brazil. PLoS Comput Biol. 2019 Mar;15(3):e1006831.
103. Marques-Toledo CA, Bendati MM, Codeço CT, Teixeira MM. Probability of dengue transmission and propagation in a non-endemic temperate area: conceptual model and decision risk levels for early alert, prevention and control. Parasit Vectors. 2019 Jan 16;12(1):38.
104. Adeola FO. Hazardous Wastes, Industrial Disasters, and Environmental Health Risks [Internet]. 2011. Available from: http://dx.doi.org/10.1057/9780230339538
105. Shultz JM. Perspectives on disaster public health and disaster behavioral health integration [Internet]. Vol. 2, Disaster Health. 2014. p. 69–74. Available from: http://dx.doi.org/10.4161/dish.24861
106. Goldstein BD. Broadening the Mandate of the Incident Command System to Address Community Mental and Behavioral Health Effects as Part of the Federal Response to Disasters. Curr Environ Health Rep [Internet]. 2020 Jun 27; Available from: http://dx.doi.org/10.1007/s40572-020-00283-4