Unorganized Machines to Estimate the Number of Hospital Admissions Due to Respiratory Diseases Caused by PM10 Concentration

Hybrid unorganized machines to estimate the number of hospital admissions caused by PM[Formula: see text] concentration

Air pollution levels exceeding the recommended limit can be the main cause of illnesses that affect human health, mainly diseases of the respiratory system. Consequently, this high exposure can impact public health management, given the increase in hospital admissions. One of the most influential air pollution parameters related to respiratory diseases is particulate matter (PM) concentrations. Thus, this paper proposes to estimate hospital admissions due to respiratory diseases caused by PM concentration with an aerodynamic diameter less than 10 [Formula: see text]m (PM[Formula: see text]), using artificial neural networks. Three hybrid neural network models are developed by combining two architectures denoted unorganized machines: extreme learning machines and echo state networks. These models also comprise extension strategies that seek to improve the generalization capability and the variation in the nonlinear outputs. Case studies explore three cities' datasets from São Paulo state, Brazil: Cubatão, Campinas, and São Paulo, to assess the quality of the hospital admissions estimations obtained by applying the proposed models. Results demonstrate that the hybrid models outperform the previously developed standard approaches in several scenarios. An overall analysis shows that the hybrid models can be a suitable strategy considering the instance particularities, especially in large datasets.

Short-term relation between air pollutants and hospitalizations for respiratory diseases: analysis by temporal association rules

This study investigates the relation between exposure to critical air pollution events with multipollutant (CO, PM₁₀, PM_2.5, NO₂, O₃, and SO₂) and hospitalizations for respiratory diseases in the metropolitan area of São Paulo (RMSP) and in the countryside and coastline, from 2017 to 2021. Data mining analysis by temporal association rules searched for frequent patterns of respiratory diseases and multipollutants associated with time intervals. In the results, pollutants PM₁₀, PM_2.5, and O₃ showed high concentration values in the three regions, SO₂ on the coast, and NO₂ in the RMSP. Seasonality was similar between pollutants and between cities and concentrations significantly higher in winter, except for O₃, which was present in warm seasons. Hospitalizations were recurrent during the transition from summer to colder periods. In approximately 35% of the total days with hospitalization greater than the annual average, one or more pollutants had a high concentration. The rules showed that PM_2.5, PM₁₀, and O₃ pollutants are strongly associated with increased hospitalizations in the RMSP (PM_2.5 and PM₁₀ with 38.5% support and 77% confidence) and in Campinas (PM_2.5 with 66.1% support and 94% confidence) and the pollutant O₃ with maximum support of 17.5%. On the coast, SO₂ was related to high hospitalizations (43.85% support and 80% confidence). The pollutants CO and NO₂ were not associated with the increase in hospitalizations. The ratio delay indicates the pollutants that were associated with hospitalizations, having concentration remained above the limit for three days, oscillating in smaller hospitalizations on the 1st day and again higher on the 2^nd and 3^rd days of delay, in a decreasing way. In conclusion, high pollutant exposure is significantly associated with daily hospitalization for respiratory problems. The cumulative effect of air pollutants increased hospitalization in the following days, in addition to identifying the pollutants and which pollutant combinations are most harmful to health in each region.

A Multi-Pollutant and Meteorological Analysis of Cardiorespiratory Mortality among the Elderly in São Paulo, Brazil-An Artificial Neural Networks Approach

Traditionally, studies that associate air pollution with health effects relate individual pollutants to outcomes such as mortality or hospital admissions. However, models capable of analyzing the effects resulting from the atmosphere mixture are demanded. In this study, multilayer perceptron neural networks were evaluated to associate PM₁₀, NO₂, and SO₂ concentrations, temperature, wind speed, and relative air humidity with cardiorespiratory mortality among the elderly in São Paulo, Brazil. Daily data from 2007 to 2019 were considered and different numbers of neurons on the hidden layer, algorithms, and a combination of activation functions were tested. The best-fitted artificial neural network (ANN) resulted in a MAPE equal to 13.46%. When individual season data were analyzed, the MAPE decreased to 11%. The most influential variables in cardiorespiratory mortality among the elderly were PM₁₀ and NO₂ concentrations. The relative humidity variable is more important during the dry season, and temperature is more important during the rainy season. The models were not subjected to the multicollinearity issue as with classical regression models. The use of ANNs to relate air quality to health outcomes is still very incipient, and this work highlights that it is a powerful tool that should be further explored.

Assessing Atmospheric Pollution and Its Impact on the Human Health

In recent decades, atmospheric pollution has become a major risk for public health and ecosystems [...]

Ambient PM Concentrations as a Precursor of Emergency Visits for Respiratory Complaints: Roles of Deep Learning and Multi-Point Real-Time Monitoring

Despite ample evidence that high levels of particulate matter (PM) are associated with increased emergency visits related to respiratory diseases, little has been understood about how prediction processes could be improved by incorporating real-time data from multipoint monitoring stations. While previous studies use traditional statistical models, this study explored the feasibility of deep learning algorithms to improve the accuracy of predicting daily emergency hospital visits by tracking their spatiotemporal association with PM concentrations. We compared the predictive accuracy of the models based on PM datasets collected between 1 December 2019 and 31 December 2021 from a single but more accurate air monitoring station in each district (Air Korea) and multiple but less accurate monitoring sites (Korea Testing & Research Institute; KTR) within Guro District in Seoul, South Korea. We used MLP (multilayer perceptron) to integrate PM data from multiple locations and then LSTM (long short-term memory) models to incorporate the intrinsic temporal PM trends into the learning process. The results reveal evidence that predictive accuracy is improved from 1.67 to 0.79 in RMSE when spatial variations of air pollutants from multi-point stations are incorporated in the algorithm as a 9-day time window. The findings suggest guidelines on how environmental and health policymakers can arrange limited resources for emergency care and design ambient air monitoring and prevention strategies.