Impact of COVID-19 pandemic in the Brazilian maternal mortality ratio: A comparative analysis of Neural Networks Autoregression, Holt-Winters exponential smoothing, and Autoregressive Integrated Moving Average models

• Published in: PloS one 2024-01, Vol.19 (1), p.e0296064
• Main Authors: Cañedo, Mayara Carolina, Lopes, Thiago Inácio Barros, Rossato, Luana, Nunes, Isadora Batista, Faccin, Izadora Dillis, Salomé, Túlio Máximo, Simionatto, Simone
• Format: Article
• Language: English
• Subjects: Autoregressive models; Births; Brazil; Brazil - epidemiology; Comparative analysis; Control; Coronaviruses; COVID-19; COVID-19 - epidemiology; Engineering and Technology; Epidemics; Fatalities; Female; Forecasting; Forecasts and trends; Health aspects; Humans; Infant, Newborn; Maternal Mortality; Medicine and Health Sciences; Mortality; Mothers; Neonates; Neural networks; Neural Networks, Computer; Pandemics; Patient outcomes; People and places; Physical Sciences; Postpartum period; Pregnancy; Pregnant women; Public health; Research and Analysis Methods; Respiratory diseases; Respiratory tract infection; Retrospective Studies; Severe acute respiratory syndrome; Trinidad and Tobago; Viral diseases; Winter; Womens health; World health
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0296064

The acute respiratory infection caused by severe acute respiratory syndrome coronavirus disease (COVID-19) has resulted in increased mortality among pregnant, puerperal, and neonates. Brazil has the highest number of maternal deaths and a distressing fatality rate of 7.2%, more than double the country's current mortality rate of 2.8%. This study investigates the impact of the COVID-19 pandemic on the Brazilian Maternal Mortality Ratio (BMMR) and forecasts the BMMR up to 2025. To assess the impact of the COVID-19 pandemic on the BMMR, we employed Holt-Winters, Autoregressive Integrated Moving Average (ARIMA), and Neural Networks Autoregression (NNA). We utilized a retrospective time series spanning twenty-five years (1996-2021) to forecast the BMMR under both a COVID-19 pandemic scenario and a controlled COVID-19 scenario. Brazil consistently exhibited high maternal mortality values (mean BMMR [1996-2019] = 57.99 ±6.34/100,000 live births) according to World Health Organization criteria. The country experienced its highest mortality peak in the historical BMMR series in the second quarter of 2021 (197.75/100,000 live births), representing a more than 200% increase compared to the previous period. Holt-Winter and ARIMA models demonstrated better agreement with prediction results beyond the sample data, although NNA provided a better fit to previous data. Our study revealed an increase in BMMR and its temporal correlation with COVID-19 incidence. Additionally, it showed that Holt-Winter and ARIMA models can be employed for BMMR forecasting with lower errors. This information can assist governments and public health agencies in making timely and informed decisions.