RESUMEN
OBJECTIVE: To analyze the underdiagnosis of COVID-19 through nowcasting with machine learning in a Southern Brazilian capital city. METHODS: Observational ecological design and data from 3916 notified cases of COVID-19 from April 14th to June 2nd, 2020 in Florianópolis, Brazil. A machine-learning algorithm was used to classify cases that had no diagnosis, producing the nowcast. To analyze the underdiagnosis, the difference between data without nowcasting and the median of the nowcasted projections for the entire period and for the six days from the date of onset of symptoms were compared. RESULTS: The number of new cases throughout the entire period without nowcasting was 389. With nowcasting, it was 694 (95%CI 496-897). During the six-day period, the number without nowcasting was 19 and 104 (95%CI 60-142) with nowcasting. The underdiagnosis was 37.29% in the entire period and 81.73% in the six-day period. The underdiagnosis was more critical in the six days from the date of onset of symptoms to diagnosis before the data collection than in the entire period. CONCLUSION: The use of nowcasting with machine learning techniques can help to estimate the number of new disease cases.
Asunto(s)
COVID-19 , Brasil/epidemiología , Ciudades , Humanos , Aprendizaje Automático , SARS-CoV-2RESUMEN
ABSTRACT: Objective: To analyze the underdiagnosis of COVID-19 through nowcasting with machine learning in a Southern Brazilian capital city. Methods: Observational ecological design and data from 3916 notified cases of COVID-19 from April 14th to June 2nd, 2020 in Florianópolis, Brazil. A machine-learning algorithm was used to classify cases that had no diagnosis, producing the nowcast. To analyze the underdiagnosis, the difference between data without nowcasting and the median of the nowcasted projections for the entire period and for the six days from the date of onset of symptoms were compared. Results: The number of new cases throughout the entire period without nowcasting was 389. With nowcasting, it was 694 (95%CI 496-897). During the six-day period, the number without nowcasting was 19 and 104 (95%CI 60-142) with nowcasting. The underdiagnosis was 37.29% in the entire period and 81.73% in the six-day period. The underdiagnosis was more critical in the six days from the date of onset of symptoms to diagnosis before the data collection than in the entire period. Conclusion: The use of nowcasting with machine learning techniques can help to estimate the number of new disease cases.
RESUMO: Objetivo: Analisar o subdiagnóstico da COVID-19 por meio de nowcasting com machine learning em uma capital do sul do Brasil. Métodos: Estudo ecológico observacional utilizando dados de 3.916 casos notificados de COVID-19 de 14 de abril a 2 de junho de 2020 em Florianópolis, Brasil. O algoritmo de machine learning foi usado para classificar os casos que ainda não tinham diagnóstico, produzindo o nowcasting. Para analisar o subdiagnóstico, foi comparada a diferença entre os dados sem nowcasting e a mediana das projeções com nowcasting para todo o período e para os seis dias a partir da data de início dos sintomas. Resultados: O número de novos casos sem nowcasting durante todo o período foi de 389, com nowcasting foi de 694 (IC95% 496-897). No período de seis dias, o número sem nowcasting foi de 19 e 104 (IC95% 60-142) com nowcasting. O subdiagnóstico foi de 37,29% em todo o período e 81,73% no período de seis dias. O subdiagnóstico foi mais crítico em seis dias, desde a data do início dos sintomas até o diagnóstico antes da coleta de dados, do que em todo o período. Conclusão: O uso de nowcasting com técnicas de machine learning pode ajudar a estimar o número de novos casos da doença.