Early Detection of SARS-CoV-2 Epidemic Waves: Lessons from the Syndromic Surveillance in Lombardy, Italy

Abstract

We evaluated the performance of the exponentially weighted moving average (EWMA) model for comparing two families of predictors (i.e., structured and unstructured data from visits to the emergency department (ED)) for the early detection of SARS-CoV-2 epidemic waves. The study included data from 1,282,100 ED visits between 1 January 2011 and 9 December 2021 to a local health unit in Lombardy, Italy. A regression model with an autoregressive integrated moving average (ARIMA) error term was fitted. EWMA residual charts were then plotted to detect outliers in the frequency of the daily ED visits made due to the presence of a respiratory syndrome (based on coded diagnoses) or respiratory symptoms (based on free text data). Alarm signals were compared with the number of confirmed SARS-CoV-2 infections. Overall, 150,300 ED visits were encoded as relating to respiratory syndromes and 87,696 to respiratory symptoms. Four strong alarm signals were detected in March and November 2020 and 2021, coinciding with the onset of the pandemic waves. Alarm signals generated for the respiratory symptoms preceded the occurrence of the first and last pandemic waves. We concluded that the EWMA model is a promising tool for predicting pandemic wave onset.

Web-based surveillance of respiratory infection outbreaks: retrospective analysis of Italian COVID-19 epidemic waves using Google Trends

“…In conclusion, using Google Trends to identify control chart-based outliers for non-pathognomonic symptoms such as fever, cough, and sore throat has high predictive power for anticipating COVID-19 epidemic waves 7–8 weeks ahead of the official reports in Lombardy. If combined with other syndromic sources like those of data from healthcare utilisation ( 8 ) and emergency visits ( 7 ), data from Google Trends searches may serve as a useful infodemiological tool for anticipating an impending outbreak, which can provide valuable buffer time to allocate the necessary supplies and personnel to hospitals expecting a surge in COVID-19 patients. Upon verification by prospective research comparing model performance in different regions of Italy, public health organisations are encouraged to take advantage of this free forecasting system to anticipate and effectively manage COVID-19 outbreaks throughout Italy.…”

“…Therefore, the amount of Google searches may be considered time series processes in which observations exhibit “natural” statistical variability ( 46 ). As a result of persistent random variability of the process and variations due to systematic and predictable reasons (e.g., Google search is expected to increase yearly, as well as to show a certain seasonal variability), the monitored process should be flagged as out-of-control whenever the observed value significantly exceeds that expected ( 7 ). The expected value is obtained taking into account the “natural” variability of the process ( 7 ).…”

“…As a result of persistent random variability of the process and variations due to systematic and predictable reasons (e.g., Google search is expected to increase yearly, as well as to show a certain seasonal variability), the monitored process should be flagged as out-of-control whenever the observed value significantly exceeds that expected ( 7 ). The expected value is obtained taking into account the “natural” variability of the process ( 7 ). Alterations in the process characteristics result in variations in the observed values, resulting in more observations exceeding the control limits and the process being flagged as out-of-control.…”

“…Predictors’ effects were estimated for the dummy variables month-of-the-year, with M 1 to M 11 representing January to December (skipping July, which was used as the reference) ( 7 ). The sine and cosine functions are used for considering seasonal effects.…”

“…Syndromic surveillance is an emerging approach in this field, defined as the ongoing systematic collection, analysis, and interpretation of "syndrome" specific data for early detection of public health threats (5). Syndromic surveillance systems seek to use existing data in real-time to provide immediate analysis and feedback to policymakers (6)(7)(8). Technologies using social media, search queries, and other internet resources are novel and inexpensive approaches for detecting and tracking emerging diseases.…”