New monitoring system can detect COVID-19 and future respiratory-virus outbreaks at airports

April 7, 2021
A new air monitoring system might stop future pandemics before they start. (AP Photo/Mark Schiefelbein)

A new air monitoring system might stop future pandemics before they start. (AP Photo/Mark Schiefelbein)

In order to prevent another global pandemic, scientists have been exploring methods to better detect and contain new viruses before they spread. To that end, a group of engineers have developed an algorithm-based monitoring system that can be integrated into airports for early detection of coronavirus and other respiratory-virus outbreaks.

In a paper published March 16 in Computers & Industrial Engineering, an international team of researchers proposed their idea for an optimized np control chart, a method used in statistical quality control for monitoring the number of nonconforming units in a sample, that monitors the infection rate of respiratory viruses, with clinical symptoms used as surrogates for infection. 

The researchers tested their optimized np chart against the traditional np chart in one real scenario at an airport and in five simulated scenarios. They found that it could "substantially outperform" its traditional counterpart in detecting a wide range of shifts in the infection rate.

"In the light of global research efforts to combat the threat of COVID-19, we studied this topic to provide policymakers with an effective monitoring scheme for early detection of outbreaks caused by coronaviruses and other important respiratory viruses," Salah Haridy, an associate professor at the University of Sharjah in the United Arab Emirates and co-author of the paper, told The Academic Times

"In particular, early detection of respiratory-virus activity at sentinel airports would allow for near real-time decisions about the potential need for specific virologic testing, patient quarantine, travel restriction, and other important outbreak investigation and mitigation measures," he continued.

Control charts serve as a visual tool that provides early identification of statistically significant changes in data. Statistical process control charts have been proved to be effective, easy to implement and inexpensive when used for virus outbreak detection, according to the authors. These charts were originally used for monitoring production processes but have been increasingly applied to health care monitoring since the 1990s. 

Haridy explained that when implemented and in use, the researchers' proposed system offers the potential to test a sample of travelers for respiratory-virus infections by measuring skin temperature with thermal cameras, thermal imaging or forehead thermometer guns; automating temperature screenings with artificial intelligence; and surveying for cough and shortness of breath. 

This system is designed to be especially useful for situations and high-volume locations, such as airports, where 100% inspection of every person is impossible, and sampling is a more realistic strategy. 

After a sample of travelers during a specific period are monitored, the number of infections detected in the sample is counted. The number of infections is plotted for each sample on a control chart. If the number of infections exceeds the upper control limit of the control chart, then a potential outbreak is declared, and a 100% inspection may be recommended. Otherwise, the infection rate is considered to be in control, and the sampling process continues. 

The researchers developed a simple code that serves as the base for the monitoring system. In addition to the code, each implementation of the system in an airport needs to establish an in-control infection rate that does not require investigation, an allowable false-alarm rate that can be managed by the airport, a sampling rate based on available resources such as manpower and virus tests, and a maximum out-of-control infection rate to be detected.

"The sample size and other charting parameters of the proposed monitoring scheme are optimized to ensure the best overall performance for detecting a wide range of shifts in the infection rate, based on the available resources, such as the inspection rate and the allowable false alarm rate," Haridy said. 

In the current study, the researchers used the average number of infections as a metric to evaluate the performance of the np chart. And in the five simulated scenarios of infection, the overall performance of the proposed optimized chart was always better than or equal to that of the traditional chart. 

The np chart was also tested between December 2019 and January 2020 — at the beginning of the COVID-19 outbreak — at a real airport with limited resources that cannot accommodate 100% inspection. The optimized np chart, which adopts inspecting a sample of 185 people every 111 minutes using a control limit of six, was able to detect a COVID-19 outbreak almost two times faster than the traditional np chart, which inspects 100 people per hour using a control limit of five. 

"The findings of this study can pave the way for more research on different statistical engineering techniques for early detection of outbreaks caused by coronaviruses and other common respiratory viruses," Haridy said. "In addition, adopting the proposed monitoring scheme in sentinel airports could help identify the origin of respiratory virus outbreaks and compare infection rates at different locations."

The authors noted that the optimized np chart is as simple as the traditional np chart to implement, particularly for public health officials without a background in statistical process control.

The study, "Monitoring Scheme for Early Detection of Coronavirus and Other Respiratory Virus Outbreaks," published March 16 in Computers & Industrial Engineering, was authored by Salah Haridy, University of Sharjah and Benha University; Ahmed Maged, City University of Hong Kong and Benha University; Arthur W. Baker, Duke University School of Medicine; Mohammad Shamsuzzaman and Hamdi Bashir, University of Sharjah; and Min Xie, City University of Hong Kong.

Saving
We use cookies to improve your experience on our site and to show you relevant advertising.