Adapting the Flexible Farrington Algorithm for daily situational awareness and alert system to support public health decision-making during the SARS-CoV-2 epidemic in England.

Abstract

The Flexible Farrington Algorithm (FFA) is widely used to detect infectious disease outbreaks at national/regional levels on a weekly basis. The rapid spread of SARS-CoV-2 alongside the speed at which diagnostic and public health interventions were introduced made the FFA of limited use. We describe how the methodology was adapted to provide a daily alert system to support local health protection teams (HPTs) working in the 316 English lower-tier local authorities. To minimize the impact of a rapidly changing epidemiological situation, the FFA was altered to use 8 weeks of data. The adapted algorithm was based on reported positive counts using total tests as an offset. Performance was assessed using the root mean square error (RMSE) over a period. Graphical reports were sent to local teams enabling targeted public health action. From 1 July 2020, results were routinely reported. Adaptions accommodated the impact on reporting because of changes in diagnostic strategy (introduction of lateral flow devices). RMSE values were relatively small compared to observed counts, increased during periods of increased reporting, and were relatively higher in the northern and western areas of the country. The exceedance reports were well received. This presentation should be considered as a successful proof-of-concept.