tetano
Editor, Senior Moderator
Influenza Other Respir Viruses
. 2024 Dec;18(12):e70050.
doi: 10.1111/irv.70050. Biases in Routine Influenza Surveillance Indicators Used to Monitor Infection Incidence and Recommendations for Improvement
Oliver Eales[SUP] 1 2 [/SUP], James M McCaw[SUP] 1 2 [/SUP], Freya M Shearer[SUP] 1 3 [/SUP]
Affiliations
Background: Monitoring how the incidence of influenza infections changes over time is important for quantifying the transmission dynamics and clinical severity of influenza. Infection incidence is difficult to measure directly, and hence, other quantities which are more amenable to surveillance are used to monitor trends in infection levels, with the implicit assumption that they correlate with infection incidence.
Methods: Here, we demonstrate, through mathematical reasoning using fundamental mathematical principles, the relationship between the incidence of influenza infections and three commonly reported surveillance indicators: (1) the rate per unit time of influenza-like illness reported through sentinel healthcare sites, (2) the rate per unit time of laboratory-confirmed influenza infections and (3) the proportion of laboratory tests positive for influenza ('test-positive proportion').
Results: Our analysis suggests that none of these ubiquitously reported surveillance indicators are a reliable tool for monitoring influenza incidence. In particular, we highlight how these surveillance indicators can be heavily biassed by the following: the dynamics of circulating pathogens (other than influenza) with similar symptom profiles, changes in testing rates and differences in infection rates, symptom rates and healthcare-seeking behaviour between age-groups and through time. We make six practical recommendations to improve the monitoring of influenza infection incidence. The implementation of our recommendations would enable the construction of more interpretable surveillance indicator(s) for influenza from which underlying patterns of infection incidence could be readily monitored.
Conclusions: The implementation of all (or a subset) of our recommendations would greatly improve understanding of the transmission dynamics, infection burden and clinical severity of influenza, improving our ability to respond effectively to seasonal epidemics and future pandemics.
Keywords: infection incidence; influenza; influenza‐like illness; surveillance; surveillance indicators.
. 2024 Dec;18(12):e70050.
doi: 10.1111/irv.70050. Biases in Routine Influenza Surveillance Indicators Used to Monitor Infection Incidence and Recommendations for Improvement
Oliver Eales[SUP] 1 2 [/SUP], James M McCaw[SUP] 1 2 [/SUP], Freya M Shearer[SUP] 1 3 [/SUP]
Affiliations
- PMID: 39617738
- PMCID: PMC11608885
- DOI: 10.1111/irv.70050
Background: Monitoring how the incidence of influenza infections changes over time is important for quantifying the transmission dynamics and clinical severity of influenza. Infection incidence is difficult to measure directly, and hence, other quantities which are more amenable to surveillance are used to monitor trends in infection levels, with the implicit assumption that they correlate with infection incidence.
Methods: Here, we demonstrate, through mathematical reasoning using fundamental mathematical principles, the relationship between the incidence of influenza infections and three commonly reported surveillance indicators: (1) the rate per unit time of influenza-like illness reported through sentinel healthcare sites, (2) the rate per unit time of laboratory-confirmed influenza infections and (3) the proportion of laboratory tests positive for influenza ('test-positive proportion').
Results: Our analysis suggests that none of these ubiquitously reported surveillance indicators are a reliable tool for monitoring influenza incidence. In particular, we highlight how these surveillance indicators can be heavily biassed by the following: the dynamics of circulating pathogens (other than influenza) with similar symptom profiles, changes in testing rates and differences in infection rates, symptom rates and healthcare-seeking behaviour between age-groups and through time. We make six practical recommendations to improve the monitoring of influenza infection incidence. The implementation of our recommendations would enable the construction of more interpretable surveillance indicator(s) for influenza from which underlying patterns of infection incidence could be readily monitored.
Conclusions: The implementation of all (or a subset) of our recommendations would greatly improve understanding of the transmission dynamics, infection burden and clinical severity of influenza, improving our ability to respond effectively to seasonal epidemics and future pandemics.
Keywords: infection incidence; influenza; influenza‐like illness; surveillance; surveillance indicators.