The generation interval is the time taken from the moment a person gets infected to the moment they infect a second person. On a population level, and as each individual typically transmits to multiple people, this forms a distribution. Along with the effective reproduction number (the average number of secondary infections from a single infected person), the generation interval distribution can be used to estimate the rate of transmission, forecast future infections, estimate the effectiveness of control measures, and potentially estimate the timing of an outbreak's peak and its final size.
If two infectious diseases have the same reproduction number but one has a shorter generation interval then case numbers for that disease will rise and fall more quickly than for a disease with longer generation intervals. Similarly, if it is wrongly assumed that two diseases have the same generation interval when one is actually shorter then estimates of the reproduction number will be biased upwards. This may result in poor quality forecasts and impact the policy interventions implemented.
Estimation of the generation interval is complex as both it and the effective reproduction number may change over time, across locations, and within subpopulations. We can think of the generation interval as being composed of both an individual's infectiousness over time and their contacts with others. Both of these are likely to differ due to demographic factors (such as age) with an individual's infectiousness likely also being altered by the effectiveness of their immune system and characteristics of the disease itself. The number, and type, of contacts for infected individuals, are also likely to vary over time and this can be impacted by mitigation efforts leading to reductions in the estimated generation interval (such as contact tracing). Finally, realised generation intervals may be dependent on the transmissibility of a disease with more transmissible diseases more rapidly depleting their local networks (such as households), and local (or global) high prevalence leading to an observed reduction due to competition effects between infectors.
There are several transmission distributions that are related to the generation time including the serial interval (which is the time between the onset of symptoms for an infector and an infectee) , and the test-to-test distribution (which similarly is the interval between the case report of an infector and an infectee). These distributions may be used as a proxy for the generation time due to the difficulty in estimating the generation time. However, both of these measures are subject to a range of additional biases especially for pathogens that can transmit before the onset of symptoms, such as COVID-19.
Does Omicron have a shorter generation interval than Delta?
Will 3, or more, of the 5 most cited studies (available on the 1st of January 2023 that refer to the generation time/interval of Omicron in the title as returned by a Google Scholar search for generation, time or interval, and Omicron) which estimate the generation interval or a transmission distribution proxy (such as the serial interval, or the test-to-test distribution) conclude (if no conclusion is made a study will be discarded and the next most cited study included) that the mean intrinsic (or realised if not distinguished) generation interval of Omicron is shorter than that of Delta? If fewer than 5 studies are found then the majority conclusion from these studies will be used.
Question composed by Nikos Bosse, Sam Abbott, and Sebastion Funk