Here’s an interesting paper that models the transmission of the virus and its outcomes based on the data from Wuhan before widespread quarantines were put in place:
2019-nCoV: preliminary estimates of the confirmed-case-fatality-ratio and infection-fatality-ratio, and initial pandemic risk assessment
The new coronavirus is contagious. The R0 is likely 2.5-2.9, meaning each person infected with it infects on average 2-3 more people, and the total number of people infected can double every 6-7 days (and therefore, increase 10-fold in a little over 3 weeks, 100-fold in 6 weeks, 1000-fold in 9 weeks, etc, without steps being taken to stop transmission).
That fits with the quick escalation in case numbers in Wuhan, Lombardy (Italy) and Daegu (South Korea), and no doubt also northern Iran if there was more data available there.
The coronavirus is much more dangerous and more often fatal than influenza. The analysis here includes asymptomatic and minimally symptomatic infections that didn’t present to medical attention as cases, distinguishing the
infection fatality ratio from the
case fatality ratio, and recognizes that both will vary depending on the availability of medical care (lots of ICU beds would lower both numbers). The analysis also accounts for cases being more likely fatal in certain populations (the elderly), and models the lag time for confirmed cases to be fatal, estimating the time it takes to die of the infection, or to recover.
The case fatality ratio has been widely discussed at 2-3%, but this analysis places the infection fatality ratio lower (as it should be, the difference being the asymptomatic infections that aren’t “cases” coming to medical attention): a little under 1% (0.94%, range of the estimate = 0.37 to 2.95%).
Here’s the key paragraph:
“
To estimate the overall infection-fatality-ratio (IFR), we turn to modeling to infer the total infection count and thus the true denominator. Alongside 761 confirmed cases in Wuhan as of January 25, Wu et al estimated that 75815 (37304, 130330) cumulative infections had occured in Wuhan through that time, with the infection count doubling every 6.4 (5.8, 7.1) days. However, many of these infections have not had time to be become symptomatic, be tested, and confirmed, and thus cannot contribute to the case count. Working backward from the Wu et al estimate, and assuming that the incubation period from exposure to symptom onset is 5.4 (4.2, 6.7) days and that it took a further 4.25 (3.8, 4.7) days from symptom onset to case confirmation in the dataset analyzed here, we infer that only 26650 (9460, 59400) cumulative infections in Wuhan were eligible to be reported as cases. The ratio of reported cases to reporting-eligible infections gives us an ascertainment rate for this time period of 2.9 (1.3, 8.0) percent. Under the assumption that most reporting-eligible infections that went unreported are not severe and will not progress to death, current evidence indicates that the likely overall infection-fatality-ratio (IFR) is roughly 9.4 (3.7, 29.5) per 1000 (or 0.94 (0.37, 2.9) percent).”
The bad news is that this coronavirus has the worst combination of transmissibility and relatively high mortality than any seasonal or pandemic influenza strain since 1918.
These numbers suggest we’re ~6 weeks away from a ~1000-fold increase in the current number of coronavirus cases in the community (excluding those already recognized and quarantined), unless measures are taken to successfully limit that spread. The NHL playoffs start 5 weeks from today.