The model used here is small (2,000) people but I believe the results are generalizable to larger networks. Social distancing is modeled as a reduction in network degree from 15 to 5. The results are from 15 iterations of each model.
The two lines represent two social distancing (SD) duration periods, one of 30 days and one for 60; both were implemented at 45 days after the initial outbreak. With the caveat that the 45 day start point may be later than is the case and noting too, that the model assumes a uniform implementation of SD (which is clearly not the case), three interesting observations may be made.
First, SD does appear to significantly reduce infection density. Second, when SD ends, infection density climbs again; and finally the second peak is much more severe when SD is short (30 days) than is the case for a 60 day period of SD (which is also lower than the initial peak.
At the end of the simulation (which was run for 18 virtual months), in the shorted SD scenario 68% of the population in the model had been infected while in the longer SD case, only 31% had; while that leaves a significant proportion of the population at risk, the hope is that within 18 months, a vaccine will have been developed.
The model should be considered as illustrative, not predictive, since it makes a number of assumptions that are clearly violated in practice. Nevertheless I think it is strongly suggestive of the efficacy of SD, though questions regarding the model's scalability have not thoroughly investigated.
No comments:
Post a Comment