top of page


Public·229 members

James Waves We Are Not In Control

Download File

The COVID-19 pandemic has had an unprecedented impact on human health and society1,2, with high-income, urban and temperate areas often the most severely affected3,4,5. The impacts of the virus are felt directly through its substantial infection-related mortality6,7 and post-infection sequelae8, as well as through the often highly restrictive public health measures needed to achieve control9.

Australia was relatively successful in controlling COVID-19 throughout 202010, with all jurisdictions of the country achieving good control of the first wave of imported cases through March and April. However, the southern state of Victoria suffered a substantial second wave of locally-transmitted cases, reaching around 600 notifications per day, predominantly in Metropolitan Melbourne in winter.

As noted previously, stage 3 restrictions were associated with a reduction in the effective reproduction number16, although significant case rates persisted throughout July, and further reductions in mobility were observed with stage 4. An agent-based model with detailed social networks, consideration of multiple intervention types, and without geographical structure was calibrated to the Victorian epidemic17. This model emphasised the importance of associations between individuals who would not otherwise be in regular contact to the epidemic. Another agent-based simulation found that earlier activation of social distancing interventions could halve the total epidemic size18. By contrast to previous work, our model captures both the temporal and spatial implementation of the policy changes in Victoria to allow inference of the effect of each intervention. As concern increased that epidemic control had not been achieved over the course of July, the policy changed rapidly in an attempt to bring the epidemic under control. Testing numbers increased following a nadir in early June and lockdown measures were implemented differently in twelve Melbourne postcodes, the remaining postcodes of Greater Melbourne, Mitchell Shire (immediately north of Greater Melbourne) and the remainder of regional Victoria. We captured these complicated geographical patterns of restriction by scaling our mixing matrices using Google mobility data, which are available at the LGA level for Victoria. School closure and face-covering policy changes were captured according to the dates of policy changes.

In earlier versions of the model we included an effect of seasonal forcing. While good fits were also achieved with this effect included, the posterior estimate of the effect of seasonality was not markedly constrained through fitting to data. The minimal information provided on seasonal forcing was likely attributable to our simulation period spanning less than four months and so covering a small proportion of the cycling period. Therefore, while a potentially important seasonal effect would be consistent with our analysis and with evidence from elsewhere19, it was not possible to draw conclusions as to its strength. The effect of face coverings was similar to or greater than is typically estimated at the individual level20,21, but is consistent with the dominant importance of the respiratory route to transmission22. The finding was also not unexpected given the marked shift in population use of face coverings at this time and the timing of the policy change in late July relative to the dramatic reversal in case numbers occurring around one week later. The significant estimated effect of behavioural changes suggests that reductions in interpersonal associations (macro-distancing) alone were not solely responsible for the dramatic reversal in the epidemic trajectory observed. However, the Google mobility functions used to capture macro-distancing simulated falls in attendance at workplaces and other non-household locations to considerably below baseline values in several services (Fig. 6), emphasising their importance. The dramatic effect of each of these interventions on t


Welcome to the group! You can connect with other members, ge...


Group Page: Groups_SingleGroup
bottom of page