Dynamic Causal Modelling of COVID-19

Technical details about the dynamic causal model used here can be found at https://www.fil.ion.ucl.ac.uk/spm/covid-19/.

The data used to produce the posterior predictions shown in this dashboard are available as the following CSV files. These data were assembled from the sources listed in the accompanying technical report (see Table 2 in Appendix).

See also the local dashboard and the long-term forecasting.

Disclaimer: The modelling and accompanying estimates are reported in these pages for purely academic (open science) purposes. This modelling has not been commissioned. In particular, dynamic causal modelling is not commissioned by the Independent SAGE (on which Prof Friston serves as a panellist). The Independent SAGE does not commit to – or engage in – any particular modelling initiative.

Last updated: 13th September 2021

Data (black dots) and predictions in terms of posterior expectations (blue lines) and associated 90% credible intervals (shaded areas) for several outcomes.

Underlying latent states generating the predictions above. The upper two panels show some outcomes from the previous figures (black dots). Here, daily confirmed cases using PCR testing, daily deaths, and critical care occupancy. The upper left panel shows the rates, while the upper right panel shows the cumulative totals. The remaining panels detail the fluctuations in the latent states of the four factors. Each factor has two panels, showing each of the accompanying levels. For clarity, some levels have been omitted because the probabilities of being in any level - of any given factor - sum to one. For a more detailed explanation of what these latent states mean - and how they can be interpreted - please see the appendix of the technical report.