Skip to main content

Table 1 Default point estimates of parameters for fitting and simulating a basic COVID-19 SCLAIV outbreak on \([0,t_\text{est}]\)). Also see Fig. 2

From: A versatile web app for identifying the drivers of COVID-19 epidemics

Parameter

Symbol (\(\hbox {math}^{\dag }\))

Value

Source/Comment

Transmission

 Contact \(\hbox {rate}^\mathsection\)

\(\kappa\)

Estimated

Incidence \(\hbox {data}^{*}\)

 Nominal/effective pop size

\(N_0=N_\text{eff}\)

\(10^5-10^7\)

\(\hbox {Varies}^\P\)

 Asymptomatic reduction

\(\varepsilon\)

0.1

Unknown: see [12]

 Surveillance

\(\delta _\text{sur}\)

0.5

\(\hbox {Unknown}^{**}\)

 Isolation/treatment

\(\delta _\text{iso}\)

0.35

\(\hbox {Varies}^{\ddag }\)

 Contact rate reduction

\(\delta _\text{con}\)

0.1

\(\hbox {Varies}^{*\ddag }\)

Progression

 Thwart period

\(\pi _\text{thw}\)

1

Normalised

 Succumb \(\hbox {period}^\mathsection\)

\(\pi _\text{suc}\)

Estimated

Incidence \(\hbox {data}^{*}\)

 Latent period

\(\pi _\text{lat}\)

4\(^{\dag \dag }\) days

[7, 12, 45]

 Asymptomatic period

\(\pi _\text{asy}\)

5\(^{\dag \dag }\) days

[7, 12, 45]

 Symptomatic/recovery period

\(\pi _\text{rec}\)

7\(^{\ddag \ddag }\) days

[12]

 Immune period

\(\pi _\text{imm}\)

1000 days

\(\hbox {Unknown}^\parallel\)

 Disease-induced mort./virulence

\(\alpha ^{\dag \ddag }\)

Estimated

Mortality data

Initial values

 Initial susceptible

\(S_0\)

\(S_0=N_{\text{eff}}-C_0-L_0-A_0-I_0-V_0\)

 Initial contact

\(C_0\)

Estimated

Incidence \(\hbox {data}^{*}\)

 Initial latent

\(L_0\)

\(\left( \frac{\pi _\text{asy} }{\pi _\text{rec}}\right) G I_0\)

Requires \(G{^+}\)

 Initial asymptomatic

\(A_0\)

\(\left( \frac{\pi _\text{lat}\pi _\text{asy} }{\pi _\text{rec}^2}\right) G^2 I_0\)

Requires \(G{^+}\)

 Initial symptomatic

\(I_0\)

Estimated

Incidence \(\hbox {data}^{*}\)

 Initial immune

\(V_0=0\)

SARS-CoV-2 immunologically naïve pop

  1. \(^\dag\)Web app symbols are: \(\kappa =\mathtt{kappa}\), \(\pi _\text{thw}\)=P_thw, \(I_0\)=I_0, etc., as made clear in text
  2. \(^\mathsection\)See Eq. B.3 in Appendix B of Additional file 1 for relationship to SEIR model parameter \(\beta\)
  3. \(^*\)Best values estimated using MLE, ranges estimated using MCMC [17, 46, 47]
  4. \(^\P\)Depends on relative outbreak size (e.g., \(<10\)% of total pop.) for reliable forecasts
  5. \(^{**}\)Values fitted to the initial conditions will inversely scale with this value
  6. \(^{\ddag }\)Varies. This value implies individuals isolated within three days of entering class I or \(\hbox {I}_\text{r}\)
  7. \(^{\ddag \ddag }\)Varies. This rate assumes that sheltering-in-place reduces contact rates by 90%
  8. \(^{\dag \dag }\)Sources are highly variable, so we selected integer-valued ball park estimates
  9. \(^{*\ddag }\)Value early into epidemic since it depends on hospitalisation/isolation protocols
  10. \(^\parallel\)If sufficiently large, its influence on the initial stage of the outbreak is negligible
  11. \(^{\dag \ddag }\alpha\), being a driver, is replaced with \(\delta _\text{vir}\) in Appendix C, Additional file 1, and Vir_const in the Web App
  12. \(^+\)Requires G which, unlike \(R_0\), is a rate of increase over recovery rather than “serial” interval [1, 48]