Dear Maintenance Team,
Thank you for making this model available to the public.
I had a number of questions about the user guide for the explicit SEEIR model at https://mrc-ide.github.io/squire/reference/parameters_explicit_SEEIR.html. I wondered whether you might be able to provide me with help in answering these questions. I apologise in advance if I have simply failed to understood your instructions.
Q.1 The explicit SEEIR model includes (time delay) variables relating to the provision of oxygen (e.g. dur_get_ox_die) or mechanical ventilation (e.g. dur_get_mv_die). The diagram labelled "Decision Trees for Healthcare Capacity" shows the logic for the provision of oxygen. Am I right in thinking that the model uses precisely the same logic for mechanical ventilation?
Q.2 In the section entitled "2. Changing parameters in the model", the third example shows the command required "To show an 80% reduction after 50 days but only maintained for 30 days". Should the vector generated for tt_contact in that example be c(0,50,80)?
Q.3 In the same section, the fourth example refers to "a changing R0, which falls below 1 after 50 days". Should the vector generated for tt_R0 in this case be c(0,50)?
Q.4 Within the section entitled "4.Calibrating the Model to Observed Deaths Data", is it necessary for the CSV file specifying daily deaths to contain a "cases" column?
Q.5 In the same section (4) it says that "Calibration occurs by scanning across a range of possible R0 values and start dates." How does the user retrieve the calibration's best estimates for the R0 and start date? Must these be inferred from the "Probability" grid shown further down the page or are these values made available explicitly via the results returned from the calibrate() function?
Q.6 Within the section describing the introduction of interventions into the calibration process, what is the format of the CSV file used to describe interventions? The example shown suggests that this CSV file is structured as a "date" column and a "change" column. Is that correct? Must the "change" column be labelled "C" within the CSV file? Is it necessary to leave a blank line between the column headings and the subsequent rows of data in the CSV file?
Q.7 Continuing with the theme of the CSV file used to specify interventions, do the "change" values in the example shown refer to: (a) the absolute value of R immediately following the associated "date"; (b) the (incremental or fractional) change in the value of R at the associated date; or (c) something else? In particular, in the example provided and assuming a starting value of Ro, is the value of R after the first "date": (a) Ro-0.900; (b) Ro*0.900; (c) 0.900; or (d) something else?
Q.8 Again on the issue of the CSV file defining interventions, must the user specify a "change" for each "date" or is the calibrate() function able to generate estimates for successive R values given just the dates on which this variable changes?
Thank you in advance for your help with these queries.
Best regards,
Mathew Davies
Dear Maintenance Team,
Thank you for making this model available to the public.
I had a number of questions about the user guide for the explicit SEEIR model at https://mrc-ide.github.io/squire/reference/parameters_explicit_SEEIR.html. I wondered whether you might be able to provide me with help in answering these questions. I apologise in advance if I have simply failed to understood your instructions.
Q.1 The explicit SEEIR model includes (time delay) variables relating to the provision of oxygen (e.g. dur_get_ox_die) or mechanical ventilation (e.g. dur_get_mv_die). The diagram labelled "Decision Trees for Healthcare Capacity" shows the logic for the provision of oxygen. Am I right in thinking that the model uses precisely the same logic for mechanical ventilation?
Q.2 In the section entitled "2. Changing parameters in the model", the third example shows the command required "To show an 80% reduction after 50 days but only maintained for 30 days". Should the vector generated for tt_contact in that example be c(0,50,80)?
Q.3 In the same section, the fourth example refers to "a changing R0, which falls below 1 after 50 days". Should the vector generated for tt_R0 in this case be c(0,50)?
Q.4 Within the section entitled "4.Calibrating the Model to Observed Deaths Data", is it necessary for the CSV file specifying daily deaths to contain a "cases" column?
Q.5 In the same section (4) it says that "Calibration occurs by scanning across a range of possible R0 values and start dates." How does the user retrieve the calibration's best estimates for the R0 and start date? Must these be inferred from the "Probability" grid shown further down the page or are these values made available explicitly via the results returned from the calibrate() function?
Q.6 Within the section describing the introduction of interventions into the calibration process, what is the format of the CSV file used to describe interventions? The example shown suggests that this CSV file is structured as a "date" column and a "change" column. Is that correct? Must the "change" column be labelled "C" within the CSV file? Is it necessary to leave a blank line between the column headings and the subsequent rows of data in the CSV file?
Q.7 Continuing with the theme of the CSV file used to specify interventions, do the "change" values in the example shown refer to: (a) the absolute value of R immediately following the associated "date"; (b) the (incremental or fractional) change in the value of R at the associated date; or (c) something else? In particular, in the example provided and assuming a starting value of Ro, is the value of R after the first "date": (a) Ro-0.900; (b) Ro*0.900; (c) 0.900; or (d) something else?
Q.8 Again on the issue of the CSV file defining interventions, must the user specify a "change" for each "date" or is the calibrate() function able to generate estimates for successive R values given just the dates on which this variable changes?
Thank you in advance for your help with these queries.
Best regards,
Mathew Davies