Post-covid recovery planning FAQs
19/03/2021by Rob Findlay
Here are some questions that people have been asking recently, as they use Gooroo Planner for covid and post-covid planning. (Updated 19 Mar 2021)
What ‘past period’ should we use to measure demand?
The pre-covid period is probably still the best reference baseline for demand, for instance 01-Mar-2019 to 29-Feb-2020, or calendar year 2019. (Remembering that the period covered by your activity and additions uploads needs to be at least 54 weeks long to capture a full year of seasonality, and your past period needs to sit within that.)
As the NHS emerges from covid restrictions, it may become apparent that pre-covid demand is no longer a reasonable guide to post-covid demand. At that point you may wish to switch from the projection method which is based on averages, to the forecasting method which is based on what is happening now. More details here.
Can we use a short covid-era period to measure demand?
Yes. If people want to explore alternative demand scenarios then you could use a different period, for instance September and October 2020 if you want to assume that demand and activity are at England’s ‘Phase Three’ levels.
When you are using short data periods, you will almost certainly want to remove seasonal variation by setting all 106 seasonality variables to the same number (for example by updating your dataset with this high level assumptions file, selecting ‘Overwrite’, and adapted for the HeadType values you are using).
Can we assume increases or decreases in demand?
Yes, using the PastExtraDem (Extra demand in past period) and/or FutExtraDem (Extra demand in future period) fields.
You will almost certainly want to make allowance for pent-up demand: those patients who developed disease during covid but avoided being referred into hospital, and who are expected to come back at some point.
You can estimate the missing demand by comparing waiting list additions since March 2020 with additions in the comparable pre-covid months, and then multiplying the missing demand by the proportion who are expected to come back. Nobody knows what that proportion will be, but a common approach is to allow for removals other than activity (e.g. by multiplying by 0.85), and also by the proportion expected to come back (e.g. 0.8, which would give an overall multiplier of about two-thirds).
Having estimated the pent-up demand for each service, put it under FutExtraDem and update your dataset with it.
You only need to do this for new outpatients, because the knock-on impact on electives is calculated automatically. And if you are creating a two-stage model, with 1) a period of covid-restricted activity, followed by 2) a multi-year elective recovery, remember to reset FutExtraDem to zero for the second stage so that you don’t count the pent-up demand twice.
We can only deliver so much activity during covid – can I model that?
Yes. Use FutActiv (Activity in future period: where a fixed activity scenario is being modelled) to specify the exact amount of activity you intend to do.
The easiest way is often to download a report from the Report Manager, keep all the header fields you are using (e.g. HeadType, HeadSpec), delete all the other columns, and create a new column with the title FutActiv. You can then put all your plans into that column.
This method creates a kind of exceptions table: more here.
What if our activity assumption is a fraction of pre-covid activity?
This is easy to do. Download a report from the Report Manager, open it in Excel, delete all the columns apart from the headers you are using and ResActPastRate, rename ResActPastRate as FutActiv, and paste-multiply all the numbers by 0.8 or 0.9 or whatever proportion of the past activity rate you want to use.
FutActiv is the future activity in your future period, so if you change the length of the future period then remember to adjust FutActiv accordingly.
How can I plan to recover to the pre-covid waiting list size?
You can use the statistical field TgtWL (target list size) to specify the waiting list size you want for each service. Or you can amend your high level assumptions table to automatically use a list size that is already loaded: -1 = start of future period, -2 = end of past period, -3 = start of past period. You will also want to change the target waiting time (e.g. in your high level assumptions table) to a high value (e.g. 99) to stop that kicking in.
Finally, remember to choose the activity scenario to match waiting list targets, when creating your report.
Can we exclude P5 and P6 patients (who don’t want to come to hospital for the foreseeable future)?
You could mark those patients as suspended in your waiting list snapshot files, by setting SuspendFlag = 1. They will then be excluded from the waiting list, but will carry on being included when demand is measured.
The question is: should you? I would argue not. Those patients are going to be treated one day (otherwise they would be completely removed from the waiting list) and you will want to include them in your post-covid recovery plan. You may get in a tangle if you exclude them in one model, and then try to add them back in again to model the recovery.
It would probably be easier to leave all those patients in the data as usual, and then – after you have created your planning model – assume that (say) only 80% of the projected waiting list size will be active at the end of March (or whenever).
I have a different question…
Then drop us a line!Return to Post Index