P
Phil C
Hi All
I am trying to develop a simple model to estimate (predict) waiting times
(medical context) based on the following input data:
IR, input rate (i.e. number of requests for test x received per week). In
other words, demand.
OR, output rate (i.e. number of appointment slots available per week). In
other words, capacity.
The build-up of patients waiting (waiting list, WL) is fairly easy to
calculate (I think!), so long as the IR and the OR remain constant:
WL=(OR/(IR/OR)+WN), where WN is the week number that the test request is
received, but the waiting time itself (in weeks) is trickier, particularly
as the IR can vary from from week to week. The OR is pretty much fixed, but
the model should also be able to cope with this varying as well, due to
(usually sudden) increase (or decrease!) in capacity.
The bottom-line question to be answered is this: "Given the number of
requests already 'in the system' (on the waiting list), for a test request
received today, how far in the future would the appointment have to be?"
This, of course, assumes that everything is done on a
first-come-first-served basis. Feels like it should just be a function of WL
and OR?
I have some ideas, but I am thinking that someone must have solved this
problem (probably many times over) so I am simply asking if you can point me
in the right direction?
Thanks for your help, Phil
I am trying to develop a simple model to estimate (predict) waiting times
(medical context) based on the following input data:
IR, input rate (i.e. number of requests for test x received per week). In
other words, demand.
OR, output rate (i.e. number of appointment slots available per week). In
other words, capacity.
The build-up of patients waiting (waiting list, WL) is fairly easy to
calculate (I think!), so long as the IR and the OR remain constant:
WL=(OR/(IR/OR)+WN), where WN is the week number that the test request is
received, but the waiting time itself (in weeks) is trickier, particularly
as the IR can vary from from week to week. The OR is pretty much fixed, but
the model should also be able to cope with this varying as well, due to
(usually sudden) increase (or decrease!) in capacity.
The bottom-line question to be answered is this: "Given the number of
requests already 'in the system' (on the waiting list), for a test request
received today, how far in the future would the appointment have to be?"
This, of course, assumes that everything is done on a
first-come-first-served basis. Feels like it should just be a function of WL
and OR?
I have some ideas, but I am thinking that someone must have solved this
problem (probably many times over) so I am simply asking if you can point me
in the right direction?
Thanks for your help, Phil