Custom arrival appointment spacing/variance in HC

Question

question

Christopher Morgan asked Mar 16 2018 at 7:29 PM Cliff King commented Apr 17 2018 at 8:02 PM

Custom arrival appointment spacing/variance in HC

I'm working on a reception area model for a large specialty center in FlexSim HC, and I'm a bit stuck for how to appropriately model my arrivals.

Patients arrive on an appointment bases, but due to significant variation in clinical scheduling, appointments can be on any 10-minute time interval per hour from 8-5pm. This seems feasible to accomplish using custom arrival tables and mapping the volume to each 10-minute interval.

The problem is, we know patients vary in their actual arrival time for appointments significantly. Our data shows slightly more early arrivals in a fairly normal distribution for arrival time relative to check in.

Is there a way to properly account for this variance using custom arrivals but maintain the actual appointment time?

Randomly spacing the data would only space the arrivals within the time interval (so a max spread of 10 minutes in this case) which would leave a bit to be desired. How would you recommend tackling this?

Thanks!

Software Version:

FlexSim HC 5.3.4

hc arrival schedule arrival arrival pattern

arrival-variance.png (46.9 KiB)

______

Cookie preferences

Your privacy is important to us and so is an optimal experience. To help us customize information and build applications, we collect data about your use of this site.

May we collect and use your data?

Learn more about the Third Party Services we use and our Privacy Statement.

Strictly necessary – required for our site to work and to provide services to you

These cookies allow us to record your preferences or login information, respond to your requests or fulfill items in your shopping cart.

YES

Improve your experience – allows us to show you what is relevant to you

These cookies enable us to provide enhanced functionality and personalization. They may be set by us or by third party providers whose services we use to deliver information and experiences tailored to you. If you do not allow these cookies, some or all of these services may not be available for you.

YES

NO

Customize your advertising – permits us to offer targeted advertising to you

These cookies collect data about you based on your activities and interests in order to show you relevant ads and to track effectiveness. By collecting this data, the ads you see will be more tailored to your interests. If you do not allow these cookies, you will experience less targeted advertising.

YES

NO

Are you sure you want a less customized experience?

We can access your data only if you select "yes" for the categories on the previous screen. This lets us tailor our marketing so that it's more relevant for you. You can change your settings at any time by visiting our privacy statement

Your experience. Your choice.

We care about your privacy. The data we collect helps us understand how you use our products, what information you might be interested in, and what we can improve to make your engagement with Autodesk more rewarding.

May we collect and use your data to tailor your experience?

Explore the benefits of a customized experience by managing your privacy settings for this site or visit our Privacy Statement to learn more about your options.

Answer 1 · 2018-03-16T22:18:56Z

Cliff King answered Mar 16 2018 at 10:18 PM Cliff King commented Apr 17 2018 at 8:02 PM

OK, I think I understand your delima a little better. You still might be able to get what you want with the Appointments table by just defining arrivals over multiple days with enough days to somewhat model the variance you see in the number of patients arriving on any given day.

Another thought I had was to use the Custom Arrivals Table with the option to dump the quantity of patients all at the start of the arrival period, and then use a Process activity as the first activity spawned in the patient's track to force a delay based on your normal distribution. The patient wouldn't be considered as "arrived" in the facility until after this initial Process activity. In order to account for the negative numbers from the normal distribution (i.e. early arrivals), you would need to have your period start time set to the earliest possible arrival time and then revise your new normal distribution to only give positive numbers. This sounds messy, and I've never tried it before, but it's an idea...

· 3

Christopher Morgan commented · Mar 19 2018 at 3:09 PM

I'm going to take a crack at the latter since it seems like that would be the most practical for what i hope to accomplish and would have the potential to be the most representative of the process.

I'll post a update in the next few days on how things go and hopefully post my results as a demo model for anyone else who might run into this issue down the line.

0 ·

Christopher Morgan Christopher Morgan commented · Apr 17 2018 at 5:41 PM

Since i always hate it when people solve a problem but forget to post results, i wanted to follow up for whomever might stumble across this post in the future.

I was able to successfully build a model for this use case which you'll find attached.

Essentially, patients arrive into a large wait area (which is hidden underground :) so its not confusing for those unfamiliar with simulation ) 3 hours early to their appointment. I mapped out the expected arrival times so that the patient waits in this area to simulate the time early/late the patient is to their appointment time and was able to curve fit each hour of day. It was found, in our case, that 3 hours early was a common enough extreme case to build our data around, so essential a wait time of 180 minutes in the "arrival area" is equivalent to arriving on time (example: patient arrives at 4 for a 7am apt, experience a wait time of 175 minutes, and arrives 5 minutes early for there appointment).

There are a few quirks I ironed out in a newer version since it is technically possible to arrive before opening/after closure in this version, but I felt this version does a great job demonstrating the concept and is easier to understand.

Arrivals are handled with a custom arrival table using a probability distribution mapped for each 5 minute interval per hour of day & day of week from using batch import from expertfit.

Hopefully this makes sense and will give others some ideas in there modeling endeavors.

-Chris M

undergroundarrivalapts.fsm

0 ·

undergroundarrivalapts.fsm (2.3 MiB)

Cliff King Christopher Morgan commented · Apr 17 2018 at 8:02 PM

Great job. Thanks for sharing Chris!!

0 ·

Answer 2 · 2018-03-16T20:36:14Z

Cliff King answered Mar 16 2018 at 8:36 PM Christopher Morgan commented Mar 16 2018 at 9:19 PM

With everything being a distribution (PCI, number of patients, and variability in arrival time) and with the number of patients being high (30) and the narrow gap between time slots (10 minutes), it seems to me that you should just use an inter-arrival time distribution and then use triggers to assign the PCI per your distribution.

· 1

Christopher Morgan commented · Mar 16 2018 at 9:19 PM

PCI's are static (one pci), only patient volume and arrival relative to appt time vary.

I don't think inter-arrival would work in this case unless i treated every hour as a different PCI and used an inter-arrival time for each one which is already sounding like a lot of trouble... and even then the majority of volume is typically distributed closer to normal times (on the hour and at 30 minute intervals), its just the fact that we have significant in between patient arrival times that complicates the matter and makes it challenging to ignore volume.

to give you an idea, the majority is distributed for common intervals like 0,10,30,&45 (this is taken from a selected hour for reference). I'd be ok leaving the odd ball cases out like 25/35/55 and grouping that data into another close interval, but the volume at these other times is significantly large

I think using an inter-arrival,even if scoped down to hour of day or a small interval, would bias us away from the true arrival pattern since it wouldn't accurately reflect the clustering of arrivals that would occur during peak arrival periods. This is why i think an arrival based on appointments with arrival variance for said appointment would be the only way to accurately simulate arrival patterns.

Maybe there is a model that demo's something like this with inter-arrivals, but i'm struggling to visualize how it could be accomplished using inter-arrival patterns.

Thoughts? Thanks for your help talking through this Cliff :)

0 ·

capture.png (12.6 KiB)

capture2.png (9.3 KiB)

Answer 3 · 2018-03-16T19:53:55Z

Cliff King answered Mar 16 2018 at 7:53 PM Christopher Morgan commented Mar 16 2018 at 8:11 PM

Not sure why you couldn't use the regulator Appointments table with your normal distribution as the variance. You're allowed to specify multiple arrivals with the same scheduled time.

· 1

question