What if you knew a bed crisis was going to happen before it happened? Could you do something to reduce its impact?
Inpatient stays are one of the most expensive areas of healthcare and managing bed capacity is a daily challenge for hospital administrators. SIMUL8 Executive Director, Health and Social Care, Claire Cordeaux introduces the new bed management solution Bed.P.A.C. Find out how better bed management can save approximately $1.8 million per quarter per hospital AND improve patients outcomes.
We hope you enjoy the workshop. If you’d like to learn more about Bed.P.A.C. please contact us
Does the simulation allow for elective patients who instead of having their operation cancelled post-admission are not admitted at all?
Yes it does - you can set thresholds for when you would expect the operation to be cancelled, so for a patient with a scheduled operation at 9am, who has waited 4 hours, you might push them into the afternoon list or decide to cancel. Or if at 3pm on Wednesday you can see that you will not be able to discharge patients on the unit until at least Friday, you could cancel elective surgeries for Thursday in advance.
Why are there 2 outliers while there are 8 empty beds?
The simulation is reporting the overall number of outliers (or boarders as we know them) over the period. If a patient could not access a bed in the unit at 10am on Tuesday morning (and our wait threshold for a bed is 4 hours) they become an outlier. In the current version, we have not pulled patients back into the unit when a bed becomes empty, so that patient remains an outlier on another unit until they have completed their length of stay. Our model allows you to experiment with different wait thresholds and also test the impact of bringing back outliers.
How complex can the scenarios be made e.g altering % discharges of patients per hour of day?
A whole range of experiments can be run including changing arrivals by hour of the day/day of the week, discharge profiles by hour and numbers of discharges, length of stay profiles etc. Please suggest scenarios you would like to run.
Does your simulation model include multiple units? PCU/ICU/MSU etc.
This is an interesting point. Our aim has been to keep patients grouped by condition/cohort rather than by unit type, but there is of course a pathway between hospital units. We can include the functionality for these arrangements although our caveat to users would be to make the model only as complex as it needs to be to answer the questions you need to solve. If the ICU is your problem, then it makes sense to include this!
>Does the model extend to the ED, or are arrivals just admissions-qualified patients?
The tool does not include the ED itself, and takes its starting point from the decision to admit. There are other tools which look specifically at EDs. This is particularly concerned with effective management of the bed resource.
From my experience, Length of Stay and service time tend to be different things (due to down-stream bed blocking). Did you use the historical LOS in your model, or did you have different service times previously calculated to populate your simulation model?
Good point. We have taken LOS from historical data arrivals by hour of the day and day of the week. From our experience the LOS profiles for a patient arriving at 6am are different from those arriving at 11am for example. This is used to show the impact if the same profiles continue, but of course the interesting thing is to change the profiles and look at the likely impact of making a change.
Are you able to look at a number of wards at any one time (i.e. at a hospital level)? Or is it constrained to on ward with one 'type' of patients at a time? How far does "long term" forecast?
Yes, we can look at multiple wards at the same time. We can forecast a year ahead based on the historic trends - and beyond that if required.
How did you input the historic data? Is it possible to set up the arrival time based on your real historic arriving time? If so, how would you do that in SIMUL8?
Yes, we use the actual arrival times, length of stays and discharge times for a defined period when the tool is using the Near Real Time functionality. This is done so that the tool reaches the current state e.g. the last patient admission that you want to predict from. The tool then uses historical data to build the arrival, length of stay and discharge trends e.g. the next day is a Monday in July, it will build a trend based on a typical Monday in July. We cannot use the actual transactional logs at this point as the aim is to predict what is likely to happen in the next 7 days.