Here is another question emailed to me by a reader, so I'm answering it here (with the reader's permission and removing any identifying details):
The question:
I am currently faced with a problem with scheduling breast biopsies. The next available appointment is over a month out. My imaging director's solution is more people, rooms, and equipment. This is a traditional solution and I won't allow it.
My question: how do I calculate patient demand? The schedule includes all mammos and must be coordinated with Ultrasound schedules. I can look at the daily schedule but we are meeting that, all the while patients continue to pile up. Can you help me get my arms around patient demand for this type of service?
If capacity (cases per day) is less than patient demand, we will see “patients piling up” in a backlog of appointments. If capacity is consistently less than demand, then the backlog of patients grows and the waiting times will grow.
Looking at the scheduled number of patients per day basically tells us capacity, not demand. However, I've seen cases where the number of patients seen per day was actually LESS than real capacity because of problems in the scheduling processes.
To estimate or measure demand, we need to know how many appointment requests are actually coming in each day and each week. How many phone calls or faxes are being received? How many patients are being added to that backlog?
Even then, we have to be careful because some offices might call and then NOT schedule an appointment based on the quoted appointment wait time. We'd call these “balks” in Industrial Engineering and queuing theory. You've “balked” if you've ever given up and waited away from a long line at a Starbucks or an amusement park.
Demand = # of appointments requested per day + # of appointments “balked” at because of long wait times
Let's say we are currently scheduling and seeing 20 patients per day, but 24 appointment calls come in each day, with one person balking. We need to figure out how to see 24 patients per day.
At the one hospital I'm thinking of, they improved the scheduling process so that the length of appointment schedule slots matched up with the length of actual appointments. Due to disconnects between departments (scheduling and the MRI area), they had scheduled some procedures for 90 minutes when they really only took 60. Changing the schedule didn't change the real capacity (which was limited by the number of machines, people, and hours per day)… but it allowed them to do more scans per day.
Once the effective capacity (the number of procedures per day) is greater than the number of appointment requests coming in, the backlog will start going down. Waiting times will get shorter, which will reduce “balks.”
The reader who asked the question is right that we shouldn't automatically jump to “more people and more machines” until we are absolutely sure that is necessary. Sometimes we can increase capacity by adding some daily overtime or some Saturday appointments… as a short-term measure for bringing our backlog down more quickly.
Lean is about making sure capacity (throughput) matches demand. That's what creates good flow and what leads to the most timely patient care.
What do you think? Please scroll down (or click) to post a comment. Or please share the post with your thoughts on LinkedIn – and follow me or connect with me there.
Did you like this post? Make sure you don't miss a post or podcast — Subscribe to get notified about posts via email daily or weekly.
Check out my latest book, The Mistakes That Make Us: Cultivating a Culture of Learning and Innovation:
Agree with Mark’s comments. Another possible issue is that the schedule may be driven by the physician’s desire to stay busy. While many mammo services do screenings daily, they may only do diagnostics on select days each week, and then may do biopsies on fewer days still all driven by physician preference. This may be the bigger hurdle in reducing lead time to biopsy. Few mammo services compete on timeliness as compared with other aspects of service such as valet parking, a premium coffee cart, fancy billboards and a garden atrium.
Another opportunity to meet patient demand might be to consider extended opening hours – One service I recall opened until 8.00 pm two evenings per week, supporting improved patient access after normal business hours. They introduced a staggered roster so overtime was not required.
Yes, that’s both a way to increase capacity AND to be more patient centered… what appointment times are convenient for patients? Sometimes that’s into the evening hours, after school or after work.
Great question. In order to evaluate demand on any system you need to use control charts to figure out not just the mean and range of the number of appointments requested, but also the length of them (actual length, not projected length). You’ll get two control charts, and these should be maintained as management documents so you can make adjustments based on the results. You need the number of appointments to understand what the natural variance is in the system, which tells you what you are likely to see on any given day. You need the appointment length mean and range to show you how much time you need to allocate to every appointment. Finally, Mark already pointed out you need the “balk” figure, which I would refer to as Failure Demand. You’ll measure this daily also using a similar control chart to the other two metrics. (The selection of control chart by the way depends on what you’re looking to do. There are several, but I tend to start with an X bar R chart with sub groups of days and weekend days)
By using these three charts, you can figure out what the likely demand is going to be on your system for any given day. This will tell you things like how much capacity you’re going to need to have to satisfy these demands, and what arrangements you need to make to accommodate the variance in both number and duration of appointments with surge capacity. This is why, coincidentally, it’s almost impossible to schedule every procedure start time in an OR. I’ve worked in facilities where the first start time was used as a performance metric, along with the start times of every subsequent procedure during the day. Needless to say the initial start time was hit quite regularly, but the subsequent ones were almost never hit.
Finally, you can use the failure demand chart to measure and fuel your PI initiatives. Failure demand is typically the great hidden capacity suck, and you can get a lot of it back by reducing the failure demand at source, ideally by the use of A3 methods in your workforce.
Hope that helps.
Lee Bryan