skip to Main Content
Improve asset utilization like Tetris

Shaping the demand curve to improve asset utilization is a bit like playing Tetris

Mohan Giridharadas

Mohan Giridharadas

Founder & CEO, LeanTaaS

As first published in MedCity News.


In a previous article, I talked about how higher “velocity” contributes to greater efficiencies in a healthcare operations setting. The key to increasing velocity is to focus on minimizing the non value-added time before attempting to improve the value-added time. It’s a bit like the video game Tetris.

Almost every asset, machine or human, in a hospital (such as providers, specialists, labs, pharmacy, operating rooms, infusion center, and diagnostic imaging center) has a defined peak period (“rush hour”) and long periods of underutilization on the same day, leading to a triangular utilization profile. The averaging effect of simple utilization metrics simply does not make sense. Establishing a 75 percent utilization target is meaningless if it is achieved by a 50 percent utilization level for half the day and a 100 percent gridlock for the other half of the day. That is akin to saying that you are at a comfortable temperature despite having your head in the freezer and your feet in the oven! Unfortunately, most appointment setting mechanisms are on a first-come, first-served basis which often leads to the triangular profile characteristic of a losing game of Tetris.

So what does a video game involving stacking blocks of varying shapes and sizes as efficiently as possible have to do with infusion scheduling, you may ask? It turns out scheduling patients with varying treatment lengths is a lot like Tetris, only the odds of winning are stacked (pun intended) much more against you for a variety of reasons.

In some instances, the vast majority of the demand curve simply cannot be “shaped”. McDonald’s cannot spread the lunchtime peak by trying to persuade customers to eat their lunch either at 10 am or at 3 pm. Cities cannot eliminate the congestion on their freeways by attempting to persuade businesses to start and end their hours of operation throughout the day and night instead of having most of them operate on similar 8 am to 5 pm schedules. In order to shape the demand curve, it must be possible to either:

  1. Provide incentives that persuade a sufficient number of individuals to alter their choice of timing.


  1. Create a scheduling template that can steer a sufficient number of individuals into the right appointment slot.

Incentives to persuade individuals to alter their choice of timing:

There are numerous examples of the effectiveness of such systems. When I lived in Singapore in the early 2000s, I was struck by the fact that the downtown freeway toll prices automatically ranged from $0.50 in the middle of the night to $5.00 at 7-9 am or 5-7 pm and $2 at other times. It was a novel idea at that time and was quite effective at reducing the rush-hour gridlock. As a more humorous example, I’ve seen a manufacturing plant spread out the 8-10 am peak in truck deliveries of parts by putting out free coffee and donuts at 7 am and again at 10 a.m. with a sign that read “When they’re gone, they’re gone!!” which was sufficient to persuade 20-30 percent of delivery drivers to swing by this particular plant either earlier or later on their designated route in order to not miss out on the donuts.

Scheduling templates that can steer individuals into a preferred appointment slot

It is possible to develop optimization algorithms that can figure out the perfect “Tetris game” configuration for the given volume and duration of appointments for each specific day. The optimization must incorporate the relevant parameters (e.g., we have 24 infusion chairs and operate from 7 am to 7 pm with nine nurses on duty) and account for the inevitable delays and variability that will occur. Once the optimal configuration of appointment slots has been figured out, it becomes possible for the scheduler to steer most patients into the right slot based on the availability of appointment slots of each duration throughout the day. This effectively flattens the midday peak which balances the workload for nurses and reduces the overall wait time for all patients. If the optimization is well constructed, it offers sufficient appointment slots for all appointment types at various times during the day, thereby providing patients with sufficient choice.

Back To Top