As a healthcare organization or chain, how do you row as smartly as possible with the oars you have? By leveraging data-driven logistics and mathematical models to optimize healthcare processes. The potential is enormous, believe Dennis Moeke and Rob van der Mei of TKI Dinalog.
"To address the lack of human capacity in healthcare, we need to design processes more efficiently," said Dennis Moeke(right), a lecturer at Arnhem and Nijmegen University of Applied Sciences. "The puzzle that needs to be laid for this is very complicated. You don't do that on the back of a beer mat. Certainly not if you want to organize an entire healthcare chain more efficiently. The biggest challenge is at the transition points. When a patient goes from one care organization to another. If the coordination is not good, waiting times arise."
Shorter waiting times
Mathematical optimization can help reduce those wait times. Rob van der Mei, professor of Applied Mathematics at the Centrum Wiskunde en Informatica (CWI) and the VU, explains how: "I lead the Dolce Vita project. In it we are tackling the logistical bottlenecks of acute elderly care, together with Amsterdam UMC, CWI, VU and regional partnership Sigra."
"Suppose an elderly woman falls down the stairs at home and breaks her hip. Then she enters a system in which she must constantly wait for the next step. First for the ambulance, then for her turn at the emergency room and then for surgery at the hospital. After that surgery, she might be able to go to a primary care center. If that fails, she has to wait in the hospital for a spot. That's an expensive solution. Or she has to wait at home, where her care needs only increase."
Identifying bottlenecks
"In Dolce Vita, we inventoried the bottlenecks and looked at where additional capacity is needed. Or where we could build in other smart solutions. Then we calculated all kinds of 'what-if scenarios'. What do extra beds mean for waiting times? What happens if we introduce primary care, such as the WijkKliniek in Amsterdam, in regions where it is not yet in place? But also: what if in a few years the group of elderly has grown by 5 to 10 percent? Where will bottlenecks occur then? How much extra capacity will we need by then? We can turn all sorts of knobs and calculate different scenarios in a split second."
Long-term care placement
"We also built a model for placement in long-term care," Van der Mei continued. "That model matches people with a Wlz indication to vacant places in care centers. It takes into account urgency and even individual preferences of clients."
"Imagine a city with four care centers. One client would prefer to go to East. North and West he finds acceptable, but South absolutely not. Someone else is currently staying in East. That was her second choice, as she preferred to go to South. The model shows that a place will become available there in a few days. This lady can then move on to her preferred South location. And client number one can go to East."
"If you do this for all clients on the waiting list, you get a mathematical chess game with enormous potential. The match between supply and demand becomes much better. Moreover, you take advantage of economies of scale, which in many cases can reduce the waiting time by a factor of three or four."
Smart capacity planning
Dennis Moeke: "You can also use such an algorithm for smarter capacity planning within an organization. In a Belgian nursing home, we analyzed 20,000 'calls' over three months. It involved an alarm system that allows residents to request help from a care worker. You would think that the calling behavior of clients is not predictable. On an individual level, that's true. But when you look at a larger scale, it turns out that clear patterns can be identified in calling behavior."
Catching up
"If employees build up a queue of clients early in the morning, it is very difficult to clear that backlog in the following hours," Moeke continued. "As a result, employees feel like they're running their legs out from under them. Once you have insight into when it is very busy and when it is quieter, you can take that into account. That can mean deploying more staff at 8:30 in the morning and fewer in the afternoon. In our project we were able to reduce the waiting time of clients in the morning from an average of 35 to about ten minutes. The staff experience more rest as a result."
Millions up for grabs
According to Moeke, there are millions up for grabs if we give data-driven logistics a prominent place in capacity planning. "Hospitals have invested a lot in this in recent years. There are even econometricians walking around there. Fantastic! It's different in long-term care. There, the knowledge and expertise in this field is often still marginal and the board lacks a clear vision. Also, data for these purposes are often difficult to obtain. Finally, systems must be linked to share data. That is often a challenge for IT departments."
Change, according to Van der Mei and Moeke, begins at all of the aforementioned points simultaneously. Moeke: "We have gold in our hands but remain stuck in pilots with fantastic results. It is now time for the next step: don't bullshit but polish."
TKI Dinalog stimulates the development of applicable knowledge for innovations in logistics and international logistics chains. During Zorg & ict 2024, TKI Dinalog will host a knowledge session on Thursday, April 11, at 14:30 in theater 3. Here Dennis Moeke and Rob van der Mei will discuss the possibilities of tackling bottlenecks in healthcare processes from a logistics perspective. Register for free below to attend the session.
