CWI, SIGRA and Amsterdam UMC join forces to improve elderly care

Elderly people sometimes experience great difficulty getting fast, adequate and suitable medical care when it is needed. Because of this, critical situations arise such as elderly people with dementia not being able to live at home anymore or developing underlying health issues that cause them to be admitted to hospital. During moments like these, parties such as the general practitioners, district nursing services, ambulances, hospitals and short term care are involved which can make transitional care difficult. In light of the proportional rise in the ageing population, the urgency is greater than ever to improve the immediate health care chain for the elderly. To tackle the problem, CWI has been leading the Dolce Vita project (financed by NWO), together with ahti, Amsterdam UMC, the VU and SIGRA.

During the project bottlenecks in elderly care are being identified and addressed. The research partners of the project are doing this by developing an innovative data model that gives insight into the structures and inefficiencies of the current immediate care system. During this interview Bianca Buurman (Amsterdam UMC), Robert Thijssen (SIGRA) and Rob van der Mei (CWI) relay how they do this, what the initial results are and how mathematics can contribute to improving health care.

Robert: “We had already noticed that there were problems developing in elderly care in Amsterdam. When Rob van der Mei gave a presentation about his work regarding the shortening of ambulance driving times, we struck up a conversation quickly. We saw a clear link with the stall in transitional care of the elderly. This was one of the biggest problems in the city and by now it is as well in the entirety of the Netherlands.”

Bianca: “We clicked immediately because absolutely everyone in health care faces this problem. With the help of applied mathematics you regard the problem entirely differently than if you would traditionally. A unbelievable amount of parties are involved in elderly care and from Rob’s perspective this number is reduced to a few neat ‘data junctions’.”

Rob: “Some people working in healthcare look at this problem through an ‘alpha’ lens. By which I mean that they focus on individual patients. Us mathematicians rather look at patients coming into the system in streams, as if they are cars going from point A to point B. This is a completely different approach. At the same time, you need to make sure that there is room for the individual in the model. Currently, we are building this into our shared system. This is why it is useful that we are working together with 10 different health care providers which gives us access to a rich set of data and practical knowledge and makes sure that we make real societal impact, rather than just producing academic articles.”

Bianca: “We are always looking at partial solutions. Take for instance elderly people who have been admitted to hospital and have to go back home. How can this be improved? We look at the activity of ambulances that got stranded in the chain because they did not have room at the hospital. In the end these are all differing communication hubs: if the hospital does not have room the general practitioners get stuck or the ambulance has to wait. When elderly people are dismissed from hospital, they should be transferred to a nursing home where they can recover. Right now, we are trying to impact the entire chain and want to paint a picture of every aspect of it, because partial solutions cause problems on other fronts. This complete overview means that from now on, it is possible to estimate how one thing can affect the entire chain. This is something that was missing in the past, when the nursing home beds were closed and budgets were centralized. Our ideal picture is that before policies are changed, the effect of these policies will be calculated beforehand. Something that is actually quite logical to do.

Robert: “Right now the collaboration between all of the parties and the partial solutions they come up with are like a waterbed that never evens out. But soon, using our model, it will be possible to simulate what a policy intervention, additional beds, decreased capacity or even the closing of a hospital will mean concretely.”

Rob: “Other parties are already interested such as the GGZ (mental healthcare organizations) and child protective services. Particular components of such a model are very generic in nature, the same kind of bottlenecks are indeed present in other health care domains.

Robert: “SIGRA, Rob and other specialists are looking for ways to copy this model to child protective services. The main goal is prevention of placing children in foster care and placing families under supervision. At the moment elderly care is facing huge issues which will happen as well in child protective services. The labor shortages will be harrowing in ten years. The bucket will definitely overflow. Huge amounts of time are already lost looking for caring facilities. This model will make sure a lot of time and emotional burden will be decreased. You could tell a patient what their next coming months will look like. Right now, there is no way to know which brings large amounts of stress.

Bianca: “There is a lot of work to do when it comes to sharing data. Now you can see that patient’s data stays within the walls of one organization while it is extremely useful for companies to share data with each other. At the moment we are working with data from the CBS but it has a two year delay. Preferably we work with realtime data.

Rob: “We need those datastreams as input for our models so we can optimize them. You can get useful insights from datastreams. For instance: we used them to develop an allocation model for long term care. This entails that based on the individual preferences of patients a smart algorithm allocates the best place for patient A, patient B, etc. The outcome is an enormous reduction of waiting times because of the optimal finetuning that is based on individual needs and placements. A great thing about the model is that it quite generic in nature and with a few adjustments it can be employed in the GGZ, child protective services, disability care and obstetrics. Simulations concretely show that waiting times can be cut in half.

Bianca: “It makes you wonder why people have not thought of this before. For instance, at the moment people with an indication are registered for one nursing home. However, it has become clear that half of the waiting time can be achieved if they are registered for two nursing homes. Then it becomes much easier to place people which makes sure there is less pressure on informal care and district nursing services. Thus, you do not need more beds, you just need to use them more efficiently.

Robert: “Daily, thousands of people are calling for a spot. That time can be won back because you are using this model to look into the future and are optimizing at the same time. It becomes more efficient and more effective leading to nurses getting more time to deliver the needed personal care. As a patient, you know where you are in the care system at any given moment. Right now, you are surrendered. The insecurity is sickening.

Robert: “What did happen in 2019 is the sudden closing of the Slotervaart hospital in Amsterdam. We had to place patients everywhere in an instance. If we had had this system then, we could have seen all of the pile ups and we could have determined how to spread them out. If either a flu-wave or an increase of COVID patients occurs, we now have an answer to those. According to me, it is cut from the same cloth mathematically.

Rob: “That is correct. If care capacity is lower, that is just one parameter for our model. You can run the same algorithm with different input. The same goes for the ambulance system, it did not matter if there was traffic or an accident for instance, the model stays the same with different numbers. That is its power. Such a pandemic does not have to be modelled in, the capacity and demand of care changes and that is what you change in the model. That is also the power of mathematical modelling. You can use the same model with different numbers in all of the other regions.

Bianca: “We want to gain insight into the origin of urgent care needs. If you want to prevent those, you have to know how to play into them. This involves issues such as dementia, if you have heart failures, if you are short of breath, it can be anything. With experts from the city we have identified how those issues develop. It does not seem to be a linear process, if A occurs, B follows. We know now that there are 60 to 70 factors that can contribute. In addition to this, the healthcare system plays a role. You cannot give urgent care to someone for 24 hours of this group. Meanwhile it can sometimes be better to give it at home. Data can help with recognizing patterns of decline. In most cases, an event occurs a month in advance which makes a prediction that someone will be admitted to hospital almost certain.

Rob: “Over the course of 2022 we expect the first dashboard that needs to be tested by multiple parties first. That will be a very exciting but tense phase. Everybody wants to share data but some parties are keeping their cards close to their chest. There are some cold feet. A small pilot would help immensely to get people on board. After that it can go very quickly. The upside is that we already have something to offer. The dashboard is here, the algorithms are here: we can show that it really works.

Bianca Buurman is professor of Acute Geriatrics at the University of Amsterdam. She is working on developing new solutions that occur in the urgent geriatric care system to prevent certain care needs, provide hospital care from home or in the neighborhood and to set this up from within the system.

Rob van der Mei is a Senior Researcher and Manager Research and Strategy at CWI and part-time professor Applied Mathematics at the VU. His approach to research blends the walls between theoretical frameworks and their practical applications. Together with his VU colleague Sandjai Bhulai, he received the prestigious Huibregtsen-prize 2021, a life-time award for the big societal impact his research has had.

Robert Thijssen is a consultant at the Sigra Innovation Lab. Here we look at the challenges of today and those to come. In this role, strategy, technology and fantasy come together. The scope and depth in this position to solve complex social issues in collaboration is very satisfying.