Spring 2024 marked the inaugural Driehaus College of Business Research Awards: two awards, given annually, to faculty whose research advances our understanding of important problems facing our businesses and our society.
In this interview, one of the inaugural winners, Sina Ansari, sits down to discuss his award-winning paper, which he authored with collaborators Mohammad Reihaneh and Farbod Farhadi. Using management science tools, they created an algorithm that significantly cuts down on missed and delayed appointments in hemodialysis centers – improving health outcomes and healthcare costs in the process.
Read on to learn about how the model works, how it can be applied to other contexts, and how spontaneous interactions fuel Ansari’s work.
What is hemodialysis and why was it important to study?
Hemodialysis is prescribed to patients with end-stage renal disease (ESRD). In the United States alone, there are more than 800,000 patients with ESRD. The majority of them are treated with dialysis.
Medicare expenditure for hemodialysis centers was reported to be close to $13 billion in 2019 -- and it is going up. It’s very much a heavy burden on our healthcare system.
Tell us about the problem you and your collaborators focused in on.
There’s a pressing issue in hemodialysis centers: the imbalance between the demand – the growing number of ESRD patients – and the supply – the number of available beds at these centers. And then there’s also this lack of an effective appointment scheduling method that takes patient availability into account. Together, this causes many patients to miss appointments and experience long delays.
What are the consequences? This dramatically increases the possibility of hospitalization or death over time. And this is a negative point for the centers’ quality scores – which are directly connected to their revenue. So it’s bad for the patients and bad for the centers, .
How did you realize that patient scheduling was a problem at hemodialysis centers in the first place?
Almost all of my research is based on a real-life problem. I don't sit down in my office and come up with a hypothetical situation. Usually, this happens only by being present and talking to experts in the field.
For this paper, we were completely motivated by a real-life situation. The wife of one of my collaborators is a nutritionist working at a hemodialysis center in Boston. She observed that patients were missing appointments often— at a rate that was startling for us. We met with the clinic manager and went from there.
Tell us more about your approach! Models can be very theoretical – how did you make sure your model was connected to what was happening on the ground?
We spent quite a lot of time understanding the current process inside the hemodialysis center. What is working? What isn’t? They gave us a tour. We looked at some of the schedules they had created. We came up with a list of the sorts of things our model could address. And we collaborated with staff to prioritize what was most important.
Establishing the inputs to our model required close collaboration with the hemodialysis center. It was only possible by looking at the data that they provided us and by having discussions with their staff.
Tell us about your findings and their significance. This model was developed for hemodialysis centers– but are there other contexts where it could be useful?
To the best of our knowledge, this is the first study to formulate and solve this specific patient appointment scheduling problem in hemodialysis centers using an exact algorithm. Our models showed that, by taking patient availability into account, we can reduce lost appointments by 98% on average. This method also reduced the hours of deviation from patients’ desired schedules by 46%.
This model has not been used in practice yet -- but it's given to those who can apply it. This model has the application to hemodialysis centers – but it’s not limited to that. It could also be applied to schedule chemotherapy, physical therapy, and even counseling sessions. Basically, it can be used to schedule patients for multi-appointment systems where patient availability is an important factor.
How does your research relate to your teaching?
I get energy from teaching to do research. I also incorporate my research outcomes into my classes.
One of the classes I teach is a graduate-level class called Decision-Making for Managers. In this class, I teach management science tools and discuss optimization models and how to develop them. I bring in examples from my research, including this paper, which gives validity to what we do in the classroom. Students get a chance to work with a simplified version of our model. This exposure shows them how they can model real-life problems and how they can make a difference in practice.
What is one takeaway from your work as a researcher that could apply to any field?
One of my collaborators on this paper, Farbod, is someone I met on a flight on the way to a conference. We ended up sitting next to each other; we didn’t know each other then. But we started talking and realized there were some commonalities to our research interests. That sparked the whole idea of this paper.
Life is about giving you chances and options. You have to be the one to take them. It's up to you to be there and use the time you’re given. I mean -- this research couldn't have been done if neither of us were willing to talk to each other.