Several years ago when SC&I’s Associate Dean for Research Mark Aakhus approached Assistant Professor Sunyoung Kim about joining a research team at SC&I examining how technology can help physicians communicate treatment choices and risks to their cancer patients in a less stressful way, she agreed to join immediately.
“My overarching research objective is to explore ways in which we can facilitate existing personal technologies to improve and promote quality of life in many aspects, but mostly in health issues and healthcare” Kim said. “So critical health is one of my major research areas. I always want to find out ways to make emerging personal technologies easier and more useful for laypeople to fulfill whatever their needs to achieve quality of life.”
Through this project, a collaboration between faculty members Kim, Associate Professor Lisa Mikesell, Aakhus, and students Sarah Fadem and Beatrice Trinidad at SC&I, and Roger Strair, M.D., Ph.D., at the Rutgers Cancer Institute of New Jersey, they tested different types of graphical representations to determine which formats would be the most effective means to present potentially devastating information to leukemia patients about their illness, such as their chances of survival and the possible difficult side effects they might suffer from as a result of Bone Marrow Transplant (BMT), the best treatment option for many of them.
The team’s ultimate goal is to use these graphical representations to create a patient portal on a website that leukemia patients can use autonomously in an outpatient setting as a comprehensive resource for clinical and educational information about their illness and treatment options.
The team decided to study BMT because “it mimics many chronic conditions and medical treatments with an accelerated treatment trajectory,” the team wrote.
The original project was spearheaded by Strair. As Chief of Blood Disorders and a Medical Oncologist at CINJ, he is responsible for delivering very difficult news to his patients and their families. He had reached out to SC&I to initiate a collaborative study to explore ways he could communicate this information in a way that would cause his patients the least amount of stress and worry.
The team began by interviewing patients who had gone through Bone Marrow Transplant (BMT) to ask what their experiences were like and to find out what information they’d wish they’d had and not had before the surgery.
The team also interviewed clinicians to learn what clinical information the doctors needed to be sure the patients were provided and thoroughly understood. “The clinicians’ objectives were really clear,” Kim said. “They said the information they need to communicate to patients can sometimes be daunting, unpleasant, and scary, and they needed a better and less stress inducing way to convey it. There are two kinds of information the clinicians need help presenting. One is progress information, based on the patients’ condition, characteristics, type of diagnosis, and projection data, such as chances of survival. The other type is educational information. This includes explaining what BMT surgery is, and its possible side effects.
“The doctors need to let patients know that getting a bone marrow transplant is a real life-changing event. Their entire lives will change after getting a transplant. Their eating, their daily behaviors. They might lose some hair or develop diabetes. There are a lot of potential side effects, so a lot of educational materials are needed to convey to patients what their prospects are.”
When Kim joined the team, the only means the clinicians employed for delivering all of this information was by holding one three-hour educational session a month. Patients attend the session together, and their families can also attend. Most of the patients interviewed said they found the session to be an overwhelming amount of information. The clinical staff therefore knew they needed to find a better approach.
Based on the initial interviews they conducted, the team “created prototype sketches in the form of a survival calculator and experience videos,” Kim said.
The idea for the survival calculator is that it would focus on clinical aspects of the disease and treatment. The team wrote, “Using patient-specific data, disease type, and donor characteristics, the algorithm developed for a survival calculator would generate predictive personalized likelihoods of the most clinically and personally relevant outcomes after BMT treatment: survival rates and outcomes relating to graft-versus- host disease (GVHD).”
Experience videos would support patients’ emotional and psychological needs and would be made by BMT survivors, explaining to newly diagnosed patients what they could expect physically and emotionally.
The team published their findings in two papers. The first, titled “Designing a Personalized Support Tool for Patients facing Bone Marrow Transplant” with co-author SC&I’s Sarah Fadem, was published in May 2019.
The second paper, “Improving prognosis communication for patients facing complex medical treatment: A user-centered design approach,” was co-authored by Beatrice Trinidad, one of Kim’s former Information Technology and Informatics (ITI) students, and was published in April 2020.
For the second paper, the team examined in much greater depth the best ways to present graphical information.
“The videos patients listened to from other patients explained what they are going to go through, and they don’t contain much scary information,” Kim said. “However, the projection data might show your chances of living might be only 30% or 15%. We faced ethical issues – do we really want to show this information to patients? This is why for the second paper we spent more time on making sure we were designing this interface not only in a user-friendly way but in an ethically appropriate way, and it’s also why we conducted another round of user studies. We had three or four different interfaces. All of the interface prototypes we tested were based on our user studies of previous sessions.”
They tested six different sketch ideas representing different ways to show numerical information, Kim said. These included a pie chart, a bar chart, an abstract graphical person, a heat map, and others.
Kim said they were surprised by two outcomes of this study for the first paper. “We thought because this information can be really unpleasant, we should make one set of images as abstract as possible so we didn’t show the patients specific numbers. We expected that patients who are under a lot of stress might not want to be exposed to direct and possibly negative information, and we thought the abstract designs might convey the numerical information, such as survival rates, in a less stressful way.
“Our presumptions were wrong. What the patients hated was uncertainty. The abstract design ideas were perceived as uncertain. The patients wanted the actual numbers. They questioned us about the colors we’d chosen. They said they don’t want to have to interpret anything – they just want to see the plain number. So that was striking to us. They wanted the interfaces to be very straightforward and simple. That’s why in the most recent study (published in the second paper), we got rid of all abstract information in the design and decided to stick with bar graphs and pie charts, those do not have any abstract components.”
Another surprising finding from their most recent study, Kim said, stemmed from an ethical issue they realized they needed to resolve. “We asked ourselves, if the critical number is really low in terms of their chances of survival, do we still want to show that to patients right away? We thought not, so we changed the design so patients can only see the actual number if they hover over the graph. In other words, if they want to see it, they have to take an extra step.
“What I found interesting about this was that even though patients were under a lot of stress – both emotional and physical stress -- they wanted to have some kind of autonomy. They liked the fact that they were given an option to choose or not to choose to see all the data pertaining to their illness at once. This way, they know the information is there, but the system is not going to show it to them unless they are ready to see it.”
The group also tested animated data which Kim said proved interesting. “Animation is not something I intended to study to be honest. Our student Beatrice was in charge of the implementation of typing as well as data collection and she had a lot of technical skills, so she made one of the prototypes animated. Since she did that, I suggested we do a few variations of them just to test them as another control factor.
“I didn’t expect anything, but it was interesting. The directional effect of animation meant the patients interpreted the numbers that appeared to be growing as a positive indication of their chances of survival. It made them feel optimistic. In other words, seeing their chances of survival grow in an animated chart from 10, to 20, to 30, to 40, instead of being shown a graph that just showed 40 percent, made them feel better, and the patients want to be positive.”
Kim said another finding that surprised them was the patients said that even though the patient portal would provide them with all the information they needed, they still wanted to review it for the first time with their doctors so they could ask questions. “The clinicians said this as well,” Kim said. “They told us it’s much easier to talk to patients when they are looking at the same screen and sitting next to each other.”
The patient portal is currently being developed. In terms of next steps, and the potential of developing additional patient portals in the future designed for patients suffering from other illnesses, Kim said, “Once this system is implemented as a working system, we can use this as a use case to show other medical domains that have similar obstacles communicating with patients and need similar tactics and strategies in that domain. But right now, my goal is to implement this and run longitudinal field studies to investigate how interacting with this personal interface will influence patients’ understanding of their treatment -- it is positively or negatively influenced. I want to measure that.”