Project Update: Data-driven shared decision making on cancer treatment for individual patients

In early November, we spoke to project leader Prof Emiel Krahmer and PhD candidate Saar Hommes from Tilburg University's Department of Communication & Cognition about this data2person project. Emiel is professor of Language, Cognition and Computation and focuses his research on gaining a better understanding of how people exchange information during communication, for example in a medical context. The current research has been running since 2018 and ends autumn 2023.

The project investigates how to make (already available) data available to (new) cancer patients to give them information and help them make choices about how to organise their lives after a diagnosis. This could be about support with treatment choices, but also about setting up rehabilitation, referral to physiotherapy or psychologist or about returning to work. There is a lot of data available on this that can be made insightful. So the focus here is on expectation management.

In this research, Saar focuses on communicating data to patients in an understandable language. Besides Saar, two other PhD candidates are working on this research. PhD candidate Felix Clouth (from the Department of Methodology & Statistics) is focusing on data - using prediction models - and Ruben Vromans (now working as an associate professor at the Department of Communication and Cognition) - is doing adjacent research around communicating numbers and risks. The subprojects have involved collaboration with many hospitals including the Antoni van Leeuwenhoek and the Elisabeth-TweeSteden Hospital. Netherlands Comprehensive Cancer Organisation (IKNL) has been involved as a partner from the application. IKNL manages the cancer registry in the Netherlands and has data on disease progression, possible side effects of treatments and (as part of the Profile project) quality of life with and after cancer.

Personalised approach

This project uses data from millions of Dutch cancer patients to empower newly diagnosed patients to make decisions during treatment. The research shows that this data-driven, personalised approach enables patients to make decisions together on how to organise their lives after cancer diagnosis. The research thus focuses on explaining this data to the individual patient. The research is already providing many insights around communication and personalisation. "Personalisation we can now apply not only to just the medical information but also to more personality traits. Within the hospital, it turned out that other departments - such as trauma - can also use our insights. So we can also look across different domains, a nice by-product."

Communicating data in comprehensible language

The research started by setting up the framework. Statistical models were built to understand individual patients' risks of experiencing different consequences by comparing them with patients similar to them. That process is called personalisation. Based on person and tumour characteristics, these models include not only predictions for (long-term) survival, but also factors such as side effects of treatment and quality of life after treatment.

End users (doctors and patients) were involved in all phases of the project from the beginning to gauge their wishes and needs, until the end of the project, to evaluate the developed systems. "To ensure that personal risk is understandable among a wide audience, we also conducted a survey with Centerdata's LISS panel." Participants of this panel are a good reflection of Dutch society with different levels of education and digital skills, for example. Moreover, a data-to-text system was built that automatically generates personalised explanations of the outcomes, using non-technical language and visualisations.

Saar: "One of the main conclusions is that we now know that patients like the personal statistics but at the same time find it difficult to understand these statistics. We are therefore investigating whether we can communicate these statistics more in story form, i.e. narrative. This was not in project proposal but gives a lot of colour to the data, so we have gradually come to that."


Within the project, personalisation proved how difficult it is, as a fairly new concept. "How do you explain statistics and then how do you communicate these statistics. This proved incredibly difficult and there was a need to really personalise this. "There are also ethical and emotional sides to communicating such information "with a good prognosis, communicating is different than with a less favourable prognosis, then this is a lot trickier and there is a lot involved."

Reaching even more patients (including those with less digital skills) and doctors is also a wish of this research team.


Developing a tool yourself is beyond the scope of the study. "However, there does appear to be a need for input from the research team. For instance, we now advise on personal statistics for various decision aids such as the Patient Journey app*. This allows us to connect directly with practice". (NED) is also currently developing a new module on 'personalised statistics' on the website following insights from Ruben's thesis. Because of these questions from practice, Emiel has been able to organise extension (through additional funding) for Saar and Felix so that they can further share their knowledge. "Hospitals and healthcare institutions like to inform their patients in a fine way, so it is a very hot topic and we are working on this at the right time." When she started, Saar walked along in hospitals to be present at doctor-patient talks. "Only there do you see how difficult it is to have these conversations and to explain data. This start was very valuable and I recommend it to everyone."

We will definitely continue with this!

The researchers are still very enthusiastic and happy with the experiences and results gained. The research is proving incredibly relevant; there are several ideas for new studies. "There are still plenty of questions to answer, so we will definitely continue with this," he says.

For more information on the study, the researchers and publications, see

* The Patient Journey app guides patients and their loved ones before, during and after treatment with the right information at the right time. The app is used by more than 100 healthcare facilities around the world, for many different treatments.

Related Project

Bekijk hieronder de Projecten gerelateerd aan het thema Project Update: Data-driven shared decision making on cancer treatment for individual patients