Summarising patient interactions creates a huge workload for the healthcare professionals. This research finds that patient interactions contain a lot of noise that is subjective of nature. To explore the problem area interviews with a summarisation prototype have been conducted to extract system requirements and validate those by implementing them in the prototype. Filtering noise, reinforcement learning and numerical factual correctness are of paramount importance to a successful summarisation system.

Colorectal cancer is a widespread disease that significantly impacts the health of individuals worldwide. Understanding the needs and concerns of those affected by this disease is crucial for improving patient outcomes and enhancing the quality of care. Patient web forums have emerged as valuable platforms for individuals to openly share their experiences and thoughts related to colorectal cancer, providing unique insights into the social, physical and emotional aspects of their patient journey. These forums offer a more comprehensive and authentic portrayal of patient experiences compared to traditional patient data collection methods, such as questionnaires and interviews, which may not capture the full scope of patients experiences in the colorectal cancer carepath.

However, analyzing the vast amount of unstructured data within these patient web forums presents a significant challenge. Traditional manual analysis by human experts is time-consuming, labor-intensive, and limited in scalability, making it impractical to analyze the sheer volume of patient-generated content. This is where the application of natural language processing (NLP) techniques becomes crucial. NLP enables the automated processing and analysis of textual data, allowing for efficient extraction and interpretation of the large amounts of patient forum posts.

Nevertheless, relying solely on machine intelligence, such as topic modeling and natural language generation, for interpreting patient forum data carries inherent risks, including the potential for disseminating misleading information. While these machine-driven techniques offer efficient and scalable ways to analyze and generate insights from the large amount of diverse and unstructured patient forums, they may lack the necessary contextual understanding and domain expertise to ensure the accuracy, relevance, and ethical implications for interpreting colorectal cancer patient experiences.

To close this gap between human experts and machine intelligence, this thesis explores the potential of hybrid intelligence (HI) for topic interpretation in colorectal cancer research. The main research question is: ``How can topic modeling, GPT-3.5 language generation and human expertise be combined to explore the interpretation of patient web forums in colorectal cancer (CRC) research?"

To address the research question, three human studies were conducted. The first study employed NMF topic modeling to compare topic interpretations created independently by medical workers and GPT-3.5. This comparative analysis discovered unique observations that differentiate human-written and AI-generated interpretations on online patient stories. In the second study, it was investigated how medical researchers collaborate with GPT-3.5 to develop hybrid interpretations on patient experience topics generated by the BERTopic model. A Flask web application served as the interactive platform for combining their knowledge with the AI model. Finally, the third study made professional human evaluators assess the topic relevance of the interpretations generated by medical researchers and GPT-3.5 to determine whether the combination of GPT-3.5 and human expertise leads to improved topic interpretations compared to individual interpretations.

The proposed solution to the research problem is to explore a hybrid workflow that compares, combines and validates GPT-3.5 language generation and human expertise, aiming for enhanced interpretations of topics extracted from colorectal cancer patient forums. The three studies provide opportunities for researchers and medical professionals to integrate machine intelligence from topic models and GPT-3.5 in their field of work. The hybrid workflow has conclusively demonstrated that human experts were successfully able to compare and enhance the relevance of human and GPT-3.5 interpretations of colorectal cancer patient experience topics. This allowed human experts to efficiently reach a more comprehensive understanding of patient forum data, which is essential for improving patient health in colorectal cancer research.

Colorectal cancer is the third most common cancer disease worldwide with highly fatal. It has accounted for 10% of the global cancer incidence (Bray et al., 2018). In the Netherlands, patients who have undergone treatment and surgery for colorectal cancer typically enter a five-year follow-up phase. During this period, patients often experience long-term health issues. Colorectal cancer not only has physical implications for patients but also significantly affects various aspects of their daily lives, including their social networks and mental health (Arndt et al., 2006). On the other hand, remote patient monitoring (RPM) can enhance efficiency and improve patient care delivery. However, there has been a noticeable lack of research and design focusing on the inclusion of social and mental health monitoring within these systems. According to the World Health Organization, health should include “Physical, Social and Mental” aspects. Therefore it’s important to focus on how to develop remote patient monitoring in these aspects.
The project first used a patient community journey map (Jung, 2023) to summarize the experiences of colorectal cancer patients from a large number of posts on an online communication platform and invited medical experts from the Netherlands to a co-creation session to validate the map and explore research directions.
The final research question was defined as: How to develop a remote patient monitoring system for patients’ social and daily activity and related emotion when they are away from the hospital?
The research question was followed by user research. Patients received diary studies, including finishing a probe based on their activities. Upon completion of the diary studies by the patients, there is a follow-up interview to obtain patients' feedback and sentiments regarding the probe. Subsequently, the insights from the user research act as input during a collaborative designers’ brainstorm session. Finally, these research findings were exported into the initial guidelines for developing the RPM in social, daily activity and related emotions for colorectal cancer patients. After evaluations, the final guidelines for colorectal cancer patients’ RPM and a toolkit summarizing the "develop" and "deliver" phases were presented. This toolkit allows future researchers to develop RPM in social and daily activity and related emotion for patients in other contexts besides colorectal cancer patients, supporting the remote patient monitoring system’s evolution of physical, social and mental health.

In this graduation thesis I investigate how to design for patient empowerment in the context of head and neck cancer (HNC) patients in their follow-up phase (the first 5 years after treatment). This project is set up with the Erasmus MC and was initiated to elaborate on their current service called the Healthcare Monitor (HM). The HM is a structure of electronic patient-reported outcome measures filled in by patients before every follow-up consultation. The answers are shown in a dashboard to the treating physician and used during patient consultations (Dronkers et al., 2020). What the current HM lacks however, is insight for the patients themselves into their health and addressing their concerns in between the consultations. Therefore this project aims to develop an additional service to the HM that empowers patients in between consults whilst also providing the doctor with a more frequent and realistic insight into their wellbeing. Patient empowerment can ultimately lead to a better quality of life. In this project it is viewed as a process that can be influenced through design. Therefore an inspirational model is created in this thesis that designers can use when designing for patient empowerment. HNC patients in their follow-up need to deal with a lot of different physical and psychosocial side effects (Aguilar et al., 2017). I investigate their concerns using the double loop method of Jung (2021). I analyze the community context through a computational analysis of online cancer community platforms. The patient context is investigated through a focus group with 6 care professionals and interviews with 9 ex-patients. 4 main concerns are found. Patients want to: •Be prepared for what they can expect and do •Have the overview of how they are doing •Be supported when they have a doubt •Have a relevant service for them The model for patient empowerment and patient concerns are combined to create a design vision. The vision is to provide patients with an overview that enables them to gain more control in dealing with issues that concern them. This overview was designed in a prototype application through 3 main elements; a logbook with previous experiences, a possibility to track current experiences and an overview of expected experiences based on peer experiences. This prototype was tested by 6 ex-patients who found value in the low threshold way to contact others, find trustworthy information and track their wellbeing. They valued peer experience exchange and better insight into themselves as this can ultimately lead to more reassurance in between consults. Based on this test further improvements, suggestions for the next steps and possibilities to scale the concept are presented. When developing an app for patient empowerment the Erasmus MC should; •Include peer patient experiences •Provide the patient with insight into their wellbeing over time •Integrate tools and databases in one complete overview For designing in the field of value based healthcare, this project stresses the importance of incorporating experiences of the community and using a multidisciplinary approach.

There are various organisations such as the European Society of Cardiology (2012) and American Heart Association (2013), which describe why physical activity is essential for the development in youth. Unfortunately, children who have a Congenital Heart Defect (CHD), may suffer from a lack of opportunity to perform physical activity, decreasing cognitive maturation, motor development and autonomy during childhood (Krol, 2003). This impediment arises due to a misunderstanding from parents, who do not know to what extent their child can exercise safely, and therefore adopt overprotective behaviours (Schwerzmann, Thomet, & Moons, 2016).In order to understand better overprotection during childhood, 305 online parental stories from various patient-association websites were analysed using Natural-Language-Processing techniques. The results exhibited the lifetime journey of these families, where an uncertain future evocated a constant search for symptoms. The findings of this phase were employed during generative interviews with seven families with a CHD paediatric patient to understand the continuous search for symptoms during exercise. The combination of the insights gathered from interviews and the lifetime journey was presented to five medical team members to inspire a co-creation session. The outcome of the meeting helped to define the problem and create a design vision which helped to ideate a potential Product-Service-System (PSS) solution to support parents to let their children safely perform physical activities. To encourage families to have a safe, ordinary sports-life, BO is introduced, a smart PSS aiming to support parents and their children with a CHD to understand better the safety boundaries of exercise during free-living conditions. With an activity tracker and his nine system modules, Bo aims to guide the child through different heart rate zones defined by doctors. Furthermore, Bo has a conversational agent function where parents can send concerns to the medical team and find relief when seeing their child’s heart rate zone.A functional prototype of the conversational agent was developed and implemented in the real context of four families to understand how could it influence overprotection. The implementation experience and overall concept of Bo were evaluated through in-depth interviews with paediatric CHD patients and their parents and three different specialities from the medical team. The results showed that Bo provides a supportive exploratory environment for the family, where the child can self-discover the safety boundaries and parents, instead of limiting the child, adopt an encouraging attitude towards physical activity.