Patient-reported outcome measures (PROMs) are increasingly used in oncology to support patient-centered care, yet their implementation in routine practice remains inconsistent. Using a communication-centered perspective which conceptualizes PROMs implementation as a process of meaning exchange between patients, clinicians, and implementation staff, this study examined PROMs implementation as a multi-stage process and explored how it unfolds in real-life oncology care, drawing on both patient and healthcare professional perspectives. A qualitative interview study was conducted in the Netherlands with patients with cancer (n = 10) and healthcare professionals and implementation staff in oncology care (n = 7). Semi-structured interviews were informed by a communication-centered framework conceptualizing PROMs implementation as a multi-stage process. Data were analyzed using hybrid inductive-deductive thematic analysis. Six interrelated stages of PROMs implementation were identified: (1) selecting and organizing PROMs use, (2) establishing purpose and motivation, (3) experiencing PROMs as a task, (4) making sense of results together, (5) applying PROMs information in care, and (6) evaluating and improving PROMs implementation. Findings showed that experiences at different stages were interconnected: processes occurring earlier in the implementation influence how PROMs were later understood, discussed, and used in care. Across roles, participants recognized PROMs’ potential to support communication beyond clinical indicators, while also describing frustration when PROMs were collected but not meaningfully used. Patient and clinician engagement influenced one another across stages, reflecting the reciprocal nature of PROMs as a communication tool. PROMs implementation should be viewed as an interdependent, sequential process. Improving connections between stages is essential to realizing PROMs’ communicative and patient-centered potential in oncology care. Practical strategies such as personalizing PROMs invitations, completing the feedback loop between PROMs results and patients, and integrating supportive tools for clinicians can help translate PROMs data into meaningful patient-centered dialogue. Patient-reported outcome measures (PROMs)—questionnaires about symptoms and quality of life—are increasingly used in cancer care. These questionnaires, often completed by patients before or between appointments, are meant to help healthcare professionals understand how patients are really doing beyond medical test results. However, in daily practice, these questionnaires are not always used in a consistent or meaningful way. This study was needed to understand why this happens. We examined how these questionnaires are introduced, completed, discussed, and used in real-life oncology care. In particular, we asked: at which points in this process do problems arise, and how do these problems affect patient care? We interviewed cancer patients and healthcare professionals in the Netherlands. We found that using these questionnaires is not a single action, but a series of connected steps. When early steps, such as clearly explaining the purpose, are not well communicated, later steps, such as discussing results during consultations, are also less effective. Both patients and professionals acknowledged the potential of these questionnaires to improve communication, but they were frustrated when answers were collected without follow-up. Our findings suggest that improving how these steps connect in practice can help turn questionnaires into more meaningful conversations and more patient-centered cancer care.

Shared experiences between deaf and hearing museum visitors are often hindered by sensory gaps, emerging from differences in sensory perception and interpretation. To better meet the needs and expectations of sensory-diverse visitors museums increasingly co-design exhibitions together with them. Which design practices support egalitarian co-design between deaf and hearing participants, however, remains unclear. Therefore, we conducted two co-design workshops focusing on shared experiences in a museum context with 18 deaf and hearing participants. To create an egalitarian setting between the participants, the use of oral and signed language was intentionally prohibited as both are considered exclusive forms of communication. Our qualitative thematic analysis of surveys, participant-generated materials, and observed behaviour showed that excluding oral and signed language encouraged more equal collaboration and joint exploration, but also slowed down the co-design process. While craft materials supported non-verbal communication during the workshops, expressing embodied experiences remained challenging without the use of oral or signed language. Our data further revealed that deaf participants often led interactions, suggesting stronger skills in body-based communication. Based on these insights, we draw four conclusions with respect to design practices that support egalitarian co-design between deaf and hearing participants and developed six recommendations for facilitating such co-design.

In the Netherlands, several public health platforms provide information and peer support to different patient groups such as people suffering from cancer or multiple sclerosis. With the rapid growth and development of Large Language Models (LLMs) in multiple sectors, these health platforms are considering the use of LLMs for personalized content generation and ease of information management. However, the sensitive nature of health information and personal experiences shared on such platforms introduces privacy and ethical risks, such as misinformation and bias. This study explores the concerns and risks associated with LLMs through interviews and focus groups with platform editors and users who use and access these platforms. Our findings show that risks related to content quality and quantity were most frequently identified. Moreover, the findings highlight the importance of disclosure if there is no human oversight, participants’ strong opposition to AI-generated blogs, and the potential of LLMs for personalization, provided users retain control over what they read. This work contributes to the ongoing discussion in human-centered computing about the ethical challenges and risks of adopting LLMs by presenting an empirical evaluation with editors and users. Moreover, the insights inform considerations and design guidelines for implementing LLMs on public health platforms.

This review gives an overview of evaluation methods for task-oriented dialogue systems, discussing the constructs, metrics and operationalisations used in previous work and highlighting the challenges in the context of dialogue system evaluation. The objective of this review is to encourage a more critical approach when evaluating dialogue systems. To that end, a systematic review of four databases was conducted (ACL, ACM, IEEE and Web of Science), which after screening resulted in 122 studies. Those studies were carefully analysed for the constructs and methods they proposed for evaluation. Four of the most occurring constructs (satisfaction, correctness, quality, and efficiency) are discussed as an example of how constructs are operationalised and measured in research. Additionally, recent developments regarding large language models are discussed for their applicability in the context of evaluation of dialogue systems. Furthermore, considerations and concerns about validity and reliability are discussed in relation to the found constructs and metrics. To improve consistency in evaluation approaches, future work should take a critical and systematic approach to the operationalisation and specification of the used constructs. To work towards this aim, this review ends with a research agenda for dialogue system evaluation and suggestions for outstanding questions.

We performed a replication study of Liebrecht and Van der Weegen on the effects of conversational human voice (CHV) and brand familiarity in chatbot conversations on social presence and brand attitude. Additionally the effect of CHV on perceptions of warmth and competence was examined. The first goal was to determine if the findings of the original study still hold, given rapid developments in chatbot technology. The study was partially replicated, with generally smaller effect sizes; employment of CHV in chatbot messages resulted in increased social presence and consequently increased brand attitude. Moreover, no significant effect of brand familiarity was found, and no significant total effect of competence. Also, no significant total effect on warmth was found. Our second goal was to explore factors researchers should consider when replicating chatbot studies and enhancing replicability of new research. Based on our experiences, we provide a checklist for future chatbot research.

Many people perceive society as more polarized than it actually is. Arguably, one prominent reason for this false polarization is the apparent hardening of online debates, raising the question whether something can be done to make online discussions more constructive. The current paper tests the effectiveness of subjective phrasing (e.g., “I think”) in reducing perceived polarization and stimulating constructive discussion online. We ask participants (N = 175; repeated-measures) to read and evaluate subjectively and non-subjectively phrased online news discussions about societally polarized and non-polarized topics. In line with our hypotheses, we find that participants perceive discussions with subjectively phrased comments as less polarized and think discussants are less disinhibited, feel more heard and experience more solidarity. Results additionally show that participants are more willing to join such a constructive discussion themselves and tend to copy the prevalent phrasing in formulating their own reaction. Auxiliary analyses show, however, that topic polarization qualifies some of these findings, which indicates interesting links between macro-societal perceptions and micro-level discussion dynamics. This study has implications for realizing deliberative democracy online: adding a simple “I think” might help counter polarization.

Personalizing digital applications for health behavior change is a promising route to making them more engaging and effective. This especially holds for approaches that adapt to users and their specific states (e.g., motivation, knowledge, wants) over time. However, developing such approaches requires making many design choices, whose effectiveness is difficult to predict from literature and costly to evaluate in practice. In this work, we explore whether large language models (LLMs) can be used out-of-the-box to generate samples of user interactions that provide useful information for training reinforcement learning models for digital behavior change settings. Using real user data from four large behavior change studies as comparison, we show that LLM-generated samples can be useful in the absence of real data. Comparisons to the samples provided by human raters further show that LLM-generated samples reach the performance of human raters. Additional analyses of different prompting strategies including shorter and longer prompt variants, chain-of-thought prompting, and few-shot prompting show that the relative effectiveness of different strategies depends on both the study and the LLM with also relatively large differences between prompt paraphrases alone. We provide recommendations for how LLM-generated samples can be useful in practice.

Implementation of patient-reported outcome measures (PROMs) in clinical practice is challenging. We believe effective communication is key to realizing the clinical benefits of PROMs. Communication processes for PROMs in clinical practice typically involve (1) health care professionals (HCPs) inviting patients to complete PROMs, (2) patients completing PROMs, (3) HCPs and patients interpreting the resulting patient-reported outcomes (PROs), and (4) HCPs and patients using PROs for health management. Yet, communication around PROMs remains underexplored. Importantly, patients differ in their skills, knowledge, preferences, and motivations for completing PROMs, as well as in their ability and willingness to interpret and apply PROs in managing their health. Despite this, current communication practices often fail to account for these differences. This paper highlights the importance of personalized communication to make PROMs accessible to diverse populations. Personalizing communication manually is highly labor-intensive, but several digital technologies can offer a feasible solution to accommodate various patients. Despite their potential, these technologies have not yet been applied to PROMs. We explore how existing principles and tools, such as automatic data-to-text generation (including multimodal outputs like text combined with data visualizations) and conversational agents, can enable personalized communication of PROMs in practice.

The rise of Generative Artificial Intelligence (GenAI) and Large Language Models (LLMs) is accelerating the integration of social robots into education. These technologies enhance robots' abilities in natural language interaction, adaptive behaviour, and personalised learning support. To advance real-world implementation, it is essential to identify the main challenges and opportunities in this field. We conducted a two-round Delphi study with 16 experts in human-robot interaction and educational technology. In the first round, participants outlined opportunities, challenges, and potential robot roles expected in the short term (1 year) and medium term (5 years). Content analysis revealed 8 opportunities, 10 challenges and 10 roles. In the second round, experts ranked their importance and feasibility across both time horizons. The results show that the most critical opportunities and challenges are also the least feasible to achieve in practice. Conversely, the proposed roles of educational robots demonstrated alignment between importance and feasibility. Experts highlighted three promising roles for robots in the GenAI era: supporting teachers in boosting learner engagement, serving as conversational interfaces for students to access knowledge and assisting teachers in supporting disadvantaged learners. These findings provide a roadmap for prioritising feasible innovations in educational robotics.

PURPOSE:To explore trauma patients' perspectives on the need, understanding and usefulness of personalized predictions after injury to assist in rehabilitation. METHODS:We performed semi-structured online interviews. Participants (N = 30 trauma patients, admitted to the hospital for an injury in the past 5 years) were exposed to a support tool that provides personalized predictions on recovery after injury. Interview data were transcribed verbatim and thematically analyzed. RESULTS:Four themes were identified. Patients expressed (1) a need for personal information and felt that information about recovery was lacking. The most important (2) reasons for needing to receive personalized predictions were reassurance and managing expectations. However, (3) understanding the prediction model was challenging. Patients expressed receiving relatively poor predictions would not undermine (4) usefulness in practice, as they would rather know about what life has in store for them so they can prepare for life after injury better. CONCLUSIONS:Trauma patients have a need for receiving personalized predictions as they perceive them to be a useful addition to clinical practice. Understanding such predictions might be challenging, so more time should be spent on how these can be communicated.

Background: Cigarette smoking poses a major public health risk, requiring scalable and accessible interventions. Chatbots offer a promising solution, given their potential in providing personalized, long-term interactions. Despite their promise, limited research has examined their efficacy and the intertwined relationship between user experience and effectiveness over an extended period of time. Methods: In this prospective, single-arm study, we developed and evaluated Roby, a 5-session chatbot intervention incorporating motivational interviewing and cognitive behavioral therapy to help smokers quit. Roby engaged Dutch adult smokers (N = 102) in conversations covering topics such as setting a quit date, managing withdrawal and cravings, and relapse prevention. The primary outcome was the continuous abstinence rate at the end of the intervention, and secondary outcomes included 7-day point prevalence abstinence, self-efficacy, and cravings. User engagement, therapeutic alliance, and interaction satisfaction were measured weekly, and the trajectory was analyzed using Linear Mixed Models. Results: Following an intention-to-treat principle, 18.6 % of participants achieved continuous abstinence, and 37.3 % achieved 7-day point prevalence abstinence. Self-efficacy significantly improved over the intervention, and cravings decreased over time. A slight decreasing trend was observed in engagement and satisfaction, likely due to a novelty effect. However, the decrease did not affect the intervention's outcomes. Conclusion: This study demonstrates the feasibility and initial usefulness of Roby, highlighting the potential for chatbots in long-term cessation support. Future research should further validate these findings with randomized controlled trials. Additional efforts should focus on monitoring and maintaining user experience in the long term to enhance effectiveness.

BACKGROUND:Patients facing a cancer diagnosis may decide to record the consultation with their clinician to ensure the recall of healthcare information. In current practice, the initiative of audio recording a consultation lies with the patient. To our knowledge, no hospital has the policy to offer routinely audio recordings. OBJECTIVE:This qualitative study aims to outline cancer patients' perspective and expectations on the barriers, facilitators, and expected clinical effects they perceive regarding the routine recording of audio recording outpatient consultations and saving these recordings in the electronic health record. METHODS:We recruited Dutch patients from an online panel on cancer. In total, 23 patients with various stages in their cancer disease were interviewed in four focus group sessions. The audio-recorded sessions were transcribed ad verbatim, and a thematic analysis was used to identify important patterns in the data. RESULTS:All patients found it difficult to remember information from their outpatient visit. Almost all patients expressed the need for routinely audio recording their healthcare consultations for storage of personalized information and later relistening to the consultation. Many patients expect that audio recordings could allow a more smooth and efficient transfer of information between healthcare specialists. Most patients stressed the need for a standard procedure, with special regards to explicit informed consent procedures and the assurance of the privacy and confidentiality of the recordings. CONCLUSIONS:There is a need to routinely offer audio recordings of outpatient visits. These recordings should be securely stored and made available through the electronic health record. Such recordings support patient empowerment by providing both informational and emotional support, helping patients to become equal partners in discussions and decision-making regarding their care. This, in turn, enhances the patient-centeredness of healthcare services. PRACTICE IMPLICATIONS:This study demonstrates that patients with cancer expect positive effects on different domains in their health care process by audio recording conversations with the clinician.

The current study evaluates the relative contributions and interaction of procedural arguments and narrative content in health interventions. Passive health interventions often fail to make health threats relevant for a young target population. While serious games have shown promise in eliciting health behaviors, the relative contribution of mechanics themselves is often hard to attribute. The current study addresses the question of whether the presence of meaningful game mechanics, in the form of procedural arguments, contributes uniquely to persuasion by heightening susceptibility and behavioral intention. Using an interactive fear-appeal the authors present the design and evaluation of a boardgame that aims to capture persuasive arguments surrounding alcohol addiction. The mechanics and narrative framing were manipulated to be able to isolate their contributions. The study supports the notion that mechanics and rules alone might not be sufficient for players to identify the content of a procedural argument, with implementation of narrative content being an important factor in making mechanics persuasive. The current study contributes to the field by being one of the few works that operationalizes the concept of procedural rhetoric, providing implications for the design of mechanics and their integration for serious games used in health behavior change.

Natural Language Processing (NLP) offers significant opportunities to support the Sustainable Development Goals (SDGs), including Zero Hunger. While many NLP applications have been documented for SDGs such as healthcare and education, its application to food security remains largely unexplored. This paper addresses this knowledge gap through a comprehensive scoping review focused on NLP for food security policymaking. Six key application areas were identified: (1) Early warning systems for food insecurity, (2) Understanding public discourse on food related issues, (3) Knowledge generation and management from food policy and program documents, (4) Understanding dietary habits, (5) Food item classification, and (6) Addressing data gaps in food security statistics and crisis response. However, limited deployment hinders real-world impact. Establishing authentic partnerships from the outset will be essential for the successful and sustained implementation of NLP projects that advance progress toward ending hunger and achieving food security and improved nutrition for all.

BACKGROUND: A healthy lifestyle can have various health benefits for cancer survivors, including positive effects for their general health and cancer-specific outcomes. In general, cancer survivors adhere more to smoking and alcohol recommendations, compared to diet and physical activity recommendations. METHODS: To better understand the differences in benefit perceptions among the four health behaviours, we conducted a cross-sectional study among 197 Dutch survivors with various cancer types. RESULTS: We found a positive association between perceiving benefits and engaging in respective health behaviours. Furthermore, cancer survivors perceived these health behaviours differently, with low alcohol intake and not smoking being perceived as more beneficial for cancer outcomes and general health, compared to having a balanced diet and engaging in physical activity. Vulnerable groups, including those recently diagnosed and those with lower perceived control, perceived health behaviours as less beneficial. CONCLUSIONS: Diet, physical activity, not smoking and limited alcohol intake are perceived as differently beneficial by cancer survivors. These results hold important implications for patient education and communication of health behaviours to cancer survivors.

In written language, demonstratives such as this and that allow writers to produce coherent texts and readers to build up a consistent mental model of the message that is conveyed. But what makes a writer decide to use one demonstrative (e.g., this) over another (e.g., that)? Here we present experimental evidence, from both Dutch and Mandarin, that discourse genre is the main predictor of writers' demonstrative use in text. Specifically, the results of a text elicitation task show that expository texts mainly elicited proximal demonstratives (this, these, here) while narrative texts showed a significant increase in distal demonstrative (that, those, there) use. This finding is taken to reflect that writers mentally position textual referents in psychological proximity to themselves or to the reader as a function of the genre of their text.