Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Maria Chatzimina has a total of 88 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Consumer Complaint Mining Through Topic Modeling and Sentiment Analysis: A Proof-of-Concept Analytical Framework for Decision Support

    by Maria E. Chatzimina, Athina Bourdena, Anitha Chinnaswamy, Nikolaos Trihas, Markos Kourgiantakis, Konstantinos Vassakis, George Mastorakis

    Consumer complaint narratives are an underused source of business and regulatory evidence. This study presents a proof-of-concept analytical framework for multi-dimensional complaint analysis and uncertainty-aware evidence synthesis, with potential use in human-in-the-loop decision support, and applies it to 99,434 Consumer Financial Protection Bureau (CFPB) complaint records spanning from March 2015 to March 2026. The empirical workflow integrates BERTopic topic discovery, two pretrained sentiment checkpoints (ProsusAI/finbert and cardiffnlp/twitter-roberta-base-sentiment-latest), exploratory clustering of complaint records, and descriptive temporal summaries. A conceptual stock–flow and feedback representation links observed complaints, analytical alerts, response capacity, and unresolved issues; these relations are not causally estimated or simulated. BERTopic produced 89 non-outlier topics. A probability-based outlier-reassignment stage assigned 80.46% of records to these topics, while 19.54% remained unassigned and were retained as an uncertainty queue. A FinBERT-prediction-stratified benchmark of 600 complaint records was independently coded by two annotators (raw agreement = 87.5%; Cohen’s κ = 0.758). Against Annotator 1 as the prespecified reference, Cardiff RoBERTa aligned better with the labels on the prediction-stratified benchmark (κ = 0.395; sample accuracy = 67.3%) than FinBERT (κ = 0.195; sample accuracy = 46.3%), although neither model supports autonomous use. The four-group K-means solution is reported as exploratory because the highest observed Silhouette coefficient was only 0.106 and favored K = 2. Temporal patterns coincided with selected external events, but no causal effect is claimed. The contribution is therefore a transparent, uncertainty-aware analytical architecture whose potential organizational uses require human review and operational validation.

  • Open Access

    An Emotion-Aware Conversational Agent for Symptom and Quality-of-Life Monitoring in Hematologic Malignancies

    by Maria Evangelia Chatzimina, Helen A. Papadaki, Charalampos Pontikoglou, Kostas Marias, Manolis Tsiknakis

    BackgroundPatients with hematologic malignancies often experience a high burden of physical symptoms together with emotional distress, especially during long disease trajectories and in palliative care settings. Continuous monitoring of symptoms and quality of life (QoL) may support timely clinical follow-up, but this is not always systematically implemented in routine care. Although digital health tools are increasingly used, many existing conversational systems mainly focus on structured data collection and do not adequately respond to the emotional state of patients during interaction.ObjectiveThe aim of this study was to develop and perform an initial expert-based evaluation of an emotion-aware conversational agent that supports symptom and QoL monitoring in patients with hematologic malignancies, while adapting its behavior to show empathy when emotional distress is expressed.MethodsWe developed a bilingual conversational agent based on a large language model fine-tuned on real Greek clinician–patient conversations and adapted to the clinical domain of hematology. A separate sentiment analysis module analyzes patient messages in real time to detect negative emotional expressions, such as fear, sadness, or anxiety. When emotional distress is identified, the system temporarily shifts from structured medical questioning to short supportive and empathic responses and resumes monitoring once emotional intensity decreases. The system supports both initial patient profiling and follow-up interactions and is available through web and mobile applications, including speech-to-text functionality. An expert-based evaluation was conducted using simulated patient scenarios. User experience was assessed with the User Experience Questionnaire (UEQ), and perceived empathy was evaluated using an adapted version of the Consultation and Relational Empathy (CARE) measure.ResultsThe evaluation showed positive user experience across all UEQ dimensions. Experts also reported high perceived empathy scores, indicating that the conversational agent was considered clear, usable, and emotionally appropriate during sensitive interactions.ConclusionsThe findings suggest that emotion-adaptive conversational agents based on large language models can support patient-centered symptom and QoL monitoring in hematologic malignancies and may complement clinical care, particularly in palliative and long-term follow-up contexts. The proposed modular architecture can be adapted to other chronic diseases where emotional support is an important component of care.

  • Assessing an eHealth Intervention on Quality of Life in Patients With Chronic Lymphocytic Leukemia and Myelodysplastic Syndromes: The MyPal Randomized Controlled Trial

    by Thomas Chatzikonstantinou, Maria Vasilopoulou, Dimitrios Kyrou, Georgios Karakatsoulis, Panagiotis Argyropaidas, Sotiria Besikli-Dimou, Panos Bonotis, Maria Chatzimina, Jana Didi, Christos Dimosthenous, et al.

    PURPOSE: The MyPal study (ClinicalTrials.gov identifier: NCT04370457 ) is a randomized controlled clinical trial assessing an eHealth intervention on the quality of life (QoL) of patients with chronic lymphocytic leukemia (CLL) and myelodysplastic syndromes (MDS). METHODS: Patients who were receiving or had previously received treatment for CLL or MDS were randomly assigned (1:1) to access the MyPal digital health platform versus standard of care. The MyPal platform included a smartphone application used to report QoL status and symptoms via standardized questionnaires or spontaneous reporting. The primary end point was QoL at 12 months, assessed by the European Organisation for Research and Treatment of Cancer (EORTC) QLQ-C30 General Questionnaire and the EuroQol EQ-5D-3L. Secondary end points included physical and emotional functioning, measured by the Integrated Palliative Care Outcome Scale (IPOS) scale, satisfaction with care, measured by the EORTC PATSAT-C33, and overall survival (OS). Additionally, the Edmonton Symptom Assessment System (ESAS), Brief Pain Inventory (BPI), and Emotion Thermometers (ET) QoL questionnaires were assessed only in the intervention group. RESULTS: A total of 171 patients (97 and 74 in the control and intervention arms, respectively) who answered multiple questionnaires were analyzed. The intervention group reported a significant decrease in pain (β 2 = –0.48 [–0.77 to –0.19], P < .001) compared with the control group (β 1 = 0.3 [0.09 to 0.5], P = .01). Communication and pain measured by IPOS reduced equally in both groups (β 2 = 0 [–0.03 to 0.02], P = .82; β 2 = –0.01 [–0.02 to 0], P = .1, respectively). Family involvement significantly increased over time only for the intervention group. The other items of EORTC QLQ-C30, EuroQol EQ-5D-3L, IPOS, and PATSAT-C33 remained unchanged in both groups. The intervention group displayed a significant improvement in all ESAS, BPI, and ET scales. OS was similar in both groups. CONCLUSION: The MyPal intervention improved several QoL aspects and led to a statistically significant decrease in pain compared with the control group.

  • Open Access

    Topic modeling and sentiment analysis of Greek clinician–patient conversations in hematologic malignancies

    by Maria Evangelia Chatzimina, Helen A. Papadaki, Charalampos Pontikoglou, Manolis Tsiknakis


  • Open Access

    A Greek Conversational Agent for Hematologic Malignancies: Usability and User Experience Assessment

    by Maria E. Chatzimina, Helen A. Papadaki, Charalampos Pontikoglou, Manolis Tsiknakis

    Enabling patients to actively document their health information significantly improves understanding of how therapies work, disease progression, and overall life quality affects for those living with chronic disorders such as hematologic malignancies. Advancements in artificial intelligence, particularly in areas such as natural language processing and speech recognition, have resulted in the development of interactive tools tailored for healthcare. This paper introduces an innovative conversational agent tailored to the Greek language. The design and deployment of this tool, which incorporates sentiment analysis, aims at gathering detailed family histories and symptom data from individuals diagnosed with hematologic malignancies. Furthermore, we discuss the preliminary findings from a feasibility study assessing the tool’s effectiveness. Initial feedback on the user experience suggests a positive reception towards the agent’s usability, highlighting its potential to enhance patient engagement in a clinical setting.


  • Open Access

    A Comparative Sentiment Analysis of Greek Clinical Conversations Using BERT, RoBERTa, GPT-2, and XLNet

    by Maria Evangelia Chatzimina, Helen A. Papadaki, Charalampos Pontikoglou, Manolis Tsiknakis

    In addressing the critical role of emotional context in patient–clinician conversations, this study conducted a comprehensive sentiment analysis using BERT, RoBERTa, GPT-2, and XLNet. Our dataset includes 185 h of Greek conversations focused on hematologic malignancies. The methodology involved data collection, data annotation, model training, and performance evaluation using metrics such as accuracy, precision, recall, F1-score, and specificity. BERT outperformed the other methods across all sentiment categories, demonstrating its effectiveness in capturing the emotional context in clinical interactions. RoBERTa showed a strong performance, particularly in identifying neutral sentiments. GPT-2 showed promising results in neutral sentiments but exhibited a lower precision and recall for negatives. XLNet showed a moderate performance, with variations across categories. Overall, our findings highlight the complexities of sentiment analysis in clinical contexts, especially in underrepresented languages like Greek. These insights highlight the potential of advanced deep-learning models in enhancing communication and patient care in healthcare settings. The integration of sentiment analysis in healthcare could provide insights into the emotional states of patients, resulting in more effective and empathetic patient support. Our study aims to address the gap and limitations of sentiment analysis in a Greek clinical context, an area where resources are scarce and its application remains underexplored.


  • Designing a conversational agent for patients with hematologic malignancies: Usability and Usefulness Study

    by Maria Chatzimina, Helen Papadaki, Charalampos Pontikoglou, Lefteris Koumakis, Kostas Marias, Manolis Tsiknakis

    Empowering patients to record health-related information can provide further insights on treatment response, disease evolution, disease burden, Quality of Life (QoL) and enables direct measurement of the experiences of patients with chronic conditions, including hematologic malignancies. Evidence suggests that using electronic tools for the collection of patient health outcomes improves symptom control and enhances patient satisfaction. In parallel, recent advancement in machine learning and speech recognition have led to conversational agents, software systems mimicking written or spoken human speech, being increasingly adopted in the health care domain. In the present manuscript we present (i) a methodology for the implementation of a conversational agent able to collect family history and symptom-related information of patients with hematologic malignancies and (ii) its initial evaluation results from a relevant feasibility study. Our approach uses deep learning algorithms trained in conversations focused on hematologic malignancies. The evaluation of the model from the user experience perspective provides promising results regarding the acceptability and comprehensibility of the system.


  • Employing Conversational Agents in Palliative Care: A Feasibility Study and Preliminary Assessment

    by Maria Chatzimina, Lefteris Koumakis, Kostas Marias, Manolis Tsiknakis

    Recording of patient-reported outcomes (PROs) enables direct measurement of the experiences of patients with chronic conditions, including cancer; thus, PROs are a critical element of high quality, person-centered care for cancer patients. A growing body of literature reports on the feasibility of using electronic tools for the collection of Patient reported Outcomes (ePROs), although the usability of available solutions does affect their acceptance and use. In parallel, recent advancement in artificial intelligence, machine learning and speech recognition have led to the growing interest in conversational agents, i.e. software applications that mimic written or spoken human speech. In the present manuscript we provide a review of current developments regarding the implementation of conversational agents and their application in the domain of palliative care for oncology patients and also present (i) a methodology for the implementation of a conversational agent able to collect ePRO health data and (ii) initial evaluation results from a relevant feasibility study. Our approach differs from other available systems since the conversational agent reported in the present work is not based on rules, but rather uses machine learning algorithms and more specifically recurrent neural networks (RNN) for identifying appropriate answers. Evaluation results of user experience provided promising results and highlight that users gave positive responds when interacting with the system. Based on the User Experience Questionnaire, pragmatic quality and overall quality were categorized as excellent and hedonic quality was categorized as good. The result of this research can be used as reference for the future development and improvement of the conversational agents in the healthcare domain.


  • Personal Health ICT Systems to Support Integrated Care Solutions

    by Dimitrios G Katehakis, Angelina Kouroubali, Ioannis Karatzanis, Dimitris Manousos, Haridimos Kondylakis, Georgios Kavlentakis, Nikos Stathiakis, Lefteris Koumakis, Maria Chatzimina, Kostas Marias


  • Open Access

    Personal health information recommender: implementing a tool for the empowerment of cancer patients

    by Galatia Iatraki, Haridimos Kondylakis, Lefteris Koumakis, Maria Chatzimina, Eleni Kazantzaki, Kostas Marias, Manolis Tsiknakis

    Ecancermedicalscience is the peer-reviewed open access cancer journal founded by the European Institute of Oncology in Milan. We consider articles on all aspects of research relating to cancer, including molecular biology, genetics, pathophysiology, epide