Explore the vast range of titles available.

[This corrects the article DOI: 10.1371/journal.pone.0305496.].

Narratives posted on the internet by patients contain a vast amount of information about various concerns. This study aimed to extract multiple concerns from interviews with breast cancer patients using the natural language processing (NLP) model bidirectional encoder representations from transformers (BERT). A total of 508 interview transcriptions of breast cancer patients written in Japanese were labeled with five types of concern labels: \"treatment,\" \"physical,\" \"psychological,\" \"work/financial,\" and \"family/friends.\" The labeled texts were used to create a multi-label classifier by fine-tuning a pre-trained BERT model. Prior to fine-tuning, we also created several classifiers with domain adaptation using (1) breast cancer patients’ blog articles and (2) breast cancer patients’ interview transcriptions. The performance of the classifiers was evaluated in terms of precision through 5-fold cross-validation. The multi-label classifiers with only fine-tuning had precision values of over 0.80 for \"physical\" and \"work/financial\" out of the five concerns. On the other hand, precision for \"treatment\" was low at approximately 0.25. However, for the classifiers using domain adaptation, the precision of this label took a range of 0.40–0.51, with some cases improving by more than 0.2. This study showed combining domain adaptation with a multi-label classifier on target data made it possible to efficiently extract multiple concerns from interviews.

Wearable devices (WDs) have evolved beyond simple fitness trackers to sophisticated health monitors capable of measuring vital signs, such as heart rate and blood oxygen levels. Their application in health care, particularly medication management, is an emerging field poised to significantly enhance patient adherence to treatment regimens. Despite their widespread use and increasing incorporation into clinical trials, a comprehensive review of WDs in terms of medication adherence has not been conducted. This study aimed to conduct a comprehensive scoping review to evaluate the impact of WDs on medication adherence across a variety of diseases, summarizing key research findings, outcomes, and challenges encountered. Adhering to PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) guidelines, a structured search was conducted across MEDLINE, Web of Science, and Embase databases, covering the literature from January 1, 2010, to September 30, 2022. The search strategy was based on terms related to WDs and medication adherence, specifically focusing on empirical studies to ensure the inclusion of original research findings. Studies were selected based on their relevance to medication adherence, usage of WDs in detecting medication-taking actions, and their role in integrated medication management systems. We screened 657 articles and identified 18 articles. The identified studies demonstrated the diverse applications of WDs in enhancing medication adherence across diseases such as Parkinson disease, diabetes, and cardiovascular conditions. The geographical distribution and publication years of these studies indicate a growing interest in this research area. The studies were divided into three types: (1) studies reporting a correlation between data from WDs or their usage and medication adherence or drug usage as outcomes, (2) studies using WDs to detect the act of medication-taking itself, and (3) studies proposing an integrated medication management system that uses WDs in managing medication. WDs are increasingly being recognized for their potential to enhance medication management and adherence. This review underscores the need for further empirical research to validate the effectiveness of WDs in real-life settings and explore their use in predicting adherence based on activity rhythms and activities. Despite technological advancements, challenges remain regarding the integration of WDs into routine clinical practice. Future research should focus on leveraging the comprehensive data provided by WDs to develop personalized medication management strategies that can improve patient outcomes.

Declining medication adherence remains a critical health care issue, often assessed through unreliable self-reporting methods. Wearable devices (WDs) may offer an objective means to improve adherence monitoring by continuously recording physiological and activity data. This study aimed to develop and internally validate personalized predictive models, utilizing objective physiological and activity data from WDs, for identifying missed medication doses. A 30-day prospective observational study was conducted with 8 participants who wore Apple Watches and used a dedicated iOS app. The app collected demographics, medication details, psychological factors, mealtimes, and daily missed dose events. WDs recorded time-series data (ie, activity, heart rate, sleep) at 3-minute intervals. Data were aggregated into 1-hour segments, and lag (6 and 12 h) as well as rolling (24 h) features were generated. Light Gradient Boosting Machine models were constructed for each individual's dosing regimen if the missed dose rate exceeded 20%. Two modeling approaches were compared: a group cross-validation (CV) model that grouped data by day to avoid data leakage from rolling features, and a nonrolling feature model that excluded rolling features and used leave-one-out CV. F1-score, accuracy, recall, and precision were assessed between the 2 models. Of the 15 enrolled participants, 8 completed the study; 4 had a missed dose rate above 20%. In these 4 individuals, the group CV model achieved F1-scores of 0.435 to 0.902, with accuracy ranging from 0.711 to 0.911, recall from 0.278 to 0.822, and a precision of 1.000 for the most robust regimens. The nonrolling feature model yielded F1-scores of 0.667 to 0.910, with accuracy ranging from 0.800 to 0.906, recall from 0.500 to 0.835, and a precision of 1.000. Morning dosing regimens generally showed higher predictive performance than evening or afternoon. Time-series features, particularly those reflecting 6-, 12-, and 24-hour patterns, emerged as key predictors, indicating that physiological and lifestyle variations prior to dosing strongly influenced missed dose events. Personalized predictive models using WD-derived data demonstrated high precision for detecting missed medication doses, especially in morning and evening regimens. These findings underscore the feasibility of employing continuous, objective physiological and activity data from WDs to forecast nonadherence events. Although the sample size was limited, restricting the generalizability of the results, this study demonstrates the potential of WD-based personalized prediction of medication adherence. Future work should involve larger populations for external validation, strategies to improve recall, especially for clinically critical medications, and careful consideration of real-world implementation challenges.

Background Patients with a history of hepatitis B virus (HBV) infection who are receiving immunosuppressive therapy are at risk of HBV reactivation and disease. Therefore, HBV screening is required prior to administering antirheumatic drugs with immunosuppressive effects. This study aimed to determine the status of hepatitis B surface antigen (HBsAg), hepatitis B core antibody (HBcAb), and hepatitis B surface antibody (HBsAb) screening prior to the initiation of drug therapy, including new antirheumatic drugs, in patients with rheumatoid arthritis. Methods This retrospective cross-sectional study used data from April 2014 to August 2022 from the Japanese hospital-based administrative claims database. The inclusion criteria were rheumatoid arthritis and first prescription date of antirheumatic drugs. Results A total of 82,282 patients with rheumatoid arthritis who were first prescribed antirheumatic drugs between April 2016 and August 2022 were included. Of the eligible patients, 9.7% ( n =7,959) were screened for all HBV (HBsAg, HBsAb, and HbcAb) within 12 months prior to the date of initial prescription. The HBsAg test was performed in 30.0% ( n =24,700), HBsAb test in 11.8% ( n =9,717), and HBcAb test in 13.1% ( n =10,824) of patients. The proportion of patients screened for HBV infection has been increasing since 2018; however, the proportion of patients screened for rheumatoid arthritis remains low. Conclusions Our findings suggest that HBV screening may be insufficient in patients who received antirheumatic drugs. With the increasing use of new immunosuppressive antirheumatic drugs, including biological agents, healthcare providers should understand the risk of HBV reactivation and conduct appropriate screening.

Effective symptom identification, a key responsibility for community pharmacists, requires patients to describe their symptoms accurately and comprehensively. However, current practices in pharmacies may be insufficient in capturing patient-reported symptoms comprehensively, potentially affecting the quality of pharmaceutical care and patient safety. This study aimed to construct a new, hierarchical symptom list derived from the Patient-Friendly Term List of the Medical Dictionary for Regulatory Activities (MedDRA) and to develop and evaluate a mobile app incorporating this list for facilitating symptom reporting by patients in pharmacy settings. The study also aimed to assess the usability and acceptance of this app among potential users. Subjective symptom-related terms were extracted from the Patient-Friendly Term List version 23.0 of the MedDRA. These terms were systematically consolidated and organized into a hierarchical, user-friendly symptom list. A mobile app incorporating this list was developed for pharmacy settings, featuring a symptom selection interface and a free-text input field for additional symptoms. The app included an instructional video explaining the importance of symptom reporting and guidance on navigation. Usability tests and semistructured interviews were conducted with participants aged >20 years. Interview transcripts were analyzed using the Unified Theory of Acceptance and Use of Technology (UTAUT) model to evaluate factors influencing the acceptance of technology. From the initial 1440 terms in the Patient-Friendly Term List, 795 relevant terms were selected and organized into 40 site-specific subcategories, which were then grouped into broader site categories (mental, head, trunk, upper limb, lower limb, physical condition, and others). These terms were further consolidated into 211 patient-friendly symptom terms, forming a hierarchical symptom list. The app's interface design limited options to 10 items per screen to assist with decision-making. A total of 5 adults participated in the usability test. Participants found the interface intuitive and easy to use, requiring minimal effort, and provided positive feedback regarding the potential utility of the app in pharmacy settings. The UTAUT analysis identified several facilitating factors, including ease of use and the potential for enhanced pharmacist-patient communication. However, concerns were raised about usability for older adults and the need for simplified technical terminology. The user-friendly app with a hierarchically structured symptom list and complementary free-text entry has potential benefits for improving the accuracy and efficiency of symptom reporting in pharmacy settings. The positive user acceptance and identified areas for improvement provide a foundation for further development and implementation of this technology to enhance communication between patients and pharmacists. Future improvements should focus on addressing usability for older adults and simplifying technical terminology.

Medication compliance, which refers to the extent to which patients correctly adhere to prescribed regimens, is influenced by various psychological, behavioral, and demographic factors. When analyzing these factors, challenges such as multicollinearity and variable selection often arise, complicating the interpretation of results. To address the issue of multicollinearity and better analyze the importance of each factor, machine learning methods are considered to be useful. This study aimed to identify key factors influencing medication compliance by applying regularized logistic regression and LightGBM. A questionnaire survey was conducted among 638 adult patients in Japan who had been continuously taking medications for at least 3 months. The survey collected data on demographics, medication habits, psychological adherence factors, and compliance. Logistic regression with regularization was used to handle multicollinearity, while LightGBM was used to calculate feature importance. The regularized logistic regression model identified significant predictors, including \"using the drug at approximately the same time each day\" (coefficient 0.479; P=.02), \"taking meals at approximately the same time each day\" (coefficient 0.407; P=.02), and \"I would like to have my medication reduced\" (coefficient -0.410; P=.01). The top 5 variables with the highest feature importance scores in the LightGBM results were \"Age\" (feature importance 179.1), \"Using the drug at approximately the same time each day\" (feature importance 148.4), \"Taking meals at approximately the same time each day\" (feature importance 109.0), \"I would like to have my medication reduced\" (feature importance 77.48), and \"I think I want to take my medicine\" (feature importance 70.85). Additionally, the feature importance scores for the groups of medication adherence-related factors were 77.92 for lifestyle-related items, 52.04 for awareness of medication, 20.30 for relationships with health care professionals, and 5.05 for others. The most significant factors for medication compliance were the consistency of medication and meal timing (mean of feature importance), followed by the number of medications and patient attitudes toward their treatment. This study is the first to use a machine learning model to calculate and compare the relative importance of factors affecting medication adherence. Our findings demonstrate that, in terms of relative importance, lifestyle habits are the most significant contributors to medication compliance among the general patient population. The findings suggest that regularization and machine learning methods, such as LightGBM, are useful for better understanding the numerous adherence factors affected by multicollinearity.

Patient-reported symptoms provide valuable insights into patient experiences and can enhance healthcare quality; however, effectively capturing them remains challenging. Although natural language processing (NLP) models have been developed to extract adverse events and symptoms from medical records written by healthcare professionals, limited studies have focused on models designed for patient-generated narratives. This study developed an NLP model to extract patient-reported symptoms from pharmaceutical care records and validated its effectiveness in analyzing diverse patient-generated narratives. The target dataset comprised “Subjective” sections of pharmaceutical care records created by community pharmacists for patients prescribed anticancer drugs. Two annotation guidelines were applied to develop robust ground-truth data, which was used to develop and evaluate a new transformer-based named entity recognition model. Model performance was compared with that of an existing tool for Japanese clinical texts and tested on external patient-generated blog data to evaluate generalizability. The newly developed BERT-CRF model significantly outperformed the existing model, achieving an F1 score > 0.8 on pharmaceutical care records and extracting > 98% of physical symptom entries from patient blogs, with a 20% improvement over the existing tool. These findings highlight the importance of fine-tuning models using patient-specific narrative data to capture nuanced and colloquial symptom expressions.

Tablets and capsules are widely used forms of oral medication, but some patients experience difficulty handling them, which can reduce medication adherence and affect health outcomes. This study aimed to identify factors contributing to perceived handling difficulty, using data from harmo ® , a nationwide electronic medication notebook system. A questionnaire was distributed to adult users who had been prescribed oral medications, and the responses were linked with personal health records to analyze medication characteristics and patient backgrounds. Among the 1,230 respondents, 24% reported difficulty with small tablets or capsules. A size threshold was identified: a combined long and short diameter of 13.3 mm or less was most associated with handling problems (ROC-AUC = 0.834). Binomial logistic regression analysis revealed that difficulty in applying force with the hands (OR = 2.64), prescription of small tablets or capsules (OR = 2.52), and medical histories of hypertension (OR = 1.69) and osteoporosis (OR = 4.99) were significantly associated with reported difficulty. These results suggest that both the physical characteristics of formulations and individual patient factors influence medication usability. Our results provide evidence to inform more patient-centered approaches to oral formulation design and prescribing practices, ultimately supporting better adherence and medication safety.

Early detection and management of adverse drug reactions (ADRs) is crucial for improving patients’ quality of life. Hand-foot syndrome (HFS) is one of the most problematic ADRs for cancer patients. Recently, an increasing number of patients post their daily experiences to internet community, for example in blogs, where potential ADR signals not captured through routine clinic visits can be described. Therefore, this study aimed to identify patients with potential ADRs, focusing on HFS, from internet blogs by using natural language processing (NLP) deep-learning methods. From 10,646 blog posts, written in Japanese by cancer patients, 149 HFS-positive sentences were extracted after pre-processing, annotation and scrutiny by a certified oncology pharmacist. The HFS-positive sentences described not only HFS typical expressions like “pain\" or “spoon nail”, but also patient-derived unique expressions like onomatopoeic ones. The dataset was divided at a 4 to 1 ratio and used to train and evaluate three NLP deep-learning models: long short-term memory (LSTM), bidirectional LSTM and bidirectional encoder representations from transformers (BERT). The BERT model gave the best performance with precision 0.63, recall 0.82 and f 1 score 0.71 in the HFS user identification task. Our results demonstrate that this NLP deep-learning model can successfully identify patients with potential HFS from blog posts, where patients’ real wordings on symptoms or impacts on their daily lives are described. Thus, it should be feasible to utilize patient-generated text data to improve ADR management for individual patients.