Journal of Medical Internet Research最新文献

Background: Health systems are increasingly offering patient portals as tools for patients to access their health information with the goal of improving engagement in care. However, understanding health care providers' perspectives on patient portal implementation is crucial.

Objective: This study aimed to understand health care providers' experiences of implementing the MyChart patient portal, perspectives about its impact on patient care, clinical practice, and workload, and opportunities for improvement.

Methods: Using an explanatory sequential mixed methods approach, we conducted a web-based questionnaire and semistructured individual interviews with health care providers at a large Canadian community hospital, 6 months after MyChart was first offered to patients. We explored perspectives about the impact of MyChart on clinical practice, workload, and patient care. Data were analyzed using descriptive statistics and thematic analysis.

Results: In total, 261 health care providers completed the web-based questionnaire, and 15 also participated in interviews. Participants agreed that patients should have access to their health information through MyChart and identified its benefits such as patients gaining a greater understanding of their own health, which could improve patient safety (160/255, 62%). While many health care providers agreed that MyChart supported better patient care (108/258, 42%), there was limited understanding of features available to patients and expectations for integrating MyChart into clinical routines. Concerns were raised about the potential negative impacts of MyChart on patient-provider relationships because sensitive notes or results could be inappropriately interpreted (109/251, 43%), and a potential increase in workload if additional portal features were introduced. Suggested opportunities for improvement included support for both patients and health care providers to learn about MyChart and establishing guidelines for health care providers on how to communicate information available in MyChart to patients.

Conclusions: While health care providers acknowledged that MyChart improved patients' access to health information, its implementation introduced some friction and concerns. To reduce the risk of these challenges, health systems can benefit from engaging health care providers early to identify effective patient portal implementation strategies.

Background: Depression is a widespread mental health issue affecting older adults globally, with substantial implications for their well-being. Although digital interventions have proven effective in high-income countries, research on the potential of internet usage to alleviate depression among older adults in high-income countries remains limited.

Objective: This study aimed to examine the impact of internet usage on depression among older adults in high-income countries by developing a comprehensive theoretical framework and testing key hypotheses.

Methods: Using data from the China Health and Retirement Longitudinal Study (CHARLS), a 2-stage instrumental variable approach was applied to address endogeneity and estimate causal relationships between internet usage and depression.

Results: The findings indicate that internet usage results in a 1.41% reduction in depression levels among older adults. This effect is mediated by four primary mechanisms: (1) enhanced social interaction, (2) increased physical activity, (3) improved intergenerational contact, and (4) expanded access to educational opportunities. A heterogeneity analysis revealed that these effects are more pronounced in urban areas, eastern regions, and regions with superior internet infrastructure.

Conclusions: Internet usage plays a crucial role in alleviating depression among older adults in high-income countries, with regional variations. The findings highlight the need for targeted policy interventions to improve internet access and digital literacy, which can mitigate depression and enhance the mental health of older adults.

Health recommender systems (HRS) have the capability to improve human-centered care and prevention by personalizing content, such as health interventions or health information. HRS, an emerging and developing field, can play a unique role in the digital health field as they can offer relevant recommendations, not only based on what users themselves prefer and may be receptive to, but also using data about wider spheres of influence over human behavior, including peers, families, communities, and societies. We identify and discuss how HRS could play a unique role in decreasing health inequities. We use the socioecological model, which provides representations of how multiple, nested levels of influence (eg, community, institutional, and policy factors) interact to shape individual health. This perspective helps illustrate how HRS could address not just individual health factors but also the structural barriers-such as access to health care, social support, and access to healthy food-that shape health outcomes at various levels. Based on this analysis, we then discuss the challenges and future research priorities. We find that despite the potential for targeting more complex systemic challenges to obtaining good health, current HRS are still focused on individual health behaviors, often do not integrate the lived experiences of users in the design, and have had limited reach and effectiveness for individuals from low socioeconomic status and racial or ethnic minoritized backgrounds. In this viewpoint, we argue that a new design paradigm is necessary in which HRS focus on incorporating structural barriers to good health in addition to user preferences. HRS should be designed with an emphasis on health systems, which also includes incorporating decolonial perspectives of well-being that challenge prevailing medical models. Furthermore, potential lies in evaluating the health equity effects of HRS and leveraging collected data to influence policy. With changes in practices and with an intentional equity focus, HRS could play a crucial role in health promotion and decreasing health inequities.

Background: eHealth interventions can favorably impact health outcomes and encourage health-promoting behaviors in children. More insight is needed from the perspective of children and their families regarding eHealth interventions, including features influencing program effectiveness.

Objective: This review aimed to explore families' experiences with family-focused web-based interventions for improving health.

Methods: Five databases were searched on October 26, 2022-updated on October 24, 2023-for studies reporting qualitative data on participating children or their caregivers' experiences with web-based programs. Study identification was performed in duplicate and studies were independently appraised for quality. Thematic synthesis was undertaken on qualitative data extracted from the results section of each included article.

Results: Of 5524 articles identified, 28 articles were included. The studies examined the experiences of school-aged children (aged 5-18 years) and their caregivers (mostly mothers) with 26 web-based interventions that were developed to manage 17 different health conditions or influence health-supporting behaviors. Six themes were identified on families' experiences: connecting with others, agency of learning, program reputability or credibility, program flexibility, meeting participants' needs regarding program content or delivery, and impact on lifestyle.

Conclusions: Families positively perceived family-focused web-based interventions, finding value in quality connections and experiencing social support; intervention features aligned with behavioral and self-management principles. Key considerations were highlighted for program developers and health care professionals on ways to adapt eHealth elements to meet families' health-related needs. Continued research examining families' experiences with eHealth interventions is needed, including the experiences of families from diverse populations and distinguishing the perspectives of children, their caregivers, and other family members, to inform the expansion of family-focused eHealth interventions in health care systems.

Trial registration: PROSPERO CRD42022363874; https://tinyurl.com/3xxa8enz.

Background: Large language model (LLM) artificial intelligence chatbots using generative language can offer smoking cessation information and advice. However, little is known about the reliability of the information provided to users.

Objective: This study aims to examine whether 3 ChatGPT chatbots-the World Health Organization's Sarah, BeFreeGPT, and BasicGPT-provide reliable information on how to quit smoking.

Methods: A list of quit smoking queries was generated from frequent quit smoking searches on Google related to "how to quit smoking" (n=12). Each query was given to each chatbot, and responses were analyzed for their adherence to an index developed from the US Preventive Services Task Force public health guidelines for quitting smoking and counseling principles. Responses were independently coded by 2 reviewers, and differences were resolved by a third coder.

Results: Across chatbots and queries, on average, chatbot responses were rated as being adherent to 57.1% of the items on the adherence index. Sarah's adherence (72.2%) was significantly higher than BeFreeGPT (50%) and BasicGPT (47.8%; P<.001). The majority of chatbot responses had clear language (97.3%) and included a recommendation to seek out professional counseling (80.3%). About half of the responses included the recommendation to consider using nicotine replacement therapy (52.7%), the recommendation to seek out social support from friends and family (55.6%), and information on how to deal with cravings when quitting smoking (44.4%). The least common was information about considering the use of non-nicotine replacement therapy prescription drugs (14.1%). Finally, some types of misinformation were present in 22% of responses. Specific queries that were most challenging for the chatbots included queries on "how to quit smoking cold turkey," "...with vapes," "...with gummies," "...with a necklace," and "...with hypnosis." All chatbots showed resilience to adversarial attacks that were intended to derail the conversation.

Conclusions: LLM chatbots varied in their adherence to quit-smoking guidelines and counseling principles. While chatbots reliably provided some types of information, they omitted other types, as well as occasionally provided misinformation, especially for queries about less evidence-based methods of quitting. LLM chatbot instructions can be revised to compensate for these weaknesses.

Background: The implementation of large language models (LLMs), such as BART (Bidirectional and Auto-Regressive Transformers) and GPT-4, has revolutionized the extraction of insights from unstructured text. These advancements have expanded into health care, allowing analysis of social media for public health insights. However, the detection of drug discontinuation events (DDEs) remains underexplored. Identifying DDEs is crucial for understanding medication adherence and patient outcomes.

Objective: The aim of this study is to provide a flexible framework for investigating various clinical research questions in data-sparse environments. We provide an example of the utility of this framework by identifying DDEs and their root causes in an open-source web-based forum, MedHelp, and by releasing the first open-source DDE datasets to aid further research in this domain.

Methods: We used several LLMs, including GPT-4 Turbo, GPT-4o, DeBERTa (Decoding-Enhanced Bidirectional Encoder Representations from Transformer with Disentangled Attention), and BART, among others, to detect and determine the root causes of DDEs in user comments posted on MedHelp. Our study design included the use of zero-shot classification, which allows these models to make predictions without task-specific training. We split user comments into sentences and applied different classification strategies to assess the performance of these models in identifying DDEs and their root causes.

Results: Among the selected models, GPT-4o performed the best at determining the root causes of DDEs, predicting only 12.9% of root causes incorrectly (hamming loss). Among the open-source models tested, BART demonstrated the best performance in detecting DDEs, achieving an F₁-score of 0.86, a false positive rate of 2.8%, and a false negative rate of 6.5%, all without any fine-tuning. The dataset included 10.7% (107/1000) DDEs, emphasizing the models' robustness in an imbalanced data context.

Conclusions: This study demonstrated the effectiveness of open- and closed-source LLMs, such as GPT-4o and BART, for detecting DDEs and their root causes from publicly accessible data through zero-shot classification. The robust and scalable framework we propose can aid researchers in addressing data-sparse clinical research questions. The launch of open-access DDE datasets has the potential to stimulate further research and novel discoveries in this field.