The patient concerns inventory (PCI) allows patients to highlight the issues they would like to discuss at their outpatient consultation. It improves patient-clinician communication and has proven benefits. While the PCI is effective, patient experiences could be improved with better access to it and the ability to more easily and frequently express their concerns. This, of course, is in the context of ever-increasing healthcare challenges and limited resources. Use of conversational artificial intelligence (CAI) represents an opportunity to improve information flow between patients and professionals remote from the consultation. This paper highlights the potential for CAI to provide an \'always-on\' platform, using natural language interface technology and based on the PCI, which patients can access via their mobile devices. We also discuss potential pitfalls and concerns, along with outlining a current clinical trial assessing, in the first instance, usability of this technology.

UNASSIGNED: Worldwide, millions suffer from cataracts, which impair vision and quality of life. Cataract education improves outcomes, satisfaction, and treatment adherence. Lack of health literacy, language and cultural barriers, personal preferences, and limited resources may all impede effective communication.
UNASSIGNED: AI can improve patient education by providing personalised, interactive, and accessible information tailored to patient understanding, interest, and motivation. AI chatbots can have human-like conversations and give advice on numerous topics.
UNASSIGNED: This study investigated the efficacy of chatbots in cataract patient education relative to traditional resources like the AAO website, focusing on information accuracy,understandability, actionability, and readability. A descriptive comparative design was used to analyse quantitative data from frequently asked questions about cataracts answered by ChatGPT, Bard, Bing AI, and the AAO website. SOLO taxonomy, PEMAT, and the Flesch-Kincaid ease score were used to collect and analyse the data.
UNASSIGNED: Chatbots scored higher than AAO website on cataract-related questions in terms of accuracy (mean SOLO score ChatGPT: 3.1 ± 0.31, Bard: 2.9 ± 0.72, Bing AI: 2.65 ± 0.49, AAO website: 2.4 ± 0.6, (p < 0.001)). For understandability (mean PEMAT-U score AAO website: 0,89 ± 0,04, ChatGPT 0,84 ± 0,02, Bard: 0,84 ± 0,02, Bing AI: 0,81 ± 0,02, (p < 0.001)), and actionability (mean PEMAT-A score ChatGPT: 0.86 ± 0.03, Bard: 0.85 ± 0.06, Bing AI: 0.81 ± 0.05, AAO website: 0.81 ± 0.06, (p < 0.001)) AAO website scored better than chatbots. Flesch-Kincaid readability ease analysis showed that Bard (55,5 ± 8,48) had the highest mean score, followed by AAO website (51,96 ± 12,46), Bing AI (41,77 ± 9,53), and ChatGPT (34,38 ± 9,75, (p < 0.001)).
UNASSIGNED: Chatbots have the potential to provide more detailed and accurate data than the AAO website. On the other hand, the AAO website has the advantage of providing information that is more understandable and practical. When patient preferences are not taken into account, generalised or biased information can decrease reliability.

BACKGROUND: Home blood pressure (BP) monitoring is recommended for people with hypertension; however, meta-analyses have demonstrated that BP improvements are related to additional coaching support in combination with self-monitoring, with little or no effect of self-monitoring alone. High-contact coaching requires substantial resources and may be difficult to deliver via human coaching models.
OBJECTIVE: This observational study assessed changes in BP and body weight following participation in a fully digital program called Lark Hypertension Care with coaching powered by artificial intelligence (AI).
METHODS: Participants (N=864) had a baseline systolic BP (SBP) ≥120 mm Hg, provided their baseline body weight, and had reached at least their third month in the program. The primary outcome was the change in SBP at 3 and 6 months, with secondary outcomes of change in body weight and associations of changes in SBP and body weight with participant demographics, characteristics, and program engagement.
RESULTS: By month 3, there was a significant drop of -5.4 mm Hg (95% CI -6.5 to -4.3; P<.001) in mean SBP from baseline. BP did not change significantly (ie, the SBP drop maintained) from 3 to 6 months for participants who provided readings at both time points (P=.49). Half of the participants achieved a clinically meaningful drop of ≥5 mm Hg by month 3 (178/349, 51.0%) and month 6 (98/199, 49.2%). The magnitude of the drop depended on starting SBP. Participants classified as hypertension stage 2 had the largest mean drop in SBP of -12.4 mm Hg (SE 1.2 mm Hg) by month 3 and -13.0 mm Hg (SE 1.6 mm Hg) by month 6; participants classified as hypertension stage 1 lowered by -5.2 mm Hg (SE 0.8) mm Hg by month 3 and -7.3 mm Hg (SE 1.3 mm Hg) by month 6; participants classified as elevated lowered by -1.1 mm Hg (SE 0.7 mm Hg) by month 3 but did not drop by month 6. Starting SBP (β=.11; P<.001), percent weight change (β=-.36; P=.02), and initial BMI (β=-.56; P<.001) were significantly associated with the likelihood of lowering SBP ≥5 mm Hg by month 3. Percent weight change acted as a mediator of the relationship between program engagement and drop in SBP. The bootstrapped unstandardized indirect effect was -0.0024 (95% CI -0.0052 to 0; P=.002).
CONCLUSIONS: A hypertension care program with coaching powered by AI was associated with a clinically meaningful reduction in SBP following 3 and 6 months of program participation. Percent weight change was significantly associated with the likelihood of achieving a ≥5 mm Hg drop in SBP. An AI-powered solution may offer a scalable approach to helping individuals with hypertension achieve clinically meaningful reductions in their BP and associated risk of cardiovascular disease and other serious adverse outcomes via healthy lifestyle changes such as weight loss.

BACKGROUND: While mental health applications are increasingly becoming available for large populations of users, there is a lack of controlled trials on the impacts of such applications. Artificial intelligence (AI)-empowered agents have been evaluated when assisting adults with cognitive impairments; however, few applications are available for aging adults who are still actively working. These adults often have high stress levels related to changes in their work places, and related symptoms eventually affect their quality of life.
OBJECTIVE: We aimed to evaluate the contribution of TEO (Therapy Empowerment Opportunity), a mobile personal health care agent with conversational AI. TEO promotes mental health and well-being by engaging patients in conversations to recollect the details of events that increased their anxiety and by providing therapeutic exercises and suggestions.
METHODS: The study was based on a protocolized intervention for stress and anxiety management. Participants with stress symptoms and mild-to-moderate anxiety received an 8-week cognitive behavioral therapy (CBT) intervention delivered remotely. A group of participants also interacted with the agent TEO. The participants were active workers aged over 55 years. The experimental groups were as follows: group 1, traditional therapy; group 2, traditional therapy and mobile health (mHealth) agent; group 3, mHealth agent; and group 4, no treatment (assigned to a waiting list). Symptoms related to stress (anxiety, physical disease, and depression) were assessed prior to treatment (T1), at the end (T2), and 3 months after treatment (T3), using standardized psychological questionnaires. Moreover, the Patient Health Questionnaire-8 and General Anxiety Disorders-7 scales were administered before the intervention (T1), at mid-term (T2), at the end of the intervention (T3), and after 3 months (T4). At the end of the intervention, participants in groups 1, 2, and 3 filled in a satisfaction questionnaire.
RESULTS: Despite randomization, statistically significant differences between groups were present at T1. Group 4 showed lower levels of anxiety and depression compared with group 1, and lower levels of stress compared with group 2. Comparisons between groups at T2 and T3 did not show significant differences in outcomes. Analyses conducted within groups showed significant differences between times in group 2, with greater improvements in the levels of stress and scores related to overall well-being. A general worsening trend between T2 and T3 was detected in all groups, with a significant increase in stress levels in group 2. Group 2 reported higher levels of perceived usefulness and satisfaction.
CONCLUSIONS: No statistically significant differences could be observed between participants who used the mHealth app alone or within the traditional CBT setting. However, the results indicated significant differences within the groups that received treatment and a stable tendency toward improvement, which was limited to individual perceptions of stress-related symptoms.
BACKGROUND: ClinicalTrials.gov NCT04809090; https://clinicaltrials.gov/ct2/show/NCT04809090.

OBJECTIVE: The purpose of our study is to develop a spoken dialogue system (SDS) for pain questionnaire in patients with spinal disease. We evaluate user satisfaction and validated the performance accuracy of the SDS in medical staff and patients.
METHODS: The SDS was developed to investigate pain and related psychological issues in patients with spinal diseases based on the pain questionnaire protocol. We recognized patients\' various answers, summarized important information, and documented them. User satisfaction and performance accuracy were evaluated in 30 potential users of SDS, including doctors, nurses, and patients and statistically analyzed.
RESULTS: The overall satisfaction score of 30 patients was 5.5 ± 1.4 out of 7 points. Satisfaction scores were 5.3 ± 0.8 for doctors, 6.0 ± 0.6 for nurses, and 5.3 ± 0.5 for patients. In terms of performance accuracy, the number of repetitions of the same question was 13, 16, and 33 (13.5%, 16.8%, and 34.7%) for doctors, nurses, and patients, respectively. The number of errors in the summarized comment by the SDS was 5, 0, and 11 (5.2%, 0.0%, and 11.6 %), respectively. The number of summarization omissions was 7, 5, and 7 (7.3%, 5.3%, and 7.4%), respectively.
CONCLUSIONS: This is the first study in which voice-based conversational artificial intelligence (AI) was developed for a spinal pain questionnaire and validated by medical staff and patients. The conversational AI showed favorable results in terms of user satisfaction and performance accuracy. Conversational AI can be useful for the diagnosis and remote monitoring of various patients as well as for pain questionnaires in the future.

BACKGROUND: Conversational agents have the ability to reach people through multiple mediums, including the online space, mobile phones, and hardware devices like Alexa and Google Home. Conversational agents provide an engaging method of interaction while making information easier to access. Their emergence into areas related to public health and health education is perhaps unsurprising. While the building of conversational agents is getting more simplified with time, there are still requirements of time and effort. There is also a lack of clarity and consistent terminology regarding what constitutes a conversational agent, how these agents are developed, and the kinds of resources that are needed to develop and sustain them. This lack of clarity creates a daunting task for those seeking to build conversational agents for health education initiatives.
OBJECTIVE: This scoping review aims to identify literature that reports on the design and implementation of conversational agents to promote and educate the public on matters related to health. We will categorize conversational agents in health education in alignment with current classifications and terminology emerging from the marketplace. We will clearly define the variety levels of conversational agents, categorize currently existing agents within these levels, and describe the development models, tools, and resources being used to build conversational agents for health care education purposes.
METHODS: This scoping review will be conducted by employing the Arksey and O\'Malley framework. We will also be adhering to the enhancements and updates proposed by Levac et al and Peters et al. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) extension for scoping reviews will guide the reporting of this scoping review. A systematic search for published and grey literature will be undertaken from the following databases: (1) PubMed, (2) PsychINFO, (3) Embase, (4) Web of Science, (5) SCOPUS, (6) CINAHL, (7) ERIC, (8) MEDLINE, and (9) Google Scholar. Data charting will be done using a structured format.
RESULTS: Initial searches of the databases retrieved 1305 results. The results will be presented in the final scoping review in a narrative and illustrative manner.
CONCLUSIONS: This scoping review will report on conversational agents being used in health education today, and will include categorization of the levels of the agents and report on the kinds of tools, resources, and design and development methods used.
UNASSIGNED: DERR1-10.2196/31923.

While earlier research in human-robot interaction pre-dominantly uses rule-based architectures for natural language interaction, these approaches are not flexible enough for long-term interactions in the real world due to the large variation in user utterances. In contrast, data-driven approaches map the user input to the agent output directly, hence, provide more flexibility with these variations without requiring any set of rules. However, data-driven approaches are generally applied to single dialogue exchanges with a user and do not build up a memory over long-term conversation with different users, whereas long-term interactions require remembering users and their preferences incrementally and continuously and recalling previous interactions with users to adapt and personalise the interactions, known as the lifelong learning problem. In addition, it is desirable to learn user preferences from a few samples of interactions (i.e., few-shot learning). These are known to be challenging problems in machine learning, while they are trivial for rule-based approaches, creating a trade-off between flexibility and robustness. Correspondingly, in this work, we present the text-based Barista Datasets generated to evaluate the potential of data-driven approaches in generic and personalised long-term human-robot interactions with simulated real-world problems, such as recognition errors, incorrect recalls and changes to the user preferences. Based on these datasets, we explore the performance and the underlying inaccuracies of the state-of-the-art data-driven dialogue models that are strong baselines in other domains of personalisation in single interactions, namely Supervised Embeddings, Sequence-to-Sequence, End-to-End Memory Network, Key-Value Memory Network, and Generative Profile Memory Network. The experiments show that while data-driven approaches are suitable for generic task-oriented dialogue and real-time interactions, no model performs sufficiently well to be deployed in personalised long-term interactions in the real world, because of their inability to learn and use new identities, and their poor performance in recalling user-related data.