Hearing loss is a rising health concern. The World Health Organization estimates that over 1.5 billion people worldwide are currently experiencing hearing loss. This estimate is projected to rise to 2.5 billion by 2050. Hearing loss is often linked with other health conditions, such as dementia, exacerbating its impact. Chronic hearing loss often worsens over time. Early intervention and effective management are therefore important. One of the first intelligent agents designed to provide personalised advice to help preserve the hearing of individuals experiencing mild hearing loss is presented. The intelligent agent functions as a smartphone app that uses natural language processing to extract information from end-users with mild hearing loss and provides evidence-based personalised advice generated through generative artificial intelligence. The advice was assessed through a tailored criterion inspired by the two validated instruments DISCERN and PEMAT scores. Assessment was undertaken across three dimensions: Relevance (measuring how well the advice is personalised to the user), Accuracy (measuring whether the provided information is accurate), and Understandability (measuring whether the text is easily understandable without specialised knowledge). Measurement across 83 AI-generated responses resulted in the following scores: Relevance (85.5%), Accuracy (80.7%) and Understandability (89.2%).

Modern advances in computation enable the use of complex machine learning algorithms and artificial intelligence to assist human decision-making. However, the lack of explainability entailing from the black box nature of complex machine learning algorithms, inhibit their adoption in real-world applications especially in fields like healthcare. To address this challenge, we explored using the Odds Ratio (OR)—a clinically well-known measure of evidence—coupled with a sorting algorithm to prototype an explainable clinical decision support system (CDSS). This CDSS intakes relevant patient information such as demographic variables, clinical variables, medical history, and so on, and ranks treatment options personalised for patients, based on OR evidence. We present in this work-in-progress paper how our algorithm performs personalised ranking of therapies, taking Type-2 diabetes as a case study. As future work, we endeavour to codesign this further with clinicians to produce a primary care CDSS and assess long-term clinical outcomes.

BACKGROUND:The World Health Organisation's (WHO) iSupport Lite program contains six short practical public health support messages for carers of people with dementia to reduce stress and improve their mental health. We culturally adapted WHO's iSupport Lite into a digital intervention-'Draw-Care'-comprising six animated films and tipsheets hosted on a website and available in 10 languages. AIM:To assess the effectiveness of 'Draw-Care' on multilingual family carers' burden, mood and quality of life. METHODS:A single-blind, parallel-group randomised clinical trial conducted with multilingual adult carers with internet access. The trial was community-based, conducted online, between 1 August 2023 and 8 November 2024, with follow-up at 6- and 12-weeks. The target was 194 participants (156 needed with 25% attrition). The primary outcome was reduction in carer burden, as measured by the Zarit Burden Interview (ZBI). Secondary outcomes were improvement in carer mood and quality of life, as measured by the Centre for Epidemiological Studies Depression (CES-D), the World Health Organisation Quality of Life Scale (WHOQOL-Bref), Care-related Quality of Life instrument (CarerQoL-7D), and productivity and activity impairment as measured by the Work Productivity and Activity Impairment Questionnaire as adapted for caregiving (WPAI:CG). FINDINGS:One hundred sixteen participants consented and 93 (46 waitlist control, 47 intervention) were included in the analyses. Participants mean age was 54.80 (SD 13.33), 72 (77.4%) were women, and 29 (31.5%), 27 (29.4%) and 22 (23.9%) were from European, Vietnamese and Chinese backgrounds, respectively. No statistically significant differences were observed between the intervention and control groups on the ZBI, CES-D, WHOQOL-Bref and WPAI:CG. Null effects for some outcomes may be due to ceiling effects or insufficient power. A statistically significantly higher mean CarerQoL-VAS score was found for the intervention group (mean difference = 0.75, 95% CL: 0.25 to 1.24, P = .003) over 6- and 12-weeks, compared to the control. INTERPRETATION:'Draw-Care' showed no significant effect on carers' burden and mood but did show significant improvements in some carers' quality of life measures.

Hearing loss has become common in the modern world. The World Health Organization has estimated that hearing loss currently affects over 1.5 billion people worldwide. Projections suggest that by 2050, this number would reach 2.5 billion. Hearing loss has also been found to being linked to other chronic health conditions, dementia for example. There is a lot of diversity among the people affected by hearing loss. This is especially true given the large number of people who are being affected. Given this diversity, hearing care needs to be more precise and personalized. However, it is well known that current approaches to hearing care often lack this required precision and personalization. An enhanced approach to hearing care is suggested in this chapter to address this challenge. In collaboration with a senior clinician, we analyzed samples of audiogram data. Following this analysis, we designed three algorithms to rate different aspects of the hearing of an individual. These scores are then used to provide personalized self-care advice to preserve the hearing of individuals experiencing hearing loss. The underpinning data-driven personalization aligns our approach with the concept of digital twin. Our approach enables early intervention to preserve hearing through personalized self-care advice.

Digital Twins (DTs) are essentially virtual replicas of physical entities. DTs have evolved significantly over time. They have been applied in various fields. Very recently, their application in the field of healthcare is also being explored. In healthcare, the creation of DTs of patients is of interest. DTs of patients show potential to perform as clinical decision support tools to enhance precision and personalization of treatment. An essential part of this role of DTs involves deriving sense from health and medical data. Personalization in this context requires looking at data of a present patient and identifying records of cohorts of past patients that are closely matching the present patient. Such matching cohorts allow for decision support on more personalized grounds. This paper presents an attempt to achieve such personalization through cohort matching. As part of an ongoing study, we do this to assist with immunotherapy treatment planning for triple-negative breast cancer as a case study.

Background:This research study aimed to detect the vocal features immersed in empathic counselor speech using samples of calls to a mental health helpline service. Objective:This study aimed to produce an algorithm for the identification of empathy from these features, which could act as a training guide for counselors and conversational agents who need to transmit empathy in their vocals. Methods:Two annotators with a psychology background and English heritage provided empathy ratings for 57 calls involving female counselors, as well as multiple short call segments within each of these calls. These ratings were found to be well-correlated between the 2 raters in a sample of 6 common calls. Using vocal feature extraction from call segments and statistical variable selection methods, such as L1 penalized LASSO (Least Absolute Shrinkage and Selection Operator) and forward selection, a total of 14 significant vocal features were associated with empathic speech. Generalized additive mixed models (GAMM), binary logistics regression with splines, and random forest models were used to obtain an algorithm that differentiated between high- and low-empathy call segments. Results:The binary logistics regression model reported higher predictive accuracies of empathy (area under the curve [AUC]=0.617, 95% CI 0.613-0.622) compared to the GAMM (AUC=0.605, 95% CI 0.601-0.609) and the random forest model (AUC=0.600, 95% CI 0.595-0.604). This difference was statistically significant, as evidenced by the nonoverlapping 95% CIs obtained for AUC. The DeLong test further validated these results, showing a significant difference in the binary logistic model compared to the random forest (D=6.443, df=186283, P<.001) and GAMM (Z=5.846, P<.001). These findings confirm that the binary logistic regression model outperforms the other 2 models concerning predictive accuracy for empathy classification. Conclusions:This study suggests that the identification of empathy from vocal features alone is challenging, and further research involving multimodal models (eg, models incorporating facial expression, words used, and vocal features) are encouraged for detecting empathy in the future. This study has several limitations, including a relatively small sample of calls and only 2 empathy raters. Future research should focus on accommodating multiple raters with varied backgrounds to explore these effects on perceptions of empathy. Additionally, considering counselor vocals from larger, more heterogeneous populations, including mixed-gender samples, will allow an exploration of the factors influencing the level of empathy projected in counselor voices more generally.

Background:Empathy is a critical component of effective mental health care communication. Positive perceptions of empathy in conversational agents (CAs) operating in the health care domain are therefore needed to enhance the quality of care provided by these emerging technologies. However, research on how users perceive empathy in CAs is limited, particularly in voice-based prototypes. Objective:The objective of this study is to identify to what extent perceptions of empathy in CA prototypes correspond with the engineered empathy levels for these voice-based prototypes. In addition, as a secondary aim, this study investigates how the demographic characteristics of participants affect their perception of empathy in a mental health helpline service context. Methods:Swinburne University first-year psychology students (N=306) were presented with 9 CA prototypes engineered to portray low, medium, or high empathy levels, and their perceptions of empathy were collected via an electronic survey. Perceptions of empathy were rated using the Perceived Emotional Intelligence (PEI) Scale and the Raters' Scale (RS10). Results:Most participants were female (233/306, 76%) with a mean age of 30 (SD 10.69) years, while a majority (194/306, 63%) were of Australian and New Zealand background. A strong positive correlation between the PEI and RS10 ratings was observed (r=0.829, P<.001). The empathy ratings across the 3 engineered empathy levels showed significant differences when using both PEI (χ22=11.865, P=.003) and RS10 (χ22=19.737, P<.001) measures. A linear mixed model for PEI showed significantly higher ratings for high rather than low engineered empathy levels (t8=-2.34, P=.048). RS10 ratings were also significantly higher for high rather than low engineered empathy levels (t8=-2.45, P=.04). However, no significant differences were detected between the CAs with engineered medium-level empathy and the CAs with low or high engineered empathy levels. The linear mixed model for PEI showed significantly higher ratings for participants of the Asian and Other ethnic categories compared to the Oceanic category (t285=2.54, P=.01 and t286=2.25, P=.03 respectively). The RS10 ratings were also significantly higher for the Other category rather than for the Oceanic category (t284=2.24, P=.03). Women showed significantly higher RS10 ratings than men (t283=1.94, P=.05). Conclusions:Recognizing empathy levels in CA prototypes proved challenging, highlighting possible complexities involved with voice-based empathy detection. The perception of empathy may also be affected by different ethnic and gender-based factors. The study findings emphasize the importance of personalized communications by CAs, with expressions of empathy tailored to key demographic characteristics of users. Future studies in a similar context would benefit from the inclusion of participants who are end users of a mental health care service with more balanced gender and age distributions. Multimodal interactions could also be considered for CA prototype development.

Digital twins are essentially digital replicas of physical entities. Their usage is becoming more common across various industries, including healthcare. However, the implementation of digital twins in healthcare is uniquely challenging. This is partly because of the sensitive nature of health data and privacy concerns. These concerns limit health data accessibility and shareability. This paper attempts to address this challenge of health data sharing. We propose a novel approach that leverages federated learning, model sharing, and digital twin-assisted clinical decision making. Our approach ensures that health data are kept federated with healthcare providers. Healthcare providers train machine learning models on their own data. Then, instead of sharing the data, the trained models are shared. This is enabled via an arrangement like a private blockchain that is accessible to subscribed healthcare providers. This approach allows healthcare providers to access and use machine learning models for clinical decision support without compromising sensitive data about patients. Certain information about machine learning models will be shared. These include indicators such as the sample size on which a model has been trained on, validation metrics, and model accuracy. Such information assists other healthcare providers in selecting the most effective models. We demonstrate the efficacy of this approach through a case study on chronic disease management (e.g., cancer) using Liquid Neural Networks. Our results show how federated learning and model sharing can enhance clinical decision making and improve patient outcomes while ensuring the privacy of data.

Background Empathy is a critical component of effective mental health care communication. Positive perceptions of empathy in conversational agents (CAs) operating in the health care domain are therefore needed to enhance the quality of care provided by these emerging technologies. However, research on how users perceive empathy in CAs is limited, particularly in voice-based prototypes. Objective The objective of this study is to identify to what extent perceptions of empathy in CA prototypes correspond with the engineered empathy levels for these voice-based prototypes. In addition, as a secondary aim, this study investigates how the demographic characteristics of participants affect their perception of empathy in a mental health helpline service context. Methods Swinburne University first-year psychology students (N=306) were presented with 9 CA prototypes engineered to portray low, medium, or high empathy levels, and their perceptions of empathy were collected via an electronic survey. Perceptions of empathy were rated using the Perceived Emotional Intelligence (PEI) Scale and the Raters’ Scale (RS10). Results Most participants were female (233/306, 76%) with a mean age of 30 (SD 10.69) years, while a majority (194/306, 63%) were of Australian and New Zealand background. A strong positive correlation between the PEI and RS10 ratings was observed ( r =0.829, P <.001). The empathy ratings across the 3 engineered empathy levels showed significant differences when using both PEI ( χ 2 2 =11.865, P =.003) and RS10 ( χ 2 2 =19.737, P <.001) measures. A linear mixed model for PEI showed significantly higher ratings for high rather than low engineered empathy levels ( t 8 =−2.34, P =.048). RS10 ratings were also significantly higher for high rather than low engineered empathy levels ( t 8 =−2.45, P =.04). However, no significant differences were detected between the CAs with engineered medium-level empathy and the CAs with low or high engineered empathy levels. The linear mixed model for PEI showed significantly higher ratings for participants of the Asian and Other ethnic categories compared to the Oceanic category ( t 285 =2.54, P =.01 and t 286 =2.25, P =.03 respectively). The RS10 ratings were also significantly higher for the Other category rather than for the Oceanic category ( t 284 =2.24, P =.03). Women showed significantly higher RS10 ratings than men ( t 283 =1.94, P =.05). Conclusions Recognizing empathy levels in CA prototypes proved challenging, highlighting possible complexities involved with voice-based empathy detection. The perception of empathy may also be affected by different ethnic and gender-based factors. The study findings emphasize the importance of personalized communications by CAs, with expressions of empathy tailored to key demographic characteristics of users. Future studies in a similar context would benefit from the inclusion of participants who are end users of a mental health care service with more balanced gender and age distributions. Multimodal interactions could also be considered for CA prototype development.

People with malignancy of undefined primary origin (MUO) have a poor prognosis and may undergo a protracted diagnostic workup causing patient distress and high cancer related costs. Not having a primary diagnosis limits timely site-specific treatment and access to precision medicine. There is a need to improve the diagnostic process, and healthcare delivery and support for these patients. This trial aims to implement and evaluate an optimal model of care for people presenting with MUO to reduce time to diagnosis, improve patient experiences and reduce healthcare costs. This is a pragmatic stepped-wedge cluster randomised trial comparing a control phase of standard practice with an intervention phase. Patient inclusion criteria are: 1) age 18 years or older, 2) presenting with suspected metastatic malignancy without an obvious primary site on imaging, 3) clinically appropriate to undergo diagnostic work-up and 4) able to provide written or verbal consent. The intervention is a new model of care comprising four key components: standardised diagnostic workup, dedicated cancer care coordinators, virtual multidisciplinary meetings and a website resource for patients, carers and clinicians. The primary endpoint is the time to completion of minimum diagnostic workup. Secondary outcomes are whether the type of tumour is diagnosed, clinical trial participation, referral to palliative care, patient-reported physical, social and mental health, patient-reported understanding and uncertainty. Implementation outcomes include acceptability, feasibility, fidelity and adoption and health care use and costs. Intervention implementation will be supported using clinical leadership, education and reinforcement. Patients who consent to having their data collected will receive the model of care active at the site at the time of recruitment. Patients will complete a patient-reported outcomes questionnaire four months after study enrolment. A health economic analysis will be included. Across 15 hospitals, a total sample size of 240 is planned. There is a lack of intervention research for people presenting with MUO. The stepped-wedge design seeks to mitigate the potential challenge of enrolling people with a poor prognosis and high symptom burden in trials. This research will generate important evidence with scalability for future research at trial completion. ACTRN12622001504707

Prediabetes presents a critical window to prevent type 2 diabetes, a rising global health crisis, yet young adults often lack engaging preventive tools. This ongoing study aims to design and evaluate a web application to enhance health knowledge, engagement, and self-management for this at-risk group. This theoretical lens combines Design Science Research Methodology (DSRM), the theory of Task-Technology Fit (TTF), and the Unified Theory of Acceptance and Use of Technology (UTAUT). The proposed solution incorporates a unique combination of features learned through a previously conducted systematic literature review (SLR). Features include Machine Learning (ML)-based recommendations, educational modules, goal setting, gamification elements, and an artificial intelligence (AI)-incorporated chatbot. The proposed design to date is presented, in addition to the planned scenario-driven use cases to highlight the relevance of the proposed solution. A pilot study will assess usability, usefulness, satisfaction, and health knowledge via initial, midway, and final surveys mapped along with the design process. The data will be analysed via descriptive statistics and thematic analysis. This work-in-progress paper offers a streamlined, user-centred approach to designing and developing digital health interventions for prediabetes prevention while contributing insights for personalised digital health interventions.

Dementia is becoming a commonly prevalent chronic disease worldwide impacting communities that are culturally and ethnically diverse. Despite the burden on family carers (hereafter, carers) of people with dementia in these communities, research on how to support these communities is lacking. Research is limited, especially regarding the design and development of multilingual online resources that are culturally appropriate and usable for these target groups. In such a backdrop, this study aimed to co-design a digital health intervention named DrawCare, including a multilingual virtual helper with carers from nine linguistic groups across Australia. Six co-design workshops were conducted online. Convenience and snowball sampling were used for carer recruitment (n=21), and data were thematically analyzed. Participants desired a helper with aesthetic and user-friendly design and problem-framing prompts in language to navigate the website for supporting resources. This feedback was used to improve the virtual helper for user testing and a randomized control trial to evaluate the effectiveness of the DrawCare intervention.

Modern healthcare services have advanced greatly due to rapid improvements in technology. The next generation of advancements requires precise and personalised treatments, especially for chronic diseases. Computational means are an effective way to achieve this through intelligent decision support assisted by superior data collection and analytics. An emerging concept to facilitate this is digital twins (DTs)—digital replicas of physical entities. DTs have evolved over the years across various industries including aerospace, control engineering, manufacturing, design optimization, and more. DTs in healthcare though, have been explored only relatively recently. One of the most interesting questions lies in creating DTs of humans to model healthcare aspects to enable intelligent decision support. Working towards this quest, this paper attempts to answer the research question: How might precise and personalised treatments for chronic diseases be planned in real-time through explainable digital twins? We attempt to answer this question in the context of breast cancer.

The World Health Organization's iSupport for Dementia program provides an online platform for carer education, yet its suitability for culturally and linguistically diverse communities remains under explored. This study evaluated the cultural and linguistic appropriateness of the adapted Vietnamese iSupport program for carers in Australia and identified factors influencing the future implementation of an iSupport Virtual Assistant (iSupport VA). A qualitative descriptive study was conducted using five focus group discussions with 30 participants, including 18 family carers and 12 formal carers from Vietnamese communities in Australia. Thematic analysis, guided by a deductive-then-inductive approach, was applied to analyse the data. Discussions were conducted in Vietnamese, recorded, transcribed, translated, and systematically coded for recurring themes. The findings showed that participants emphasized the need for culturally sensitive language and visual representation in the adapted iSupport program, stressing the necessity for translations that align with context, incorporate relatable examples, and feature realistic video content. They expressed a strong preference for accessible multimedia formats, favouring video content with voice-over and interactive features over text-heavy materials, particularly for those with limited literacy. The importance of culturally tailored caregiving scenarios was highlighted, with a preference for real actors over animated characters to enhance emotional authenticity. Despite recognising the program's value in improving caregiving skills, carers cited time constraints, competing responsibilities, and digital literacy challenges as barriers to engagement, emphasising the need for a clear value proposition and targeted support mechanisms, including introductory tutorials and peer-based community interaction. Adapting iSupport to align with cultural and linguistic needs enhances its relevance and accessibility for Vietnamese carers in Australia. Refining translations, incorporating culturally familiar multimedia elements, and addressing usability concerns are crucial to optimising engagement and effectiveness.

Addressing the needs of ethnically diverse multilingual people can be challenging in environments that are non-native to them. The consequences of this issue become more significant in healthcare contexts. Insights from the DrawCare study-an Australian study that explores the effectiveness of a web-based intervention for multilingual family carers of people with dementia-are presented illustrating the enabling role of digital health.

Abstract Efforts to limit the impact of the coronavirus disease (COVID‐19) pandemic led to the implementation of public health measures and reallocation of health resources. To investigate trends in blood pressure (BP), hypertension and BMI in the Australian population during the COVID‐19 pandemic, data from publicly accessible health stations were analyzed. Average BP and BMI measured by the SiSU Health Station network in Australia in over 1.6 million health screenings were compared between the years 2018 and 2021. Additionally, paired trajectories for BP and BMI development before and during the COVID‐19 pandemic were calculated. Comparisons between pre‐COVID years and post‐COVID years of 2018 versus 2020, 2019 versus 2020, 2018 versus 2021, and 2019 versus 2021 showed increases in average adjusted systolic BP of 2.0, 1.7, 2.6, and 2.3 mmHg, respectively. Paired analysis of longitudinal data showed an overall increase in the trajectory of systolic BP of 3.2 mmHg between pre‐ and post‐COVID years. The prevalence of hypertension in users of the health stations increased by approximately 25% in the years 2020–2021. Similar trends were seen for BMI. Data from public Australian health stations indicated a strong trend toward higher BP during the COVID‐19 pandemic. At the population level, BP increments have been shown to markedly increase cardiovascular disease risk. Anti‐pandemic measures need to be carefully evaluated in terms of secondary public health effects and health support systems extended to effectively target cardiovascular risk.