Hong Choon Oh

Background: Type 2 diabetes mellitus (T2DM) is a chronic condition that requires ongoing self-management. While chatbots offer scalable monitoring and management, evidence on their integration into structured hospital tele-support workflows, where patient engagement and digital readiness may differ from those in standalone digital interventions, remains limited. Objective: This study aimed to describe the design of a clinical protocol–guided chatbot, Sharing Information with Users on Diabetes Actions and Interventions ( Siu Dai ), and evaluate its feasibility, usability, and acceptability among patients and telecarers within a hospital tele-support program. Methods: We used an embedded mixed methods design (QUAL [quan]) guided by the World Health Organization (WHO) monitoring and evaluating digital health interventions and the theoretical framework of acceptability. Through a co-design process involving clinicians, researchers, telecarers, and developers, the chatbot was aligned with established clinical protocols. It featured 7 precall sessions designed to collect patient data, generate summary reports for staff, and deliver targeted educational content. Patients (≥21-year-old individuals with T2DM) and telecarer nurses (with experience ≥6 mo) were recruited from an existing tele-support program. Data sources included interviews, online surveys (5-point Likert scale), and chatbot usage logs. Qualitative data were analyzed thematically; quantitative data were summarized descriptively to supplement and contextualize the qualitative findings. Results: Fourteen participants (7 patients and 7 telecarers) completed both interviews and surveys and were included in the analysis. Patients (mean age 45.3, SD 15.4 y) had a median diabetes duration of 10 (IQR 3-25; range 1‐30) years, and all were digitally literate. Telecarers had a median of 5 (IQR 4-9) years of telehealth experience. Quantitatively, patients showed high engagement, completing 79% of started sessions with a median completion time of 2 minutes and 58 seconds. Patient satisfaction was high, with 86% (6/7) willing to recommend the tool; however, telecarer endorsement for universal rollout to patients in other clinical programs or outpatient clinics was more cautious (3/7, 43%). Qualitative findings revealed 5 themes: (1) perceptions of chatbot usability, (2) alignment and integration into telecarers’ processes, (3) factors influencing acceptability of the chatbot, (4) perceived efficacy of the chatbot for T2DM self-management and care delivery, and (5) suggested improvements and enhancements. Integrated findings revealed that the chatbot was usable and functioned as a precall support tool that optimized the personalization of tele-support sessions. However, its perceived efficacy in self-management behavior was limited by this cohort’s extensive experience with T2DM management. Key barriers to adoption included administrative burden for telecarers due to a lack of electronic medical record integration and content redundancy for patients. Suggested enhancements included multilingual support and adaptive content catering to different levels of disease-specific literacy. Conclusions: A clinical protocol–guided chatbot is feasible for hospital tele-support, complementing rather than replacing human care. It automates routine assessments and reinforces education, but future iterations must prioritize workflow integration and adaptive content to sustain engagement.

Preventing inpatient falls remains challenging for healthcare institutions globally, including in Singapore. Integrating technological innovations into fall prevention measures may optimize inpatient care and improve health outcomes. A multiphase study was conducted from 2019 to 2022, employing a human-centred design (HCD) approach to develop a technology-based inpatient fall prevention system (IFPS). The four phases include (1) pre-design observations and focus groups, (2) feature prioritization and wireframe development, (3) prototype testing and safety assessments, and (4) post-design staff training and feedback collection. The developed IFPS integrated artificial intelligence (AI) video analytics for bed-exit prediction with communication devices and autonomous commode delivery to facilitate ward communication and reduce staff workload. This paper describes the development process and user evaluation of the IFPS to assess its operational usability and safety. Potential users of the IFPS, such as ward nurses and patients, suggested features for the IFPS during the pre-design phase and thereafter evaluated the system through focus group discussions and/or feedback surveys. Pre-design focus group participants (n = 24) emphasized durability and user-friendliness requirements, informing system design. When evaluating the system, nurse users (n = 39) perceived the IFPS as effective in reducing falls (65%), enabling them to perform other duties (85%), and allowing them to remain with patients without searching for a commode (64%). Patient users (n = 21) found pre-recorded messages effective (91%), though communication clarity varied. Engaging healthcare workers in IFPS development offered valuable context-based insights, highlighting the importance of addressing technology acceptance factors early to promote adoption of fall prevention technologies in acute care settings.

Background: Social isolation and loneliness significantly impact the physical and mental health of older adults. To address these issues, telephone-based care services such as CareLine, a 24/7 service in Singapore for seniors living alone, provide emotional support, social interaction, and emergency response. This study evaluates the utilization and experiences of CareLine users and providers to understand how effectively the service meets the needs of older adults. Methods: A mixed methods approach was used, combining a retrospective analysis of CareLine’s call log dataset (November 2016 to January 2024) and in-depth interviews with users and staff. Quantitative data was analyzed using descriptive and multivariable regression analyses to explore participant characteristics and engagement patterns. Qualitative data from 20 interviews were thematically analyzed to explore user and provider perspectives. Results: The call log data included 9,512 users, with 318,950 recorded calls over the study period. Most users were female (64.6%), Chinese (82.1%), and aged 80 years or older (33.2%). Call volume peaked during the COVID-19 pandemic, with social care and emotional support being the most common call topics. Male users, those not currently married, and individuals living alone were more likely to engage with the service. Frequent callers (1% of the users) contributed to 68.5% of the inbound call volume. Thematic analysis highlighted the value of cultural and linguistic compatibility in service delivery, but also pointed to operational challenges, particularly with staffing and call management. Conclusion: CareLine effectively provides social and emotional support to a vulnerable population of older adults, particularly those living alone. However, frequent callers and high call volumes during the pandemic have stretched the service’s capacity, highlighting the need for targeted user engagement strategies, increased staffing, and resource optimization. The findings have important implications for the design and implementation of telephone-based care services for older adults globally.

Background: Family caregivers play a crucial role in providing home palliative care for non-cancer patients with advanced illnesses. Given the complexity and heterogeneity of end-of-life caregiving experiences, there is a need to understand factors influencing caregiver outcomes for different profiles of caregivers. It is also crucial to consider the influence of cultural values on caregiving experiences, particularly in Asian settings. Aim: To explore the lived experiences of caregivers of non-cancer patients with advanced illnesses in Singapore. Design: A secondary qualitative analysis of transcripts from interviews conducted as part of a realist evaluation of a non-cancer home-based palliative care programme. Framework analysis was carried out using a combined theoretical framework to identify determinants, mediators and outcomes of caregiving. Analysis was stratified according to five caregiver profiles characterised based on the perceived burdens and benefits of caregiving. Setting/participants: Twenty interview transcripts of family caregivers of non-cancer patients with advanced illnesses were analysed. Results: Two, eight, three and seven family caregivers were profiled to have ‘Intensive’, ‘Balanced’, ‘Relationship’ and ‘Satisfied’ caregiving experiences respectively. Across all profiles, caregivers’ psychological outcomes were found to be more prominent than physical outcomes. Caregivers with ‘Intensive’’ and ‘Balanced’ profiles experienced greater caregiving intensity and poorer psychological well-being, as compared to caregivers with ‘Relationship’ and ‘Satisfied’ profiles. Caregiver appraisal and quality of relationship with others were found as key mitigating factors impacting caregiving demands on well-being. Conclusions: Insights gleaned from this study can be used to develop resources for home palliative providers, tailored to the different caregiving profiles.

Background: GPFirst is a primary care partnership program designed to encourage patients with nonurgent conditions to seek care at participating general practitioner clinics instead of visiting the emergency department. In 2019, a digital media campaign (DMC) was launched to raise awareness and knowledge about GPFirst among residents in eastern Singapore. Objective: This study aims to assess the DMC’s impact on awareness and knowledge of GPFirst across different age groups, and the acceptability and satisfaction of GPFirst. Methods: The DMC, comprising Facebook posts and a website designed using the Andersen behavioral model, was evaluated through 2 repeated cross-sectional surveys. The first cross-sectional survey (CS1) was conducted with eastern Singapore residents aged 21 years and older, 2 1 year before the campaign’s launch, and the second survey (CS2) 4 months after. Satisfaction was measured on a 5-point Likert scale (very poor to excellent) about GPFirst experiences. Acceptability was assessed with 3 yes or no questions on decisions to visit or recommend GPFirst clinics. Analyses used tests of proportions, adjusted multiregression models, and age-stratified secondary analyses. Results: The Facebook posts generated 38,404 engagements within 5 months, with “#ThankYourGP” posts being the most viewed (n=24,602) and engaged (n=2618). Overall, 1191 and 1161 participants completed CS1 and CS2 respectively. Compared to CS1, CS2 participants were more aware (odds ratio [OR] 2.64, 95% CI 2.11-3.31; P<.001) and knowledgeable of GPFirst (OR 4.20, 95% CI 2.62-6.73; P<.001). Awareness was higher among married individuals (OR 1.31, 95% CI 1.04-1.66; P=.03), those without a regular primary care physician (OR 1.79, 95% CI 1.44-2.22; P<.001), and with higher education levels. Similarly, knowledge was greater among individuals with secondary (OR 2.88, 95% CI 1.35-6.17; P=.006) and preuniversity education (OR 2.56, 95% CI 1.14-5.70; P=.02), and those without a regular primary care physician (OR 1.54, 95% CI 1.02-2.34; P=.04). For acceptability, among participants who visited a GPFirst clinic, 98.2% (163/166) reported they would continue to visit a GPFirst clinic before the emergency department in the future, 95.2% (158/166) would recommend the clinic, 60.2% (100/166) cited the clinic’s participation in GPFirst as a factor in their provider’s choice and 87.3% (145/166) were satisfied with GPFirst. Among those unaware of GPFirst, 88.3% (1680/1903) would consider visiting a GPFirst clinic before the emergency department in the future. Conclusions: The DMC improved awareness and knowledge of GPFirst, with high satisfaction and acceptability among participants. Age-dependent strategies may improve GPFirst participation. The “#ThankYourGP” campaign demonstrated the potential of user-generated content to boost social media engagement, a strategy that international health systems could adopt.

Background/Objectives: This study aimed to explore the feasibility of predicting short stays among urgent admissions to an acute hospital in Singapore. With an increase in the average length of stay (LOS) in hospitals in recent years, accurately predicting short stays could enable hospitals to better manage inpatient demand and reduce emergency department (ED) overcrowding. Methods: This was a retrospective study of urgent admissions to Changi General Hospital, Singapore, from 1 January 2016 to 30 June 2022. To identify potential short stayers, a total of 25 features comprising demographic characteristics, admission and clinical characteristics, and healthcare utilization history were analyzed for each admitted patient at the point when the ED physician decided to admit the patient. The dataset was further split into a development dataset and an external validation dataset based on the year of admission. A CatBoost classifier was trained using 75% of the development dataset. Apart from reporting the model’s prediction accuracy, we conducted various analyses and simulations to study the effects of the features crucial to the prediction output. Results: The prediction accuracy of the model was evaluated on both the development test dataset (25%) and the external validation dataset. On the former, the area under the receiver operating characteristic (AUROC) and the area under the precision-recall curve (AUPRC) were 0.803 (95% CI: 0.799, 0.808) and 0.755 (95% CI: 0.749, 0.762), respectively, with the precision = 0.700 (95% CI: 0.694, 0.707) and recall = 0.692 (95% CI: 0.685, 0.699). On the external validation dataset, the performance was similar. The ED diagnosis and whether the admission required a surgical procedure were the most important features for making the prediction. Conclusions: The LOS prediction model could help providers to identify short stayers early in the course of their inpatient journeys so they could make interventions to better manage the overall utilization of hospital beds.

This study documents the experience of implementing an adaptation of the Hospital-at-Home (HaH) model to alleviate the constraints in available hospital beds and manpower amid a surge in infection rates in Singapore during the Omicron and XBB COVID waves, addressing challenges and proposing insights for scalable implementation. HaH substitutes inpatient hospitalizations by leveraging existing community healthcare services and remote healthcare technologies. This HaH adaptation was designed to be activated in during surges and deactivated when bed and manpower demands stabilize, making it less intensive on hospital resources. HaH demonstrated success in facilitating safe early discharge and admission avoidance for high-risk patients, reducing hospital bed utilization without reducing care quality. However, challenges including lack of technological literacy, language barriers, and miscommunications resulting from clerical errors were experienced. Our findings suggest that hospitals with internal resource constraints can make adaptations to leverage existing providers and assets within the community where necessary. We also observed that HaH shifts many aspects of healthcare responsibility to patients and their caregivers, which may be beyond their expected capabilities. Clear communication of expectations and limitations from all parties involved is paramount to upholding the quality of care in HaH.

Background: Telehealth interventions where providers offer support and coaching to patients with chronic conditions such as heart failure (HF) and type 2 diabetes mellitus (T2DM) are effective in improving health outcomes. However, the understanding of the content and structure of these interactions and how they relate to health care utilization remains incomplete. Objective: This study aimed to characterize the content and structure of telecare conversations on lifestyle management for patients with HF and investigate how these conversations relate to health care utilization. Methods: We leveraged real-world data from 50 patients with HF enrolled in a postdischarge telehealth program, with the primary intervention comprising a series of telephone calls from nurse telecarers over a 12-month period. For the full cohort, we transcribed 729 English-language calls and annotated conversation topics. For a subcohort (25 patients with both HF and T2DM), we annotated lifestyle management content with fine-grained dialogue acts describing typical conversational structures. For each patient, we identified calls with unusually high ratios of utterances on lifestyle management as lifestyle-focused calls. We further extracted structured data for inpatient admissions from 6 months before to 6 months after the intervention period. First, to understand conversational structures and content of lifestyle-focused calls, we compared the number of utterances, dialogue acts, and symptom attributes in lifestyle-focused calls to those in calls containing but not focused on lifestyle management. Second, to understand the perspectives of nurse telecarers on these calls, we conducted an expert evaluation where 2 nurse telecarers judged levels of concern and follow-up actions for lifestyle-focused and other calls (not focused on lifestyle management content). Finally, we assessed how the number of lifestyle-focused calls relates to the number of admissions, and to the average length of stay per admission. Results: In comparative analyses, lifestyle-focused calls had significantly fewer utterances (P=.01) and more dialogue acts (Padj=.005) than calls containing but not focused on lifestyle management. Lifestyle-focused calls did not contain deeper discussions on clinical symptoms. These findings indicate that lifestyle-focused calls entail short, intense discussions with greater emphasis on understanding patient experience and coaching than on clinical content. In the expert evaluation, nurse telecarers identified 24.2% (29/120) of calls assessed as concerning enough for follow-up. For these 29 calls, nurse telecarers were more attuned to concerns about symptoms and vitals (19/29, 65.5%) than lifestyle management concerns (4/29, 13.8%). The number of lifestyle-focused calls a patient had was modestly (but not significantly) associated with a lower average length of stay for inpatient admissions (Spearman ρ=-0.30; Padj=.06), but not with the number of admissions (Spearman ρ=-0.03; Padj=.84). Conclusions: Our approach and findings offer novel perspectives on the content, structure, and clinical associations of telehealth conversations on lifestyle management for patients with HF. Hence, our study could inform ways to enhance telehealth programs for self-care management in chronic conditions.

BACKGROUND GPFirst is a primary care partnership program designed to encourage patients with nonurgent conditions to seek care at participating general practitioner clinics instead of visiting the emergency department. In 2019, a digital media campaign (DMC) was launched to raise awareness and knowledge about GPFirst among residents in eastern Singapore. OBJECTIVE This study aims to assess the DMC’s impact on awareness and knowledge of GPFirst across different age groups, and the acceptability and satisfaction of GPFirst. METHODS The DMC, comprising Facebook posts and a website designed using the Andersen behavioral model, was evaluated through 2 repeated cross-sectional surveys. The first cross-sectional survey (CS1) was conducted with eastern Singapore residents aged 21 years and older, 2 1 year before the campaign’s launch, and the second survey (CS2) 4 months after. Satisfaction was measured on a 5-point Likert scale (very poor to excellent) about GPFirst experiences. Acceptability was assessed with 3 yes or no questions on decisions to visit or recommend GPFirst clinics. Analyses used tests of proportions, adjusted multiregression models, and age-stratified secondary analyses. RESULTS The Facebook posts generated 38,404 engagements within 5 months, with “#ThankYourGP” posts being the most viewed (n=24,602) and engaged (n=2618). Overall, 1191 and 1161 participants completed CS1 and CS2 respectively. Compared to CS1, CS2 participants were more aware (odds ratio [OR] 2.64, 95% CI 2.11-3.31; P<.001) and knowledgeable of GPFirst (OR 4.20, 95% CI 2.62-6.73; P<.001). Awareness was higher among married individuals (OR 1.31, 95% CI 1.04-1.66; P=.03), those without a regular primary care physician (OR 1.79, 95% CI 1.44-2.22; P<.001), and with higher education levels. Similarly, knowledge was greater among individuals with secondary (OR 2.88, 95% CI 1.35-6.17; P=.006) and preuniversity education (OR 2.56, 95% CI 1.14-5.70; P=.02), and those without a regular primary care physician (OR 1.54, 95% CI 1.02-2.34; P=.04). For acceptability, among participants who visited a GPFirst clinic, 98.2% (163/166) reported they would continue to visit a GPFirst clinic before the emergency department in the future, 95.2% (158/166) would recommend the clinic, 60.2% (100/166) cited the clinic’s participation in GPFirst as a factor in their provider’s choice and 87.3% (145/166) were satisfied with GPFirst. Among those unaware of GPFirst, 88.3% (1680/1903) would consider visiting a GPFirst clinic before the emergency department in the future. CONCLUSIONS The DMC improved awareness and knowledge of GPFirst, with high satisfaction and acceptability among participants. Age-dependent strategies may improve GPFirst participation. The “#ThankYourGP” campaign demonstrated the potential of user-generated content to boost social media engagement, a strategy that international health systems could adopt. CLINICALTRIAL

Objective This study aims to explore the feasibility of employing convolutional neural networks for detecting and localizing implant cutouts on anteroposterior pelvic radiographs. Materials and methods The research involves the development of two Deep Learning models. Initially, a model was created for image-level classification of implant cutouts using 40191 pelvic radiographs obtained from a single institution. The radiographs were partitioned into training, validation, and hold-out test datasets in a 6/2/2 ratio. Performance metrics including the area under the receiver operator characteristics curve (AUROC), sensitivity, and specificity were calculated using the test dataset. Additionally, a second object detection model was trained to localize implant cutouts within the same dataset. Bounding box visualizations were generated on images predicted as cutout-positive by the classification model in the test dataset, serving as an adjunct for assessing algorithm validity. Results The classification model had an accuracy of 99.7%, sensitivity of 84.6%, specificity of 99.8%, AUROC of 0.998 (95% CI: 0.996, 0.999) and AUPRC of 0.774 (95% CI: 0.646, 0.880). From the pelvic radiographs predicted as cutout-positive, the object detection model could achieve 95.5% localization accuracy on true positive images, but falsely generated 14 results from the 15 false-positive predictions. Conclusion The classification model showed fair accuracy for detection of implant cutouts, while the object detection model effectively localized cutout. This serves as proof of concept of using a deep learning-based approach for classification and localization of implant cutouts from pelvic radiographs.