Proposed Topic (Most preferred): :
Research and Innovations (new projects / technology / innovations / service models)
Proposed Topic (Second preferred): :
Enhancing Partnership with Patients and Community (Projects initiated to engage patients / carers / community to improve efficiency / quality of care)
Authors (including presenting author) :
Cheng HL(1), Ku BBS(1), So BCH(1), Wan JCH(1), Lui KWC(1), Wong JTK(1), Yeung BCY(1), Cheng DLL(1), Yu VMO(1), Lee DPK(1), Poon JCM(1), Lansley P(1), Chan JKY(1), Pang JYW(1), Cheung NT(1)
Affiliation :
(1) Information Technology and Health Informatics Division, Hospital Authority Head Office
Introduction :
In an era where technology is advancing at an unprecedented rate, Hospital Authority recognized the need to innovate and harness the power of artificial intelligence (AI) to enhance patient engagement. Consequently, we developed the first AI-powered feature within HA Go Apps, which serves as a digital front door, catering to a wide range of population. Notably, a large number of our users are elderly, and this innovative AI application has been specifically designed to facilitate smoother and faster health value input for this demographic. This abstract highlights the key features and outcomes of our visual AI-powered HA Go Apps, showcasing its potential to revolutionize the healthcare experience for all users, particularly the elderly.
Objectives :
The primary objectives of this project were to empower patients, improve the mobile experience, maximize efficiency in healthcare delivery, and support regular measurement of health values. Specifically, we aimed to integrate AI technology into the HA Go to automate the recognition of vital signs, such as blood pressure readings, from images captured using the phone's camera. Additionally, we sought to expand the applicability of the AI model to other measurements, including weights and blood glucose levels.
Methodology :
Utilizing a multi-platform approach, our team integrated the Hospital Authority (HA) built AI model with HA Go Apps which support Apple's iOS, Android and Huawei systems. To enhance accessibility and engage a wide range of patients, we prioritized the creation of a user-friendly interface. Our application's functionality included the ability to capture images of blood pressure monitors using the phone's camera. We developed our own AI model that recognizes specific 7 segment characters commonly found on various health machines. To train our model, we utilized a data set of over 14,000 images. The AUROC (Area under the receiver operating characteristics) for our AI model was an impressive 0.971, indicating its high predictive accuracy. Additionally, our model demonstrated an overall accuracy of 95%. These results demonstrate the success of our approach in effectively utilizing AI technology within the healthcare sector. By leveraging the HA's expertise and resources, we have developed a robust AI model that performs at a high level of accuracy, enabling us to provide reliable and efficient healthcare services to our patients.
Result & Outcome :
The project was implemented since end of September 2023. Total over 83,000 images were processed. The implementation of the AI-powered HA Go Apps yielded several positive outcomes. Foremost, patient empowerment was achieved as individuals were able to monitor their own vital signs conveniently, enabling greater involvement in their own healthcare. The application's multi-platform support and ease of use contributed to a seamless and enhanced mobile experience for the users. Furthermore, the successful integration of the AI model for blood pressure recognition paved the way for future expansions to other vital sign measurements and range of motion assessments. In conclusion, the AI-powered mobile application served as a transformative tool in our healthcare organization, facilitating patient empowerment and delivering an improved mobile experience. With the demonstrated success of the initial implementation, we foresee a brighter future where AI technology continues to enhance healthcare delivery and patient engagement
