- 7274
With the ultimate aim of reducing the mortality and severity of Cancer within the US and Japan; (a) what evolving factors surrounding the diagnosis, care and treatment of cancer should be incorporated into insurance products to predict treatment, better cover policyholders and/or appropriately price coverage options and;(b) Could the wealth of structured and unstructured claims data that exists in Aflac be leveraged to help determine and predict indicators for severe health issues such as cancer, heart attack, stroke etc?
What evolving factors surrounding the diagnosis, care and treatment of cancer should be incorporated into insurance products to predict treatment, better cover policyholders and/or appropriately price coverage options? With the ultimate aim of reducing the mortality and severity of Cancer within the US and Japan could the wealth of structured and unstructured claims data that exists in Aflac be leveraged to help determine and predict indicators for severe health issues such as cancer, heart attack, stroke etc?
- 7277
What is the predictive accuracy of wearable technology data (e.g., fitness trackers, heart rate monitors) in assessing individual health outcomes specifically related to Cancer, and how can this information be used to design more personalized supplemental insurance policies?
What is the predictive accuracy of wearable technology data (e.g., fitness trackers, heart rate monitors) in assessing individual health outcomes specifically related to Cancer, and how can this information be used to design more personalized supplemental insurance policies?
- 7179
How can generative Artificial Intelligence be effectively integrated into healthcare processes to enhance patient care, streamline medical workflows, and support clinical decision-making, whilst ensuring compliance with regulatory standards and addressing ethical concerns around data privacy and bias? What is currently happening in Northern Ireland and the opportunities that Encompass will bring?
- 7279
AI-driven models have had growing success in the discovery and optimisation of therapeutic proteins and antibodies, for example, in protein structure prediction. However, challenges remain in how best to use such insights to create an integrated platform to accelerate the development process and optimise therapeutic efficacy and safety. What strategies, AI/ML/statistical techniques and computational tools could be used to create such a platform, integrating real-world data from pre-clinical and clinical studies, to develop highly specific and functional therapeutic candidates?
AI-driven models have had growing success in the discovery and optimisation of therapeutic proteins and antibodies, for example, in protein structure prediction. However, challenges remain in how best to use such insights to create an integrated platform to accelerate the development process and optimise therapeutic efficacy and safety. What strategies, AI/ML/statistical techniques and computational tools could be used to create such a platform, integrating real-world data from pre-clinical and clinical studies, to develop highly specific and functional therapeutic candidates?
- 7198
(a) What are the latest advancements and challenges in applying single-cell genomics for predictive health analytics, particularly in forecasting disease progression, treatment response, and drug resistance in precision medicine? and
(b)What are the latest breakthroughs in predicting tumor evolution, and how does Multiomics contribute to improving predictive models and patient outcomes in precision oncology?
- 7180
(a) What cutting-edge techniques within statistics and AI/ML exist for integrating multimodal data (e.g., combining eye tracking and speech patterns) into a cohesive model for autism diagnosis? How can we ensure the quality and reliability of data collected from VR/XR environments, particularly when it comes to sensitive indicators like facial expressions and body movement? Are there any relevant existing datasets that we can leverage, in addition to the data we are currently collecting, when building models for autism diagnosis?
(b)2. What are the ethical implications of using AI for autism diagnosis, and how can we ensure our models are developed and validated in an ethically sound manner? What regulatory challenges are there when developing such approaches and are there any recommended pathways for obtaining regulatory approval across different regions and countries?
- 7215
Data governance and the access to high quality healthcare data is of high importance to SMEs working in predictive healthcare modelling. How can SMEs best balance the need for information governance with effective data sharing when developing an analytical healthcare tool, maintaining high standards of privacy and accounting for any ethical concerns throughout the process? What tools or frameworks are available to help with this and how might the integration of Encompass in NI change current practices?
Joint Data Governance Session.
- 7129
How could we conduct a proof of concept trial on music therapy in either alzheimers or parkinsons?
Lú Music Technology Limited (Lú) is a Startup focused on exploring the use of music therapy to improve mental health and assist in the treatment of neurological conditions. Specifically, Lú has been developing an engine, the Unlimited Harmonic Expansion Seeder (UHES), designed to support Emotional Regulation. UHES is based on a proprietary algorithm and analyses harmonics and subharmonics to tailor music therapy sessions, gradually shifting an individual’s emotional state through various pathways. For example, UHES could be used to predict subharmonic resonances of particular songs to optimize auditory responses along different bands of the brainwave spectrum, potentially aiding in the treatment of conditions like Alzheimer’s and Parkinson’s disease. Our UHES builds a map of resonances, which are tuned ( and pruned) to reveal potential areas of treatment for both mental health and neurological conditions. Using UHES, we could then utilise recorded music and modulate it to suit therapeutic needs. UHES is on track for trial release in Q4 2024. As of 2023
- 7237
What are the most effective strategies to engage patients who are at high risk of non-adherence, particularly in populations with low technological proficiency, and how can AI-driven solutions be leveraged to address these barriers?
Medication Non-Adherence Impact: Medication non-adherence is a widespread issue that leads to poorer health outcomes, increased hospitalisations, and significant healthcare costs. It is estimated that 30-50% of patients with chronic conditions do not take their medications as prescribed, resulting in billions of pounds lost annually across healthcare systems.
Barriers to Adherence: Non-adherence can result from various factors, including complex medication regimens, side effects, lack of understanding about the importance of the treatment, and forgetfulness. In particular, patients who are less familiar with technology or have cognitive limitations can struggle with adherence, as many existing digital solutions are app-based and may not cater to these groups effectively.
Role of AI in Addressing Non-Adherence: AI-driven solutions have the potential to personalise patient support and deliver reminders through accessible channels, such as telephone calls, rather than solely relying on digital apps. Leveraging AI to anticipate adherence risks and adapt interventions based on patient behaviour can significantly enhance patient engagement and improve health outcomes.
Question Focus: The question aims to explore effective strategies to address non-adherence in high-risk populations, particularly where technological proficiency is a barrier, and seeks to understand how innovative AI solutions can be tailored to meet these needs. This topic is critical for developing inclusive, scalable solutions that support diverse patient groups in adhering to their prescribed treatments.
- 7234
What is the most effective way to allocate funds and research facilities for developing multi-omic AI applications in cancer prediction – prognosis, diagnosis, patient identification (for clinical trials) and treatment outcome? How new advancements in AI may be utilised for cancer prediction?
We are particularly excited about the opportunity to partner with Queen’s University Belfast. Their expertise in bioinformatics tools for cancer drug discovery, including target identification and machine learning approaches, complements our platform perfectly. Together, we plan to develop novel AI approaches in metabolomics and proteomics for precision oncology. These techniques will enable precise metabolite and protein mapping, facilitating accurate phenotype categorization and true activity measurement. This partnership will empower Meta-Flux to provide cutting-edge solutions to biotech and pharmaceutical companies, achieving a minimum of 25% savings in time and expenditure
compared to traditional methods. The integration of bioinformatic tools will provide the necessary infrastructure for technological growth, positioning Meta-Flux as a global competitor in a $2.4 trillion market. Furthermore, the enhancement of Meta-Flux
technologies through our collaboration will have a broader impact on the border regions. By utilizing this advanced technology, aspiring biotech companies in these areas can develop assets faster and more cost-effectively, with enhanced insights. This will not only accelerate their growth but also contribute to the overall economic development of the region, fostering innovation and creating new opportunities in the biotech sector. We look forward to the opportunity to collaborate with Queen’s University Belfast and
contribute to the advancement of drug development technologies.
- 7258
The NI Ambulance Service has very high levels of staff sickness, with musculoskeletal injuries and work-related stress being two of the primary reasons. All staff members at any point in time are either not sick, on short term sick or long-term sick. NIAS has many related datasets on what calls staff answer and attend, shift patterns, historical absence at an individual and organisational level, training records, late finishes, violence and aggression against staff, demographic data and pay/conditions including levels of overtime worked. Can these data be used to identify staff at risk of becoming ill and to inform absence reduction strategies?
The NI Ambulance Service has very high levels of staff sickness, with musculoskeletal injuries and work-related stress being two of the primary reasons. All staff members at any point in time are either not sick, on short term sick or long-term sick. NIAS has many related datasets on what calls staff answer and attend, shift patterns, historical absence at an individual and organisational level, training records, late finishes, violence and aggression against staff, demographic data and pay/conditions including levels of overtime worked. Can these data be used to identify staff at risk of becoming ill and to inform absence reduction strategies?
- 7212
Automating the triage and reporting of radiographs through AI is often an attractive option. Despite this, adoption of AI in radiology has been slower than anticipated. What factors contributes to the slower adoption of AI in radiology and how these issues can be addressed? Are there new advancements in AI that could translate to radiology that haven’t been tested yet which may help with such adoption?
ChestlInk is an algorithm that removes healthy normal (clinically insignificant) chest xrays from the reading list by creating a final report. It is the only solution certified to do so. Are there any radiologists interested in trialling the solution at their institution, in order to find out what percentage of their workload could be automated?
- 7192
We are developing an AI driven patient focused circadian lighting system for people living with dementia, building a database of behaviour patterns in order to improve quality of life, wellbeing and reduce falls. We propose to develop our system by capturing data on environmental criteria such as temperature, humidity, air quality or other variable. Which would be most valuable to capture in a clinical trial in order to enhance patient engagement?
- 7216
How can clinical research units / QUB assist SMEs validate their predictive AI platforms to confidently identify and recruit patients for commercial clinical trials?
Accessing health data from public health systems like HSCNI is crucial for SMEs aiming to build AI-based predictive models for oncology care and clinical trials. However, challenges such as data privacy concerns, strict regulatory frameworks, and navigating the transition from legacy systems to new platforms like Encompass create barriers. SMEs require access to anonymized or securely managed datasets to train their AI models, but the process can be lengthy and complex due to data protection regulations like GDPR. Overcoming these challenges requires strategic collaboration and a clear plan to ensure compliance while gaining necessary data access.
- 7183
What are the biggest challenges in combining real time self-report patient contributed health data?
We are interested in combining big data and small data for the purpose of comparison of predicting personal health.
- 7231
How Can We Build a Secure, Unified Data Network for Healthcare Research and Product Development in Northern Ireland?
This session will explore the creation of a secure, integrated data network for healthcare research and product development across Northern Ireland, in alignment with the new Encompass system. Given the complexity of healthcare data access in Northern Ireland, which involves various regulations and agreements, the goal is to develop a “one-stop shop” for anonymised patient data, similar to the MIMIC dataset or NHS England digital datasets. By consolidating data from different systems into a secure network, we aim to enable the mapping of full patient journeys and longitudinal pathways, that can be easily accessed and utilised. Participants will identify challenges they have encountered when accessing healthcare data or users for research and product testing, and collaborate to outline actionable next steps for building this data network to benefit both academia and industry partners.