Sessions

An exciting lecture where we delve into the current hype surrounding ChatGPT and large language models. Are they truly the holy grail in healthcare, coming sooner than we think, or are they just another tool in the toolbox? What do we need to be cautious about?

The landscape of healthcare is rapidly evolving, with patient safety and the efficiency of patient monitoring gaining increased importance. Traditional spot-based patient surveillance methods have notable limitations, and there is substantial documentation on the benefits of more continuous monitoring approaches. Vitalthings, a Norwegian health technology company, is at the forefront of this evolution, developing contactless monitoring solutions that significantly enhance patient safety, operational efficiency, and overall cost-effectiveness. Dr. Thu will elaborate on how the continuous data harvested by this system serves as a foundation for future research and the development of AI models. This shift towards more proactive healthcare modalities, empowered by enhanced patient and staff safety, paves the way for the integration of AI into predictive medicine.

This presentation offers a perspective to some of the successes and failures in the Finnish ecosystem for sharing health data over the last 10 years.

Discover the future of healthcare through the lens of artificial intelligence. In this presentation, we delve into two major AI approaches: tabular and generative. We'll guide you through a complete solution, from sourcing data to module creation and validation, with the goal of delivering actionable insights for streamlined workflows. Learn how our services can accelerate your journey to value in healthcare AI.

AI has been hailed as one of the instrumental changes that can help healthcare systems tackle current and future challenges. But to what extent can AI help? Milan Shah, NHS transformation specialist, talks you through a study that was conducted on a Swedish primary care health center to assess the impact of AI-enabled triage on labour productivity.

Implementing AI in healthcare safely and effectively presents a unique set of challenges. From data quality, information security, systems change, evidence generation, clinical safety and workforce transformation this session discussions the key challenges in the adoption of AI and how they might be tackled.

Successful AI deployment within healthcare relies on the trust and confidence of healthcare professionals.  Based on the findings of the NHS AI lab and Health Education England report on Healthcare workers’ confidence in AI, Dr Painter explores which key factors influence confidence in healthcare AI technologies.

Dermatologists are uniquely trained and experienced in managing all aspects of skin health, making them invaluable healthcare providers for individuals seeking treatment for dermatological concerns. Their medical training is to "see" and "remember" dermatology pathologies. During their lifetime they see about 200 000 dermatology pathologies. In 2018 we trained a deep neural network on 55,364 smartphone images of dermatology pathologies in 8 hours and used the artificial intelligence model in a primary care setting. Referrals to a dermatologist could have been prevented in 34% of cases, in 60% of the cases the AI model helped the primary care physician diagnose correctly and 92% of the primary care physicians loved the AI decision support tool.

This study examines the availability of information on welfare technologies or people living with dementia in Sweden. We conducted a scoping review of 21 counties and 290 municipalities to assess the degree of information provided on welfare technologies. 

Earlier detection of interval cancer with AI-powered personalized breast self-checks. 

This presentation explores the complexities and costs associated with implementing new systems in healthcare, focusing on the challenges of migrating patient data and maintaining interoperability between systems. It proposes a novel approach using lightweight FHIR messages and OpenEHR standards to streamline data integration and preserve the integrity of electronic patient records. By adopting these methods, healthcare providers can enhance patient care, reduce clinical risks, and facilitate research, ultimately leading to more efficient and patient-centric healthcare solutions.

Artificial intelligence (AI) has the potential to improve the efficiency and safety of medical history-taking and triage in primary care. However, there is a lack of knowledge about the practical implementation of AI systems for these purposes. This study explores healthcare leaders’ experiences of barriers to implementing an AI system for automated history-taking and triage in primary care, and the actions they took to overcome these barriers.

The healthcare sector often utilizes various methods to track patients' conditions, symptoms, and progress. A critical aspect of this tracking is the identification and documentation of Adverse Drug Reactions (ADRs), which include symptoms like nausea and drowsiness caused by medications. These ADRs, typically categorized by severity and frequency in databases, require medical professionals to meticulously connect them to specific drugs, a process that demands considerable expertise and time. Given the complexity of drug interactions and the limited time available for thorough investigation, this crucial area often remains under-explored. Artificial Intelligence (AI) emerges as a pivotal tool in addressing these challenges, offering decision support to doctors in pinpointing the causes of ADRs and facilitating swift and accurate drug follow-ups. This presentation will delve into a research study from a Master's Thesis project at Karolinska Institutet, exploring AI's potential in identifying rare ADR cases and aiding physicians in efficient medication management.

Staff welfare has a profound impact on emergency services' abilities to meet targets, deliver appropriate levels of care and implement long-term strategies. Poor staff morale and high employee turnover are disruptive, expensive and result in sub-optimal service provision. As such, more organisations are turning to advanced modelling solutions to better balance operational performance and employee well-being. This presentation draws on first-hand experience from the Welsh Ambulance Service NHS Trust (WAST) and details how data-driven modelling software enables the organisation to implement effective changes that prioritise employee welfare, respect operational targets and empower the team to do more with less.

Case study describing AI Chest Radiograph the approach to evaluation to support decision making to adopt

InfCare HIV quality register and decision support tool has now been used nationwide in Sweden and in other Nordic countries for 20 years. In Sweden, a great amount of structured clinical data has been collected during the years and the Swedish HIV physicians have included the decision support system as an integral part of all patient-related workflows in the clinic. The system gives clinicians in their daily practice a holistic view of the patient’s state of health over time as well as real time access to key performance indicators regarding the care given at the clinic.Data from InfCare HIV has been used in numerous research projects and studies to support developing treatment guidelines and improve care pathways for HIV patients. BMJ Open recently published an article describing the first 20 years with the Swedish InfCare HIV cohort, Christina Carlander as the main writer.Beyond the InfCare HIV system, the presentation will cover some examples of similar clinical information management systems for other medical specialties.

Currently, an individual’s opportunities for privacy protection are limited by their digital skills and external factors such as illness, residence, and socio-economic status. Technological development in the field is happening very quickly, but coordination and common platforms are needed to increase accessibility. The HEIDA project have developed software examples using homomorphic encryption for end-to-end privacy protection. The technology enables processing of encrypted data and increase the individual’s opportunities for privacy protection in digital health, which can be of interest to health care decision makers and IT professionals.

The roll-out of digital innovation across all tiers of healthcare provision can inherently result in some patients finding it harder or indeed impossible to use new digital models of accessing care. You will hear about national UK NHS work undertaken to understand how to overcome barriers to creating digitally equitable pathways, in both hospital and primary care. Find out what tools, knowledge and guidance is needed to achieve Inclusive Digital Transformation across clinical pathways.