Digital Health and AI 2024

It is essential to ensure that investments in AI technologies enable accurate identification of the health risks to inform and direct planners towards appropriate sources of support. Importantly, more diverse data must be collected to address the current biases. At the same time, more investment is needed to achieve the desired long-term outcomes – sustained lifestyle behavioural changes. Uptake of digital health opportunities is often most challenging in communities with some of the most significant health challenges and existing health disparities. For employers, integrating pricing strategies into procurement processes that correlate with levels of employee engagement achieved and health outcomes delivered has significant benefits for employer and employee alike over fixed pricing approaches to access the technology. Data privacy, information security and ethical concerns AI systems rely on access to personal data in order to function. Such data is already subject to international and national regulations to protect the privacy of the individual. In harnessing the benefits of AI in occupational-health settings, the same requirements apply to ensure that data cannot be accessed by unauthorized users nor used for purposes other than those agreed with users. Failure to address these concerns in workplace settings can lead to privacy violations, including identity theft. There is a deeper philosophical question as to whether companies should track all of the personal health data they are interested in. This is not just a data privacy issue, it is also about workers’ agency in looking after their own health. While controlling exposure to workplace hazards is clearly an issue of management accountability, it is the workers themselves who are surely responsible for many of the lifestyle changes and better health habits needed to address their own health risks. However, employer support – for example, by providing healthy choices where food is available – can help bring about positive lifestyle choices. Therefore, finding a balance between employer support and workers’ agency is crucial here.In addition, if AI models predict the early onset of cancers/ illnesses through early scanning programmes at work, what are the implications in terms of current and future employment decisions and access to health insurance, notwithstanding the equality regulations currently in place to protect employees from discrimination? Quality of evidence for sustainable behavioural change and health outcomes from using wellness tools While the issue of data quality is well known, the risks and costs of using poor data to shape AI development cannot be overestimated. Not only might some people receive inappropriate or potentially harmful advice through poor-quality data, but others might be excluded altogether from access to care. In the end, such exclusion leads to significant harm and a costly erosion of trust in the tools and the employers who designed and implemented them.23 Developers of AI tools and those who choose to promote them, including employers, must ensure that the claims made for their benefit, such as changing health outcomes, are trustworthy. The European Union AI Act specifically requires this in order to protect the health, safety and human rights of those who rely on them.24 With the race to deploy AI interventions in health and healthcare, the importance of clinical evidence for positive health outcomes is critical. Chief health officers are well placed to contribute to governance in workplace settings, and to advocate for a more equitable, participatory approach to the development of AI algorithms. I’m excited about AI’s potential across drug discovery, diagnostics, personalized health and safety, all aimed at delivering more effective health and well-being. We are actively integrating AI into drug discovery and development to accelerate the creation of new molecules and medicines. We’re also leveraging AI to enhance employee health and safety: wearable devices provide real- time feedback on movement speed in sterile gowning areas, with haptic alerts guiding optimal speed to reduce contamination and injury risks. This is just the beginning – AI has the power to transform diagnostics and therapeutic care as well, not by replacing clinicians but by supporting them in delivering precision care. We’re only beginning to see how AI will shape the future of patient care and the making of medicines. Michael A. Kaufmann, Associate Vice-President, Global Employee Health Services, Eli Lilly and Company, USA