Digital Health and AI 2024

Early detection of exposure to harmful workplace substances and improved safety standards The International Labour Organization (ILO) estimates that nearly 3 million people globally lose their lives each year to work-related incidents or illnesses, while 395 million people suffer a work-related accident annually.8 AI has the potential to analyse vast quantities of complex safety and health data, highlight the root causes of work-related accidents and illnesses and facilitate the targeting of evidence-based opportunities for prevention and improvement. In cases where adverse health outcomes can be reliably predicted by particular unsafe conditions and worker and/or supervisor behaviour, preventative action can be better targeted. Modifications to workplace design needed to prevent recurrence can then be identified to reduce the risks of worker harm in the future.The provision of wearables as a part of corporate wellness programmes can also serve as a means of articulating a company’s concern for its employees’ health and well- being. Financial and non-financial incentives can be aligned with the degree to which employees participate in physical activity, healthy eating, psychological well-being and other wellness programmes to further encourage commitment. In addition to the benefits to employee health, employers report greater engagement in wellness programmes and increased returns, particularly for mental health investments.6 Group programmes also have the potential to improve social connectedness, which is increasingly reported as protective of mental health and well-being.7 AI and digital technologies are significantly changing the way we work. From a health and safety perspective, integrating data metrics can support the prediction of potential accidents and support our people to be better equipped for future challenges. One exciting development might be the use of digital twins to simulate factories and workplaces, enabling us to create predictive models. These models could forecast ways of helping to prevent accidents and work-related illnesses. Ralf Franke, Executive Vice-President, Environmental Protection, Health Management and Safety, Siemens, Germany The use of AI to reduce ergonomic risks is another area of promise. Using AI software to analyse workspace ergonomics and identify poor postures and repetitive tasks has the potential to reduce the risk of musculoskeletal injury. Musculoskeletal conditions are the leading contributor to disability and human suffering worldwide11 and are estimated to affect approximately 1.71 billion people globally. With the impacts of climate change increasingly apparent, the number of people exposed to extreme heat is growing exponentially in many parts of the world.12 One consequence of this is that more workers are at risk of exposure to prolonged and excessive heat. With its capacity for big data analysis and predictive modelling, AI can play an important role in preventing work-related heat illness. Algorithms can trigger interventions when the environmental risks may be highest, based on the analysis of past weather conditions. Current data can be used to optimize ventilation and cooling systems in built and working environments. By combining environmental data with individual worker data, workers and their supervisors can be alerted when physiological signs (e.g. body temperature, hydration, heart rate) suggest a change is needed in activity or worker behaviour to reduce the likelihood of heat-related illness – for example, communicating a reminder to take a break, seek shade or rehydrate. Enhanced access to affordable healthcare With many populations geographically distant from health experts, advances in technology are enabling individual health- risk data to drive personalized advice or interventions remotely. Generative AI (GenAI) has joined the list of guided (coached) and unguided (self-help) digital tools that aim to deliver health advice, especially early mental health interventions.AI can also improve occupational-health surveillance, enhancing the effectiveness of early detection and the targeting of health interventions. For example, a key application of AI in occupational health surveillance is in the control of noise-induced hearing loss at work. With exposure to workplace noise estimated to be one of the most common occupational hazards in the USA and Europe,9 the opportunity is compelling, particularly for those working in construction, manufacturing and mining. In addition to the workplace factors (including noise exposure and the design of engineering protections), the risks of hearing loss in workers are affected by a complex mix of personal factors (hypertension, hyperlipidaemia, etc.) and individual health-related behaviour (selection and appropriate use of personal protective measures, smoking, drinking, etc.), all of which can be prevented and mitigated using AI’s predictive power. AI and machine learning are well placed to combine the focus on organizational prevention strategies (implementation of engineering controls) with personal behaviour change (better condition management, use of personal protective equipment and lifestyle changes), pinpoint areas for interventions and highlight accountabilities for better health outcomes. AI-enabled imaging holds great potential for early detection of many diseases affecting workers, such as cardiovascular diseases, cancers and respiratory diseases.10 For instance, AI- interpreted radiographic imaging is already showing promise in improving the early diagnosis and management of work-related lung diseases. With early detection of the preclinical health effects of dust exposure in populations with limited access to professional advice in low- and middle-income countries (LMICs), the risks of silicosis and other lung diseases can be reduced. Similarly, AI can be used to detect the early onset of cancers.