JMIR Publications

ABSTRACT

Background:

The escalating global scarcity of skilled healthcare professionals is a critical concern, further exacerbated by rising stress levels and clinician burnout rates. Artificial intelligence (AI) has surfaced as a potential resource to alleviate these challenges. Nevertheless, we should not presuppose that AI will innately augment human performance, as improperly designed healthcare technology may inadvertently impose new burdens on healthcare workers. An in-depth understanding of how AI can effectively enhance rather than strain work conditions is therefore needed.

Objective:

This research investigates the efficacy of AI in alleviating stress and enriching work conditions, using Intensive Care Units (ICU) as a case study. Through the lens of a socio-technical system perspective, we delineate how the integration of AI and the professional responsibilities of ICU physicians and nurses should be co-designed to foster a motivating, stimulating, and health-promoting work environment.

Methods:

We employ the socio-technical system framework COMPASS to analyze and assess five fundamental work characteristics in present and prospective work conditions: autonomy in decision-making, skill diversity and competence development, flexibility in time and location, problem-solving opportunities, and task identity and variety. The analysis is underpinned by 250 hours of workplace observation in six European ICUs (N= 562 physicians and nurses), 12 in-depth interviews with work unit leaders, and an expert evaluation of optimal levels of human-AI collaboration.

Results:

The results indicate that AI holds the potential to positively impact work conditions for ICU physicians and nurses in several key areas. Firstly, decision-making autonomy is vital for stress reduction, motivation, and performance improvement. AI systems that ensure transparency, predictability, and human control can reinforce or amplify autonomy. Secondly, AI can encourage skill diversity and competence development, empowering professionals to broaden their skills, achieve polyvalence, and deepen interprofessional cooperation. However, careful consideration is required to avoid deskilling of experienced professionals. Thirdly, AI automation can expand flexibility by relieving healthcare professionals from administrative duties, thereby concentrating their efforts on patient care. Remote monitoring and improved scheduling can additionally improve work-life integration. Fourthly, while AI may reduce problem-solving opportunities in certain areas, it can open new pathways, particularly for nurses, thereby fostering interprofessional information sharing and decision-making. Lastly, maintaining task identity and variety is crucial to sustaining intrinsic motivation and engagement.

Conclusions:

This research underscores the potential of AI in alleviating stress and enhancing work conditions for overburdened ICU physicians and nurses. By adopting a socio-technical viewpoint, technology designers can engineer AI solutions aligned with the principles of good work design, while hospital management can strategically choose AI solutions that boost motivation, job satisfaction, and well-being. This research thereby indirectly contributes to addressing the shortage of skilled professionals and enhancing the quality of care.