Procuring AI-systems in Healthcare: More Than a Technology Contract  

Artificial intelligence is increasingly being introduced into healthcare, often with promises of improved efficiency, decision support, and reduced administrative burden. Yet purchasing AI is not simply a matter of selecting the best technology. It involves balancing clinical needs, patient safety, work environment considerations, legal requirements, organizational priorities, and economic constraints. 

As part of the AI-PROCARE project, we have so far conducted 68 interviews with stakeholders involved in AI procurement across Denmark, Finland, Iceland, and Sweden.  

A diverse set of voices 
To understand AI procurement from different perspectives, we interviewed: 

  • Clinicians and healthcare managers 
  • Procurement officials and advisors 
  • National and regional regulatory actors 
  • Vendors and system management representatives 
  • Trade union representatives 

This has helped us go beyond viewing procurement as a purely administrative process and instead explore how different actors contribute to decisions about introducing AI into healthcare practice. 

A broad variety of AI applications 
The interviews covered a wide range of AI technologies, including: 

  • Image analysis tools such as mammography and radiology algorithms,  
  • Documentation support such as transcription, summarising and translation solutions 
  • Decision support such as triage systems, fracture detection tools and stroke detection 

This diversity provided valuable opportunities to compare procurement challenges across different types of AI solutions. 

What happens next? 
We aim to conduct a total of 100 interviews and are currently recruiting the remaining participants. The ongoing analysis focuses on how different stakeholders perceive and navigate tensions between: 

  • Efficiency and work environment 
  • Clinical value and economic value 
  • User influence and managerial decision-making 
  • Standard procurement procedures and the unique characteristics of AI 
  • Innovation and regulation 

The next work package will involve in-depth case studies of successful AI procurement initiatives, with a particular focus on the experiences and perspectives of end-users. By examining procurement as a socio-technical process rather than merely a purchasing activity, the project seeks to generate knowledge that can support more sustainable work envrionment and human-centered implementation of AI in healthcare. 

Written by Maral Babapour Chafi