Medical imaging and Picture Archiving and Communication Systems (PACS) in particular, have appeared to be one of the promising areas for Open Source Software (OSS). Open source medical imaging solutions do exist, including PACS, but are not widely deployed at hospitals and health care establishments, which prevents them from achieving their full potential. In most cases where medical OSS systems exist (i.e. not necessarily PACS), it is to a very limited degree, and typically out of sight of the common user. Research we have conducted on medical software companies in North Norway and three hospitals in Europe suggests that if open source medical software is to become a useful alternative to proprietary software, that firstly, the initiative must be taken by the public health services and secondly, that it will require a shift from software companies (from sale-value oriented to service oriented). However, it would be naïve to rely on existing proprietary companies to initiate such a change. Interviews revealed that some companies considered the possibilities of developing using an OSS model, but did not deem it to be profitable, whereas others stated that it was simply out of the question. They are not willing to risk their successful business models, because historically (and perhaps at the cost of quality) it pays to keep the inner workings of their software secret. Other reasons revealed for not using OSS were: poor support, prejudices and the unwillingness of proprietary companies to accept a new business model. We suggest that these problems can be overcome with the emergence of competence centers for OSS, and that if open source medical imaging, PACS projects in particular, are to get started, they are more likely to succeed if a hospital is involved. However, our suggestions can only be tested thoroughly if more implementations are done.