The lack of accurate, reliable and consistent patient information is a major issue in healthcare, despite a relatively high Health Information System (HIS) adoption level worldwide. The main reason for this appears to be patient records lacking accurate particulars, including links to associated care programs, disease classification and treatment plans. The causes for this are multiple, including incompatibility of healthcare standards between version releases, inconsistent HIS implementation, lack of effective data input / validation, and the rapid evolution of and absence of a single ‘universal’ technological solution. Sustainable, stable and long-term architectural solutions are required. This research builds on previous work identifying major challenges and root causes of the problem and proposing essential non-functional requirements for HIS architectures. The paper elaborates on non-functional requirements and proposes an evaluation framework (based on a new international standard) that can be used to assess aspiring HIS architectures for long term stability and self-evolution and thus to support strategic decision making from within the evolving HIS.