Design studies are a core methodology in visualisation for solving real-world problems, but applying them in complex domains, such as clinical visual analytics, encounters well-recognised challenges. Existing frameworks provide rigour but often fall short in guiding systematic long-term, cross-disciplinary collaborations and sustainable tool adoption in high-stakes settings. This paper introduces a two-phase framework combining extended domain-characterisation methods and grounded in the established design study methodologies to frame the industry-level precondition analysis from project-specific design. Validated through AI-Enabled Clinical Decision Support Systems (AI-CDSS) case studies, our approach standardises domain constraints upfront, accelerates project onboarding, and lays the groundwork for cross-project comparison for sustainable, scalable visualisation research.