There are divergent views among scholars and policymakers about the nature of permissible evidence for policymaking. It is often not feasible to construct a policy system exclusively based on objective research findings, particularly for rare diseases where conventionally accepted evidence remains a rarity. Evolutionary theories in such cases offer an overarching framework to represent the various heterodox understandings of what constitutes evidence and how evidence-based policies can be formulated under knowledge uncertainty. We conduct an empirical investigation of India’s rare disease policymaking endeavour in evolutionary perspective. The existing rare diseases policy architecture in India, in our view, reflects a ‘rationalistic’ framework. It intends to act only on ‘hard evidence’ to make, what may be called, an optimum decision, rather than initiating a ‘good enough’ policy decision based on existing (limited, soft) evidence and improving it incrementally through learning and trial-and-error. Our findings suggest that in the presence of ‘evidentiary vacuum’ and knowledge uncertainty, broadening the contours of epistemic communities, to include ‘lived experiences’ of the ‘lay’-stakeholders, can be effective in formulating an adaptive policy framework, which would ‘learn’ to better fit with the dynamic environment through inclusive deliberations, and trial-and-error.