Change programs convert strategic intent into altered operating reality, yet how their success should be established remains unresolved. Evaluation practice stays anchored to delivery conformance, while program value is generated through adoption, behavioural embedding, capability accumulation and strategic repositioning occurring largely after delivery closes. This paper reviews multi-dimensional evaluation frameworks for change programs through an integrative narrative review of a consolidated corpus of applied program management, governance, assurance and sector implementation research published in the five years to 2022. Thematic coding yields five evaluation domains, namely delivery and governance performance, adoption and behavioural embedding, benefits and financial value realisation, stakeholder and workforce outcomes, and capability, maturity and sustainment, together with two cross-cutting conditions of validity, namely measurement infrastructure and causal inference. Evidence is drawn from healthcare infrastructure, energy and procurement, enterprise information technology, occupational safety, financial analytics and assurance, pharmaceutical regulation, education, aviation oversight, the built environment and agricultural development. Four weaknesses recur: adoption is systematically unmeasured, so implementation failure is misread as design failure; attribution is asserted rather than tested; sustainment is treated as a continuation of delivery rather than a distinct problem with its own decay dynamics; and measurement systems are weakly governed, so indicators used for accountability lose validity as measures. An integrative architecture of five domains across three horizons is set out, with governance conditions allowing measurement to survive program closure, and implications are drawn for theory, practice, governance and policy, together with limitations and directions for further research.