In late 2025, the Centers for Medicare & Medicaid Services (CMS) released updated guidance for Section 1115 demonstration evaluations. The guidance raises expectations for rigor, transparency, theory-driven design, and policy relevance across the full evaluation lifecycle.

Importantly, the updated guidance emphasizes the value of early evaluation planning and alignment of design decisions with CMS expectations to help states avoid costly rework, meet timelines, and produce meaningful findings. Drawing from NORC at the University of Chicago’s applied evaluation work on Medicaid Section 1115 demonstrations, below is a summary of key aspects of CMS’s updated guidance and considerations for states undertaking demonstration design or renewal. This summary reflects the updated CMS evaluation guidance and does not represent NORC’s positions or policy endorsements.

The updated CMS guidance underscores the importance of engaging an experienced evaluation contractor, such as NORC, before waiver implementation. Early collaboration between states and evaluators supports the joint development of logic models, research questions, data strategies, and feasible comparison group approaches that align with both policy goals and operational realities.

These steps require careful consideration of opportunities and limitations, with equal focus on policy impact and feasibility. This early investment can mean the difference between a seamless evaluation process and multiple rounds of re-work risking timeline, budget, and compliance with CMS requirements. Engaging evaluators early reduces the risk of proposing designs that later prove infeasible due to data limitations, timing constraints, or misalignment with implementation plans, and supports stronger coordination between evaluation and implementation teams.

CMS’s updated guidance moves away from driver diagrams and toward conceptual frameworks or logic models that articulate how a demonstration is expected to produce change. Compared to driver diagrams, which often list high‑level activities and outcomes, logic models explicitly map policy mechanisms, short‑term and long‑term outcomes, contextual factors, and potential unintended consequences.

For example, a waiver focused on behavioral health integration might theorize that care coordination and workforce enhancements will lead to improved outpatient engagement, reduced emergency department use, and longer‑term cost reductions. Contextual factors, including provider capacity, rural access barriers, or workforce shortages, may influence whether these mechanisms operate as intended. Without a clear conceptual framework at the outset, evaluation teams may later discover that key outcomes, mediating pathways, or contextual variables were not measured, potentially leading to conclusions that underestimate a demonstration’s true impact.

In updated guidance, CMS charges states to employ data sources that go beyond Medicaid claims when appropriate, including Supplemental Nutrition Assistance Program (SNAP), Temporary Assistance for Needy Families (TANF), child welfare data, all‑payer claims databases, and national surveys. These data present valuable opportunities to broaden research questions and capture cross-system impacts, but they require substantial early planning to secure data use agreements and establish record linkages.

CMS also highlights the importance of planning for subpopulation analyses and being transparent when sample size or data limitations constrain causal estimation. Addressing these considerations during the design phase helps ensure that evaluation plans are both comprehensive and feasible.

CMS continues to emphasize that 1115 evaluations must explain not only whether observed outcomes changed, but why they did or did not change. The updated guidance goes further by providing more detailed and explicit expectations for implementation research questions, signaling that understanding how a demonstration operates in practice is now a core evaluation requirement. This requires careful attention to how policies were implemented on the ground and whether implementation aligned with stated intentions.

For example, the same care management model may produce very different results depending on whether it is embedded within managed care organizations, operated through provider‑led entities, or administered centrally by the state, with each approach carrying distinct implications for provider uptake, beneficiary engagement, and consistency of service delivery.

Understanding these distinctions requires integrating quantitative outcome analyses with qualitative information from providers, beneficiaries, and other stakeholders. Working in partnership with the state, an evaluator can apply this framework to integrate quantitative outcomes with qualitative insights, providing a clearer picture of how and why a waiver performs differently across settings.

The updated guidance strengthens expectations for cost analyses by emphasizing transparency and alignment with the demonstration’s conceptual framework. CMS recommends analyses that examine specific pathways through which demonstrations may affect spending, such as reductions in avoidable hospitalizations or emergency department visits, improvements in chronic disease management, or better care transitions.

Evaluations should clearly state cost assumptions, explain observed cost drivers, and distinguish between service expenditures and administrative costs (such as salaries, contracts, monitoring, and evaluation) to ensure findings are relevant for waiver renewals, scaling decisions, and broader Medicaid policy.

CMS guidance indicates that interim and summative evaluation reports should include actionable lessons learned and policy implications intended to inform future decision‑making. Rigorous mixed‑methods evidence is critical because it connects what the demonstration achieved with the implementation challenges and contextual factors that shaped those results, helping CMS and states understand how waiver design, implementation strategies, and stakeholder engagement should differ across settings or subpopulations. Together, these insights inform Medicaid policy, targeted supports, and resource allocation in future demonstrations.

CMS’s updated guidance includes specific expectations for Section 1115 reentry waivers focused on improving care transitions for individuals leaving correctional settings. Although the population and service systems are distinct, the reentry evaluation guidance reflects core principles of the broader evaluation framework. CMS expects reentry evaluations to present clear logic models linking pre-release services to post-release outcomes, incorporate mixed methods to examine implementation and context, and plan early for cross‑system data integration across correctional, Medicaid, and community‑based systems.

Attention to population‑specific contextual factors, such as continuity of care, access barriers, and unintended consequences during reentry is essential for producing accurate, policy‑relevant findings. Early coordination is particularly critical for reentry demonstrations given the complexity of linking data across systems and addressing barriers unique to criminal-legal system‑involved populations.