Section 1115 demonstrations give states flexibility to test new approaches to improving health care access, quality, and outcomes for Medicaid beneficiaries. Evaluating these initiatives often requires bringing together information from multiple agencies and data systems. While Medicaid claims and encounter data are often the foundation, states frequently need information from behavioral health agencies, provider licensing systems, public health surveillance programs, workforce databases, and other administrative sources for a more complete picture of program performance.

As federal guidance increasingly emphasizes early evaluation planning, implementation monitoring, and continuous learning, strong data integration has become critical. The need for cross-agency coordination has grown as states pursue demonstrations that extend beyond traditional health care delivery. For example, newer Medicaid reentry demonstrations require collaboration among Medicaid agencies, departments of corrections, behavioral health programs, and community providers, each of which may maintain separate data systems.

The goal of data integration is to produce information that states can use to manage programs, monitor implementation, and understand whether reforms are achieving their aims. Policymakers and program leaders need evidence to understand whether services are reaching intended populations, whether provider capacity is meeting demand, and whether reforms are producing meaningful improvements. The strongest evaluations will connect information across systems in ways that help states understand what is working, where challenges remain, and where course corrections may be needed.

Bringing together data from multiple systems is often one of the most challenging aspects of a Section 1115 evaluation, requiring planning, coordination, and collaboration across agencies that collect and use data for different purposes. Below are common challenges states and evaluators encounter when working across multiple data systems, along with practical considerations for building a strong foundation for evaluation from the outset.

States need information from multiple systems to understand whether a demonstration is being implemented as intended and achieving its goals. Medicaid claims and encounter data can show who is receiving billable services and how utilization is changing over time. However, understanding changes in provider capacity, implementation progress, or community need may require information from behavioral health agencies, licensing systems, program records, public health surveillance data, and other administrative sources.

Together, these sources provide both measures of outcomes and the context needed to interpret them, creating a more complete picture of how a demonstration is functioning and where challenges or successes are emerging.

For example, in a state implementing reforms to expand access to substance use disorder treatment, Medicaid claims may show increases in treatment utilization, but they cannot explain whether growth was driven by new provider capacity, expanded outreach efforts, changing population needs, or other factors. By combining claims data with provider licensing records, implementation tracking data, and public health surveillance information, evaluators can move beyond describing what changed to developing a better understanding of why those changes occurred.

Many states have invested substantially in data collection and reporting systems, but those systems are often maintained by different agencies for different purposes. Medicaid, public health, behavioral health, workforce, corrections, and licensing systems may all contain valuable information, yet use different identifiers, reporting schedules, and governance processes.

This challenge is particularly apparent in newer demonstrations, such as Medicaid reentry initiatives, where evaluators often need to bring together information from correctional systems, Medicaid enrollment and claims data, behavioral health programs, and community providers. Because many of these systems were developed to support specific programmatic and operational needs, additional coordination may be required for evaluation purposes.

Documentation gaps, inconsistent definitions, and difficulties linking data across systems can create challenges for evaluators. To avoid delays, states and evaluators should begin planning for data integration early in the evaluation process by identifying key data sources, engaging data stewards, establishing data-sharing agreements (which are necessary for protecting enrollee data), documenting measure definitions, and assessing data quality at the outset.

These activities also help ensure that evaluation questions are aligned with available data and that agencies have sufficient time to prepare data extracts and reporting processes. In many demonstrations, the success of the evaluation depends as much on early planning and coordination as it does on the analytic methods used to answer the research questions.

Independent evaluators are in a unique position to work across agencies and programs to synthesize disparate data sources. Building relationships with data stewards helps evaluators understand the strengths and limitations of individual datasets, protect enrollees’ identities, identify issues early, and develop realistic data collection and reporting plans.

Equally important, evaluators should be mindful of agency burden—many state staff support ongoing operations, federal reporting, performance monitoring, and evaluation activities simultaneously. Coordinating data requests and making use of existing reporting processes can reduce burden while still generating the information needed to answer key evaluation questions.