The following is the first in NAMD’s Working Paper Series, “Advancing Medicare and Medicaid Integration.”
New guidance and assistance issued by the Centers for Medicare and Medicaid Services (CMS) are good first steps for improving access to Medicare data for the Medicaid program. However, in practice, the requirements and procedures remain a barrier for states seeking to improve care coordination and outcomes for dually eligible individuals in ways that will also better control costs. While states understand the sensitivity of Medicare data, it is also widely acknowledged that this data is critical to achieving the shared goals of the federal government and states. States are partners in the care of 9 million dual eligibles and have a program imperative to obtain and use the data. States have a programmatic right to Medicare data, and the focus should be on how to facilitate, not impede this process.
To address the ongoing operational challenges for states, three steps are needed:
- Minimize the operational barriers to access of Medicare data.
- Standardize the practice for obtaining Medicare data.
- Translate and diffuse the successful experiences with obtaining and utilizing Medicare data.
Minimize the operational barriers to access of Medicare data.
Usability. States such as Ohio have asked for and were advised that they could obtain a one-time historical “dump” of data for Medicare Parts A, B, and D – the data most critical for identifying the dual eligible population. While some states have received data, it is in a format that does not readily lend itself to manipulation. States report that the current format for the data requires costly and time consuming efforts to extract the needed information in order to identify the complete universe of dual eligibles and use it for care coordination activities. Some states report that it will take them one to two years to have useable data.
For example, Washington state has received ongoing A and B data and Part D historical data. However, each data set is in a different format. Medicare A and B ongoing data files are received in standard HIPAA transaction format which requires considerable technical skill to reformat into a usable data structure. Medicare A and B historical data comes through a different CMS organization and process, and Washington state has not yet been able to receive that file. Medicare Part D files are extremely limited and contain no claim identifiers and no payment information, making these data much less valuable in states’ care coordination activities.
The federal government must commit to work with states on a process to ensure that data can be provided in a useable format in a timely manner. Without the data, states cannot obtain a true picture of the population with whom they are dealing nor can states know with any degree of certainty the clinical needs and experience of this population.
Content. There are significant gaps in the scope of Medicare information provided to states. For example, states have reported that the Medicare Part D data for dual eligibles will not contain costs. Part D cost data is an integral component for how states will calculate the total expenditures for dual eligibles. This denies states full information about expenditures for this population and undermines Medicaid’s ability to develop clinically and fiscally sound integrated care programs. The lack of access to Medicare Part C data also creates barriers to states’ care coordination initiatives.
Cost. Historical Medicare data for the dually eligible population is an essential component for conducting trend analysis. However, the cost to obtain historical data is a barrier for states.
As policymakers seek to improve coordination for the dually eligible population, Medicare and Medicaid must create a culture of trust that facilitates the flow of information across payers. States should have access to all Medicare data sets through ResDac or another single, secure data hub, at no cost and with minimum required administrative paperwork.
Standardize the practice for obtaining Medicare data.
Standardize justifications for data. States that request Medicare data sets must submit granular justifications for the use of this information. The current practice requires states to provide a unique and specific justification for each data field that is requested. Requiring states to provide this level of detail at the front-end is counterproductive. For example, states have proposed to use Medicare data sets to develop performance and quality measures for care coordination initiatives for the dually eligible population. However, it is challenging for states to justify data fields before they have an opportunity to review and evaluate the data.
While explanation and periodic reporting may be appropriate, the current procedures are overly complex and have required states to repeatedly re-submit their requests to address additional CMS questions. States believe a standardized check list of uses for the Medicare data should be sufficient to facilitate the transfer of Medicare data sets to states. States should not be treated as researchers – requesting data only for pre-specified analytic purposes. States and the federal government are partners in the care of 9 million dual eligibles and states have a program imperative to obtain and use the data.
CMS should recognize and support other mutually beneficial uses for sharing Medicare data sets with states, including program integrity-related initiatives and savings. For example, one leader state, Washington, has recent experience in attempting to coordinate Medicare/Medicaid prescription drug claims with CMS. CMS processed an algorithm that utilized Medicaid Statistical Information System (MSIS) state data and Medicare Part D data to identify potential duplication of payments across the two systems. The initial Medicare D data set sent to the state did not include any claim identifier, making comparison and analysis difficult. CMS coordinated the effort and requested the addition of data fields that would augment the Medicare data set and facilitate the review. While the state sent data back to CMS in an attempt to match data sets and complete this algorithm, to date there has not been no further communication regarding any algorithm results.
Standardize recognized user requests. Each data set — Medicare A, B and D as well as the historical and ongoing – has a separate set of parameters and restrictions around the use of the data. Medicare A and B ongoing data have a different set of user restrictions than Medicare A and B historical, making it necessary to segregate data according to its approved use. States report that Medicare Part D prescription drug data is the most restrictive, with states required to identify every user that might access the Part D data. In Iowa, for example, this essentially requires them to identify each individual working for the Medicaid agency. Other states have been informed that for any user outside of the Medicaid agency, a separate business associate agreement and an individually signed and CMS counter-signed data use agreement is required. Business processes for managing these agreements have yet to be developed.
Streamline reporting requirements. States report that CMS requires extensive reporting concerning the utilization and experience outcomes derived from the data. For example, one state that has sought to obtain the data was subsequently notified it would have to comply with ongoing six month reporting requirements about how the data was used and results of care coordination activities.
Reporting should be periodic rather than on an ongoing basis at intervals that allow sufficient time for reporting outcomes.
Translate and diffuse the experience with obtaining and utilizing Medicare data.
The process for obtaining and utilizing Medicare data remains a “black box” for all states. Several states are farther along in this process but continue to experience barriers to obtaining the information and utilizing it for a range of purposes. Recent policy changes could ease the process for states, but currently there is no mechanism in place to translate and diffuse the experience of states in this crucial area.
Translate the experience. The federal government and states would likely benefit significantly from a resource that describes a consistent process for obtaining data. CMS could assist states by cataloguing these experiences for other states going forward and clearly identifying the successful elements of state approaches.
Diffuse the experiences. In addition to translating how to obtain the data, states are thinking about different uses or ways to incorporate the Medicare data. CMS could assist states by helping to diffuse information about how stats have manipulated the data and incorporated it into care coordination initiatives, including those related to care integration efforts for dual eligibles and health homes, outcomes from these early experiences, and other “lessons learned.”
- Standardize broader uses for data sets. States have reported that the use of the Medicare data is currently limited to “Care Coordination” and “Quality Improvement” only. States have or may propose other uses for the data, but are currently limited to how the data is to be used.