Evidence-based medicine goes beyond choosing the right clinical pathway, prescribing the right medication, or following procedural standards. Evidence-based decision-making also extends to targeting the right patients at the right time, employing preventive programs, and using diverse clinical teams to serve patient populations.
Reviewing Medicare claims data is a valuable first step for accountable care organizations (ACOs) that are asking themselves: “Where do we begin?” Using claims data made available through the Medicare Shared Savings Program (MSSP) will arm providers with the information they need to make critical decisions about delivering high-value care to their beneficiaries.
New ACOs are inundated with many activities, from strategic decisions, like organizing the board, to technical requirements, such as updating the website. But managing and using data should also rank highly on one’s list of priorities.
The Centers for Medicare & Medicaid Services (CMS) provides ACOs with robust, claims-level information about their assigned population of beneficiaries. This information includes beneficiary-level Medicare claims for hospital services, physician services, post-acute care, and other covered Medicare services, and enables ACOs to evaluate beneficiary costs across the continuum of care.
ACOs receive monthly Claims and Claims Line Feed Files (CCLF) from CMS for assigned beneficiaries, and quarterly reports, which include an updated ACO benchmark, list of attributed beneficiaries, and expenditure/utilization reports. Data and information files are securely delivered through an online portal. Given the volume of data, and the fact that CMS includes beneficiary-level detail, the data files can be quite large and complicated. To use the information effectively, organizations will need staff or external expertise with basic data science experience.
Three Things to Do with Data as a New ACO
Getting started with something new is often the most difficult step. Here are suggested first steps for ACOs new to managing and using claims-level data for Medicare beneficiaries. Each of these steps is a high-impact activity.
1. Normalize and organize. There’s no way to sugarcoat the fact that the data files from CMS are complicated. They will take time and experience to organize. For the lay-person without claims data management skills, this task will be difficult.
We suggest using—or hiring—a skilled data analyst(s) that understands data science and has experience working with claims data. This person (or team) will need to merge different datasets, organize the combined data so as to easily update each month, identify data issues, and more.
It can be quite cumbersome and time intensive to “extract, transform, and load” (ETL) the numerous claims data files month after month. Adopting a method to automate these procedures and streamline this process is vital. Ultimately, the data analyst(s) must be able to query this robust dataset based on the questions you want answered.
A data team with healthcare experience is essential. Such a team can flag data abnormalities and errors, and merge disparate datasets (from a variety of data sources) to create meaningful metrics. Having someone with knowledge of these various data sources, along with plans for best integrating them with the claims data, is key to harnessing the power of claims data.
2. Create a “Top 25” beneficiaries list. Everyone can appreciate a “Top X” list—it is easy to understand and creates a sense of urgency. We recommend every ACO create a list of the 10 to 25 highest utilizers for each ACO participant. This list, which can be based on a number of factors, will serve as a simple, introductory risk-stratification tool for each participant within the ACO. We suggest including each beneficiary’s name, hierarchical condition category (HCC) score, annual Part A and B expenditures, and percentage of those payments made to ACO participants.
Using this list, you can start to ask questions like, “Are 20 percent of my beneficiaries accounting for 80 percent of all expenditures?” This list may also serve as justification for ACO-wide education, and a review of HCC coding and documentation best practices.
3. Create a simple dashboard. Once the data is clean and organized, and you’ve created some basic lists, we suggest developing a simple dashboard. At this point, the ACO board, management team, and providers should have a sense of the data capabilities.
We suggest allowing the relevant ACO board committee, with input from other ACO stakeholders, to choose the most important metrics that should be tracked in the early phases of operation. This committee should be accountable for ACO financial performance; creating a dashboard is a good step toward evidence-based decision-making.
Basic metrics may include: expenditures by beneficiary type (end-stage renal disease, dual, disability, aged), expenditures by type of services, expenditures by ACO participants compared to expenditures by non-ACO participants (i.e., network leakage), and emergency department admissions. Separate dashboards can be created for each participant/taxpayer identification number, supporting transparent and comparative culture.
It is normal for the dashboard to change over time. While some metric consistency is important to identify trends, don’t be afraid to add new metrics or remove those that aren’t beneficial. Every ACO will have different priorities and opportunities for improvement; these dashboards should reflect those priorities.
Conclusion
These three steps are almost universally applicable to ACOs that are new to working with claims data. Organizing and cleaning the data is necessary for anyone wanting to make the data meaningful. Creating “Top 25” lists and developing simple dashboards are attainable first actions; both provide meaningful information, and push participants to start asking more advanced questions of the data.
An ACO’s quality and finance committees will be charged with various responsibilities to improve quality while lowering the cost of care. Using the Medicare claims data to its potential is one tactic to help any ACO reach these goals.
When working with your Medicare claims data, there are a few additional items to consider:
- Be vigilant for data errors and abnormalities (both from CMS’ files and your outputs from the data). Mistakes can happen, so implement steps to quickly identify and mitigate these potential errors.
- Consistently review data with participants to improve accountability. Creating a routine for reviewing updated information ensures that the reports are used to their fullest extent for evidence-based decision-making.
- Expect growing pains. Do not be afraid to experiment and improve your processes, the questions you ask of the data, and the format of your outputs. Conquer the basics, then start incorporating more complex analyses.
- Don’t forget about data privacy, security, and use requirements! As with any patient-level health data, it is important to follow all HIPAA (Health Insurance and Portability and Accountability Act) and other applicable policies and procedures to ensure proper data privacy and security. Note that CMS also has specifications under its Data Use Agreement that every ACO must sign. Make sure you’re complying with all terms.
Aaron Elias, a healthcare consulting manager with professional services firm PYA, helps healthcare providers navigate the transition from volume-based to value-based payment and delivery models. Jason Hardin, manager, healthcare consulting, PYA, contributed to this article.
Healthcare Informatics’ “Industry Voices” articles provide a platform for industry experts to weigh in on the latest healthcare IT trends and best practices. All Industry Voice submissions (submit here) are subject to editorial approval and cannot include explicit mentions of vendor products. More information on our submission guidelines can be found here.
