Becker's Hospital Review

22 vendor business practices, according to a 2016 AAMC study. The challenge is two-fold. For one, providers comprising ACOs use 13 different EHR platforms on average, according to the Health Affairs study. This complicates both the flow of information and the format in which it is stored. Further intensifying the issue, the AAMC study also found an inverse correlation between ACO growth and data integration. As ACO networks added new provider teams and expanded their continuum of care, the likelihood of data returning to the primary care team, which is responsible for coordinating care and planning interventions, fell off significantly. Poor interoperability between clinical sources means data acquisition processes, for most ACOs, are highly manual. "To get practice data outside the hospital, ACOs have hired a lot of [full-time employees] to do the legwork — physically going to provider groups to track down claims and clinical information," Dr. Underberger says. Maintaining the internal employee infrastructure necessary to conduct acquisition at scale across a network simply isn't in the budget for many ACOs, driving provider groups to increasingly outsource this component. 3. Time sensitivity. It is no easy task for ACOs to get the right data from disparate systems to data analysts, and then to care management teams, in a timely manner to positively affect patient care. is is especially difficult when it comes to ingesting claims data. Claims data drawn from an ACO's member population can provide a wealth of insights. EDI 837 (pre- adjudicated) claims can be particularly valuable. For a single patient-provider interaction, 837 claims include patient demographics, the services and patient's condition(s) for which treatment was administered, the provider and date of service(s). is information helps ACOs track patient care across the network to ensure patients requiring greater degrees of medical attention, such as those with chronic diseases, receive appropriate care when and where they need it. Despite claims data's value, few health systems are experienced in using it, largely because they wait (30 – 60 days) to receive adjudicated claims information from payers. By the time that data is presented to care coordinators, the window of opportunity for patient intervention has already passed or is no longer relevant. 4. Data security. Controlling access to data is essential for healthcare organizations, as some data are appropriate for one organization to see at one time, but may be illegal for another partner organization to see at a different time. Consider EHR data shared between a hospital and physician group participating in an ACO. It's helpful and appropriate to share patient information for that specific population when making healthcare decisions, such as planning interventions or offering new services. However, it is inappropriate for a hospital physician to access all of the patient records in the physician group's EHR. Protecting physicians' financial information is equally important when sharing data. "In addition to gaining physician buy-in, you need to have physicians' confidence that you have the appropriate data security policies in place," Dr. Underberger says. "Some independent practices are reluctant to share their claims information with other institutions because of the risk their billing information may be exposed or leaked." 5. Data aggregation and traceability. Capturing data is the first step toward evidence-based medicine. But data is only as helpful as the insights it yields, and it oen requires a significant amount of rework before it can be combined and used to make clinical, operational or business decisions. While aggregating information is key for analytics, being able to trace the results back to the original data source is imperative for creditability. Different types of information exchanged for ACO care coordination may require the use of different types of documents. For example, a behavioral health assessment may be captured and sent as a PDF, a nurse's patient notes may be stored as unstructured free-text, and a laboratory result may be sent in a HL7 message or in a cCDA (via an EMR). As is, data scientists at health systems spend a substantial amount of time making data usable. e granular nature of the job — what data scientists call "data wrangling" or "data janitor work" — requires enormous labor and cost. Data scientists report spending up to 80 percent of their workday collecting and preparing unruly, disparate healthcare data, according to a survey conducted by e New York Times. at leaves only 20 percent of the day to use and analyze data to drive insights that bring positive change to organizations. To help support data analysts and acquisition teams, ACOs see value in partnering with third-party vendors. Advantages of partnering with third-party vendors ACOs are strapped for resources when it comes to supporting data analytics. ACOs reported spending an average of $600,000 on operating expenses for health IT, analytics, and reporting, which is relatively low compared to the reported average investment of $1.1 million on care management, according to Health Affairs. "Our clients say their biggest problem is internal IT staff being totally overwhelmed with the task at hand and in need of help," Dr. Underberger says. "Most of these organizations only have a handful of data analysts, and the data acquisition teams they do have are mostly concerned with their own network facilities. ey haven't even begun wrapping their arms around what it takes to process external data. at's why many ACOs are going to third-party vendors [for support]." Change Healthcare helps ACOs prepare their data for improved risk management. e IT and consulting company is in the process of building and rolling out a data-agnostic platform with security policy infrastructure that accepts data from any source, including individual physician data and claims information from a multispecialty out-of- network facility. e platform then aggregates and normalizes the data, applies patient identifiers to map and trace patients across discrete IT systems, and enforces rules to improve data presentation for end users. In addition to helping ACOs establish secure, standardized data processes, Change Healthcare offers analytics support services to better mine data for actionable health insights. "We have an analytics team that helps ACOs analyze their own data in a way that is more visual and meaningful," Dr. Underberger says. "For instance, we are developing a measure service engine that calculates a patient's quality measure compliance score based on the type of data we receive. In addition, our analytics team can help configure the solution to attribute these patients to the most appropriate providers within various practices." Conclusion In an ACO, several organizations practicing along the continuum of care capture various types of information, in distinct systems at different times. Organizations need a defined plan to put their data to work across the enterprise if they are to survive in the challenging healthcare landscape. is includes building the skills and processes to transform raw data into information that drives earlier patient interventions and better health outcomes. Understanding known data barriers and viable IT management solutions are the first steps in choosing the right analytics strategy to achieve specific care coordination objectives. ird-party vendors can be extremely valuable partners in helping ACOs control analytics and better manage risk by supporting manual data acquisition processes, overcoming interoperability challenges, unclogging the flow of information, ensuring security compliance and supporting data aggregation and lineage traceability. n SPONSORED CONTENT

• Cover