A Selected Case Study in Healthcare Innovation and Transformation
• Create a streamlined process for accountable care organizations’ (ACOs) annual mandatory reporting to the Centers for Medicare & Medicaid Services (CMS) on 33 nationally recognized clinical quality measures.
• Maintain status as an ACO, while gaining valuable insights from clinical data obtained.
• Map clinical data from multiple electronic medical records (EMRs) into a single cloud-based dashboard providing real- time outcomes for clinical quality measures across the ACO.
• Obtain powerful information from a dashboard to promote population health management (PHM).
Although the Affordable Care Act (ACA) does not formally define the term “population health,” CMS defined it within the context of ACOs. While it is still unclear how ACO care delivery systems will intersect with a broader community definition of PHM, what has become clear is the value of clinical data to connect the dots in integrating, measuring and improving the health of a population.
This is reflected by the mandatory ACO reporting to CMS on 33 nationally recognized clinical quality measures in the domains of patient/caregiver experience; care coordination/patient safety; preventive health; and at-risk populations for diabetes, hypertension, ischemic vascular disease, heart failure and coronary artery disease.1 Thus, ACOs have come face to face with employing clinical data in PHM.
Three Medicare Shared Savings Program (MSSP) ACOs in distinct settings—Indiana Lakes Quality Partners (ILQP), Northwest Physicians Network/Rainier Health Network in Washington State (NPN) and Cumberland Center for Healthcare Innovation in Tennessee (CCHI)—found innovative ways of leveraging clinical data for PHM.
The collaboration of these three ACOs began due to their shared use of an ACO vendor, Clinigence, to achieve the mandatory annual ACO quality measure reporting. The three ACOs represent a diverse collection of hospital, health system and primary care ACOs and at the time of their introduction to one another in the fall of 2013, they represented more than 50,000 MSSP ACO lives, with an additional 50,000 commercially and self-insured beneficiaries.
From a population health perspective, the ACOs’ stories were much the same in that they were each starting at ground zero. For example, the ambulatory care physicians had no idea how many diabetics were in their practices, much less a registry of who these patients were and their respective outcomes. In addition, each of the ACOs had multiple different outpatient EMRs across both owned and independent practices, which greatly complicated the task of compiling clinical quality metric reports.
Adding to the mayhem were physicians in the ACOs who did not have a clear understanding of the ACO reporting measures by which their quality would be scored. Underpinning the confusion was inaccurate patient attribution logic utilized by CMS, which made it difficult to assign the correct patient to the correct primary care provider.
To develop a streamlined reporting process, the Clinigence tool ingested the list of randomly selected annual reporting beneficiaries as provided by CMS, populated each beneficiary’s profile with the measures that applied to them and consumed claims data for the same beneficiaries in order to pre-populate the tool with as much data as possible and reduce the time and burden of ACO reporting.
After CCHI adopted the strategy, the other two ACOs followed suit by achieving 100% submission of data for all 616 assigned beneficiaries per measure (the minimum requirement is 411) in their first years of reporting—an accomplishment they credited to the ACO reporting tool. This is a valuable achievement in light of the fact that five MSSP ACOs were unable to complete the mandatory reporting in 2013, and two of the five forfeited shared savings as a result of their inability to complete reporting.2
CCHI found an additional benefit of the reporting tool by using it to gauge quality improvement throughout the year— something the ACO accomplished by populating the tool with the same ACO patients that were required for CMS reporting and resubmitting it to providers for attesting to updated clinical data part way through the reporting year.
This semi-annual reporting strategy identified deficiencies in clinical documentation processes, as well as detected practices that were falling behind in terms of quality improvement. It also reminded the practices of ACO measures that CMS would be using for quality scoring. Both CCHI and ILQP found that engaging the practices in reporting for their own populations had tremendous value in alerting practices to areas of quality improvement.
The new strategy also mapped clinical data from multiple EMRs into a cloud-based dashboard that provided real-time outcomes for the clinical quality measures across the ACO. This relieved the pain point of inaccurate patient attribution because logic was supplemented with the EMR location in which the greatest amount of clinical data was found. In other words, clinical data could trump claims data in assigning a patient to a specific primary care physician.
“The result was a dashboard of real- time clinical data that allowed the ACO to view how each of the practices were performing on the ACO measures.”
ILQP began the mapping process by identifying practices most likely to benefit from automated data extraction, practices with a locally hosted EMR and those with more complete clinical documentation methods. A clinical data extraction application was installed on the EMR servers and ran at regularly scheduled intervals to automatically extract clinical data that applied to the ACO measures, including vital signs, lab test results, immunizations, diagnoses codes, allergies and medications.
The result was a dashboard of real-time clinical data that allowed the ACO to view how each of the practices were performing on the ACO measures. Ross Family Medicine of CCHI, who began the EMR mapping process one year prior to ILQP, found that for 2013 ACO reporting 18 of the 22 clinical measures were automatically extracted and thus populating the CMS reporting registry. This greatly decreased the time required for the ACO reporting process.
Results: Once the ACOs were able to pull clinical data into a dashboard from multiple practices and across various EMRs, they had powerful information to promote PHM. Examples of how this data was leveraged included physicians using the diabetes composite measure data as a registry of not-to-goal diabetic patients to coordinate with both onsite and centralized practice care teams for outreach and follow up.
In terms of broadening the onsite practice care teams, Ross Family Medicine found that it could easily pay for a full-time embedded care coordinator based on using the dashboard to identify and follow up on gaps in care. Shipshewana Family Medicine of ILQP employed a full-time registered nurse health coach to coordinate a team-based approach that used the dashboard to supplement identifying gaps in care for both chronic and preventive measures. The practice simultaneously applied standing orders for preventive care, such as mammograms, flu shots and colonoscopies. This off-loaded the burden of care coordination from the physician and engaged multiple practice team members in PHM.
ILQP provided each of its primary care providers with access to the dashboard clinical data, which resulted in improved engagement of the physicians in PHM. For example, a physician work group at ILQP created an ACO-wide strategy for quality improvement of known hypertension patients. ILQP found that 40% of its diagnosed hypertensive patients were not to goal, meaning that their most recently obtained blood pressure readings were above normal. Measuring patients’ blood pressure at both the beginning and the end of every patient appointment, a pilot practice increased its performance from 58% to 67%—the ACO set a goal of 66%—between April and November of 2014 by implementing this strategy alone.
Since the inception of the ACO collaborative, it has grown from three initial ACOs to 10, and represents more than five million beneficiaries across nine states. In terms of decreasing costs, one of the initial three ACOs achieved more than $4.5 million in savings in 2013. Two of the ACOs were recently awarded lucrative commercial payer contracts. On target for 2015, the ACOs plan to leverage the clinical data in the dashboard to identify patients most likely to benefit from the new CMS Chronic Care Management Medicare Physician Fee Schedule for 2015, outlined on Oct. 31, 2014.3
· One hundred percent automation of EMR clinical data for PHM is impossible; ACOs must have a backup tool in place for manual attestation.
· Engaging practices in reporting for their own populations promotes innovation and quality improvement.
· Assessing all the determinants of health for a given population is at the intersection of accurate clinical and claims data.
· Patient attribution logic can be supplemented with EMR data to assign patients to the correct provider.
· Automating the aggregation of the greatest amount of clinical data possible identifies trending and promotes physician engagement.
1 “Guide to Quality Performance Scoring Methods for Accountable Care Organizations.” Medicare Program; Medicare Shared Savings Program: Accountable Care Organizations; Final Rule, 76 Fed. Reg. 67802. Nov. 2, 2011.
2 Petersen M, Muhlestein D. “ACO Results: What We Know So Far.” HealthAffairs Blog. May 30, 2014.
3 “Fact Sheets: Policy and Payment Changes to the Medicare Physician Fee Schedule for 2015.” CMS.gov. Oct. 31, 2014.