Value-based care is defined by quality (numerator) over total cost of care (denominator) by the National Academy of Medicine. Houston Methodist Coordinated Care Accountable Care Organization is a Medicare ACO responsible for more than 52,000 traditional Medicare beneficiaries attributed to approximately 300 primary care physicians within its high-value primary care network within the greater Houston area.

One of the key challenges the ACO faced was providing real-time feedback to the physicians on their CMS ACO quality performance. If primary care physicians had this information, they would be able to identify and close quality gaps in real time at the patient visit, said Dr. Julia D. Andrieni, senior vice president, population health and primary care, at health system Houston Methodist, and president and CEO at Houston Methodist Coordinated Care Medicare ACO and HM Physicians Alliance for Quality.

She also is an associate professor of clinical medicine at Weill Cornell Medicine and an adjunct associate professor of medicine at Texas A&M College of Medicine.

A standardized approach to support excellence

"By identifying and bringing awareness to potential quality gaps at the point of care, physicians were able to improve depression screening, fall risk screening, immunizations, and appropriate cancer screenings," Andrieni explained. "Our goal was to provide a standardized approach to support excellence in quality by developing an ACO quality metrics dashboard integrated in the EHR that updates in real time.

"Each physician has a visual of the open quality gaps at the point of care and can drill down to the patient level or drill up to their aggregate performance in each quality metric," she continued. "In addition, this aggregate CMS ACO quality gap information enabled physicians to send bulk EHR patient portal messages for specific quality metrics with the 'why' this metric is important to improve health."

This CMS ACO quality metric dashboard gave the Houston ACO team a deeper understanding of the challenges and complexities of each CMS ACO quality metric. The ACO team could develop and implement customized workflows for practices with individualized patient outreach to tackle specific population needs. The hope was that the ACO team could mine ACO patient quality data to identify opportunities to improve care.

Another key challenge was how does one identify the "impactful" patients who need additional resources when caring for more than 50,000 patients within a program?

"One way in which we solved this puzzle was in working with a vendor who developed a predictive analytics tool and integrated within our EHR for transparency with the medical team," Andrieni noted. "The data used to develop the predictive analytical tool included Medicare claims data, social determinants of health, real-time EHR data of admissions and ED utilization, Zip code demographics, ethnicities, and behavioral health assessments.

"Three risk scores were developed to include the risk of admission/ED utilization, complex care risk score (patients with multiple comorbidities), and advanced illness care risk score for patients with end-of-life choices," she added.

Patient care programs

Nursing teams developed patient care programs for high and rising-risk patients in each of the risk scores. For patients admitted to a hospital, the nursing transition in care team outreached to patients identified as high and rising-risk and followed these patients for 30 days post-discharge to review medications, post-discharge appointments, and patient empowerment/education.

This program's primary goal is to prevent unnecessary readmissions and ED utilization, if possible. A nursing team for complex care patients was developed with the goal of preventing avoidable admissions by outreaching and following patients identified by risk scores for two to three months to provide resources, education, and follow-up appointments and tests.

"Our nursing team for advanced illness care are nurses with palliative care experience who can navigate end-of-life issues and honor patients' choices," Andrieni explained. "By identifying and understanding which patients would benefit from specific nursing interventions, the ACO can budget for staffing and resources to support patients identified as high risk.

"All ACO nursing programs have the goal to improve quality of care with the byproduct of a reduction of cost from preventing an admission, readmission or ED visit," she continued. "In value-based care programs, shared savings are realized by demonstrating the highest quality at appropriate cost."

Understanding ACO-observed outcomes

Currently, Houston Methodist also is using another method to identify high and rising-risk patients by "twinning" or matching complexity of patients to similar patients within the Medicare population with claims data to understand the ACO-observed outcomes for admissions, ED utilization, costs per beneficiary per year, utilization of post-acute facilities, readmissions, and utilization of other health resources to the expected outcomes for a similar population.

This type of comparison gives staff insights into ACO performance compared with national data and also insights into the individual team members' performance compared with other team members.

Before the implementation of the quality dashboard, the vision was that this tool, which would be integrated within the EHR, would engage and support busy clinicians at the point of care. With multiple competing priorities at the point of care, staff wanted to facilitate an easy, unobtrusive way to point out gaps in quality of care so each patient visit was an opportunity to improve care.

"It was important to obtain physician buy-in and feedback in working with IT to implement a usable tool," Andrieni stated. "In addition, the ACO team with our IT team needed to develop reports for tracking and trending quality from this dashboard. Further, we needed to be able to meet CMS requirements for submission of quality data with easy abstraction and submission, which could be scalable and sustainable.

"Part of the implementation would be educating clinicians of the discrete fields to place specific data in order to receive credit," she added.

Closing quality gaps

Clinicians use the quality dashboard integrated in the EHR at the point of care to close quality gaps. The Houston Methodist ACO leadership team pulls monthly reports for individual physicians with their patient list and quality gaps remaining open.

Physicians have found these reports helpful in managing their patients and can pull these reports themselves. The analyst team audits the quality reports and reviews workflows to be sure the quality metrics are placed in the appropriate discrete fields in for the data to be captured in the dashboard and in the reports.

"An executive dashboard was developed to aggregate all ACO quality metrics for overall ACO quality performance with drill-down to the practice level, physician level and patient level," Andrieni noted. "Nursing teams use this dashboard when outreaching to patients enrolled in clinical programs to identify quality gaps in care.

"Our population health advisors review practice and individual physician quality performance with the goal of strategizing with each practice and physician to achieve the 90th percentile nationally in CMS ACO quality metrics," she continued. "Our population health advisors also review primary care practice workflows to look for opportunities to improve closing quality gaps of care when a particular quality metric is lower than expected."

Big results

From 2017 to today, Houston Methodist Coordinated Care Medicare Accountable Care Organization has achieved the 90th percentile in CMS Medicare ACO quality metrics to include tobacco screening and cessation, influenza immunization, fall risk screening, depression screening, hypertension control, diabetes control, and breast cancer and colorectal cancer screening.

Most recently, Houston Methodist Care Coordination ACO received from CMS 99 quality points, which included quality performance, patient satisfaction and clinical data integration. In the initial year of the ACO (2017), it performed in the 40th percentile in some of the CMS ACO quality metrics, which included depression screening and fall risk screening.

"By bringing awareness to our clinicians, to our ACO team, we were able to build workflows and infrastructure to support tracking, trending and feedback to our clinicians," Andrieni said. "Physicians also received CMS shared savings based on their individual quality percentile performance.

"HMCC ACO was No. 3 nationally for savings per patient and No. 13 nationally for earned shared savings from 482 CMS Medicare ACOs nationally," she continued. "In addition, HMCC ACO beat the CMS national average quality percentile for each CMS ACO quality metric."

Advice for other ACOs

In the lessons learned from this journey, it is important to know stakeholders early in the process and engage and involve them in the decision making for choosing a particular vendor, Andrieni advised.

"This work would not be possible without an engaged IT team, an engaged ACO team, engaged clinicians, and health system support," she said. "Infrastructure and investment in technology is necessary before the return on the investment is realized and before improvement in quality is trended. Change management is important in the implementation of any new technology, which requires listening, patience and compassion."

When choosing a technology vendor, Andrieni said look for a true partnership. Here is her advice:

• Ask the vendor to share examples with outcomes for the type of work you want to accomplish with dashboards and/or predictive analytics.
• Ask about annual updates to the vendor's predictive analytical tool.
• Ask for evidence the vendor is identifying the right patients to drive the desired outcome. Speaking with vendors' clients is also important to understand the vendor's expertise and to identify any blind spots.

Follow Bill's HIT coverage on LinkedIn: Bill Siwicki
Email him: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.