Exploring Best Practices in Serious Illness Care for Medicare Advantage and Accountable Care Organizations

By Allison Silvers, Torrie Fields, Florence Kariuki, Dana Lustbader, and Ana Tuya Fulton, Center to Advance Palliative Care


Health plans and accountable care organizations (ACOs) have the opportunity to implement specific strategies to ensure high-quality and cost-effective care for people living with serious illness. To accelerate adoption of these strategies, the Center to Advance Palliative Care (CAPC) launched the Medicare Advantage (MA) and ACO Learning Communities, providing education and peer support to expedite needed changes and expand access to palliative care among high-need populations. Forty-three organizations participated, covering over three million Medicare beneficiaries across 32 states.

Following are summaries from the Learning Communities’ experiences with four recommended strategies for improving serious illness care, as well as resources that can be used by health plans and ACOs looking to drive better value and improve quality of life for those living with serious illness.

1. Proactively Identify the Population in Need of Palliative Care

While many people in the U.S. live with a serious diagnosis, a diagnosis alone does not indicate need for palliative care. The definition of serious illness includes “either a negative impact on a person’s daily function or quality of life, or an excessive strain on their caregivers” — information not often found in claims or electronic medical records. Thus, finding the population in need of palliative care requires a proactive and creative approach.

The Learning Communities confirmed that diagnoses, while the right starting point, need additional variables to narrow down the high-need, at-risk population. To start, most organizations use a “short list” of diagnoses, which includes: metastatic cancer, COPD with oxygen, heart failure (with ejection fraction requirements), renal failure, advanced dementia, and six or more chronic conditions. Patients with high disease burden and multiple chronic conditions can also be identified with tools like the Charlson Comorbidity Index.

Most Learning Community participants agreed that adding in an unplanned hospital admission or recent emergency department visit within the past three months appropriately narrows down the population for targeted additional assessments and conversations. This approach is particularly useful since studies have shown that patients and families are more likely to engage in additional supports during a crisis. Interestingly, most Learning Community participants felt that polypharmacy — while it represents a medical risk factor — is not a useful variable to identify a seriously ill population, because so many Medicare beneficiaries are prescribed 10 or more medications.

A few organizations, particularly those eager to identify at-risk individuals prior to unnecessary utilization, use indicators of frailty to narrow the population. This approach is now even more viable, with recent work by the National Committee for Quality Assurance to define value sets for frailty symptoms, encounters, diagnosis, and devices. While earlier recommendations had suggested using both utilization and frailty variables to identify the population with serious illness, some are now raising concerns that this approach is too specific and unduly narrows the population to be assisted.

In terms of non-data-driven strategies, many were skeptical that clinician opinion was a reliable mechanism to identify the population in need of palliative care. Instead, the emerging best practice involves clinician review of the data-derived patient lists. More recently, a promising new approach is to use self-reporting for proactive identification by asking the patient: “How likely do you think it is that you’ll find yourself in the hospital or emergency room in the next month or two?” A patient reporting a high likelihood may be driven by insufficient caregiver support, symptoms, or a fragmented and unreliable health care system. Lastly, broadening the sources of member/patient identification to include referrals from behavioral health, pharmacy, and community health workers can expand opportunities to collect valuable subjective information.

The final advice? Start somewhere, and as you learn more about your members/patients with unmet palliative care needs, continue to adjust your predictive models and processes. 

2. Build Additional Care Management Capabilities

Once a member/patient has been identified, the next steps are to assess for unmet palliative care needs, establish trust, and hold meaningful patient-centered conversations to clarify expectations, options, and preferences. Successful organizations use their existing care managers for these responsibilities. The Aetna Compassionate Care Program is one of the early pioneers in supportive care management for people living with serious illness, and their approach still shows significant benefits.

Success lies in changing the role of care management — turning the focus from treatment adherence to uncovering and addressing potential suffering, and ensuring that any treatment is aligned with the patient’s goals. Care managers must be on the lookout for patients in need of palliative care services, and skillfully refer when warranted.

Faculty and participants of the Learning Communities highlighted the following:

  • While care managers often conduct comprehensive assessments, issues such as symptom burden and caregiver burden are less frequently assessed, yet are essential to identify gaps and risks for people living with serious illness. Functional and cognitive assessment must be included if not done already.
  • Holding meaningful conversations with patients and families living with serious illness is a skill that must be taught and practiced, and investments in staff training, role play, mentoring, collaborative case reviews, and other supports are needed to ensure competency. Care managers working with this population should be trained in the expected progression and likely symptoms for the more common diagnoses they will be working with.
  • Serious illness conversations are emotionally challenging work. Organizations should consider case conferences, buddy systems, and other emotional supports for their care managers to avoid burnout.

Creating a dedicated team to work with patients with serious illness poses pros and cons that are important to consider (see table below). Even if the organization creates a dedicated team, Learning Community faculty recommended that all care managers have ongoing training in having serious illness conversations and recognizing unmet palliative needs to enable successful warm transfers.

3. Support Home-Based Care for People with Serious Illness

A subset of the seriously ill population requires interdisciplinary, specialized home-based care to manage the complexity of their situation. Many plans and ACOs begin their interest in improving the care of people with serious illness because of the perceived strong business case for home-based care. Although some of the Learning Community organizations had already been pursuing this strategy, there is significant variation in how organizations implement home-based care, which drives mixed results. Some deliver home primary care over years, while others deliver home-based care with intended short enrollment periods. The distinction between home palliative and home primary care may be less meaningful for the frail population, so ensuring key clinical competencies in serious illness care is first priority.

The first two strategies — proactive identification and the successful engagement and support from a specially trained care manager — are key to implementing effective home-based care for those with serious illness. Other essential elements that drive outcomes include:

  1. Expert pain and symptom management;
  2. Reliable and meaningful 24/7 clinical response for urgent issues; and
  3. Skilled communication with patients and loved ones about what matters most as the disease progresses, identifying areas of concern, and aligning treatments with values.

For the MA Learning Community, conversations about filing for supplemental benefits for home-based palliative care landed on a note of caution. Organizations must first learn from an administrative program before committing to the filed benefit. Despite low numbers of home-based supplemental benefits, many MA plans are supporting access to home-based care for their eligible members, outside of the standard structure.

Across the Learning Communities, there was a growing recognition that one size does not fit all, and the optimal length of time that patients can benefit from home-based care is unknown. Some participants expressed concern about the cost-effectiveness of ongoing home-based services, and a few organizations modified their programs into short-term models, with some noting ongoing patient and home-based team touches to ensure that needs are met.

Lastly, there was strong interest in improving the coordination between home-based palliative care services with other services, especially complex care management, certified home health services, and primary care medical homes, and the approach of regularly scheduled meetings between care management and the home-based service providers was appreciated.

4. Incentivize Palliative Care in Acute and Post-Acute Networks

This last serious illness strategy has not yet been discussed among the Learning Communities, nor been pursued by any of the participating organizations to date. The strategy involves using incentives and/or quality reporting to encourage palliative care within hospitals and provider networks. Currently, CAPC is aware of only two health plans and one ACO that have included palliative care capabilities in their quality incentive or network assessment programs. Nonetheless, this small sample demonstrates that network incentives for palliative care can have a meaningful impact on timely access to palliative care services for people with serious illness. Early conversations in the Learning Communities confirmed that incorporating appropriate quality measures into value-based contracts (for MAs) or network dashboards (for ACOs) was an area for further consideration.

One word of caution did emerge: incentives and reporting should not be tied to the completion of an advanced directive or POLST/MOLST form. Focusing on form completion too often results in a check-box mentality and can lead to poorer communications between patients and providers. If documentation completion is a necessary process measure, consider using the selection of a health care proxy for all adult patients.

In Closing

The MA and ACO Learning Communities have refined a set of practical implementation strategies needed for organizations responsible for the quality and cost of Medicare populations. CAPC will be updating its tools and guidance based on the Learning Communities, but in the meantime, the Guide to Integrating Palliative Care into Population Management provides comprehensive guidance and useful tips.