Timing is Everything: Defining the Serious Illness Population

Betsy recently died after a sixteen-year battle with amyotrophic lateral sclerosis (ALS). ALS is a terminal illness that slowly paralyzes adults – meaning she progressively lost the ability to move her body while her creative and fun-loving mind remained intact. Her paralysis required around-the-clock care, and her frailty resulted in broken bones, chronic infections, and pain that eroded her quality of life.

Meanwhile, she had four thriving children and three grandchildren and wanted to celebrate as many weddings and birthdays as she could. In the last months of her life, she received hospice care to enable her to die peacefully at home. For years before that, her adult children were left wondering how to best care for her – which doctor should they visit for her skin infections? What about her digestive pain? When does a fall necessitate an emergency department trip?

Individuals like Betsy must navigate a system that sees dying as a “yes”/“no” rather than a “maybe” or “eventually.” To receive team-based palliative care focused on symptom relief, they often must elect hospice—but in order to do so, they must be within 6 months of death and they must stop curative treatments. Patients who are facing a terminal illness but not immediately dying know that coordinated care for serious illness is hard to find.

What’s the Potential Solution?

In the last several years, providers have developed new models of serious illness care within testbeds both public and private. These innovations largely integrate community-based palliative care, which focuses on quality of life and symptom relief, into serious illness care at many stages of illness.

The expansion of palliative models to Medicare, the largest payer for American healthcare, is in process. The Centers for Medicare and Medicaid Innovation (CMMI) has implemented, and is considering, several models for patients experiencing serious illness. Other groups have proposed new models to CMMI to target patients with serious illness who may still be pursuing curative care. These models vary depending on the types of patients they will serve or when in the course of a serious illness they target patients (more on this below).

The Challenge: Identifying the Right Patients

As with innovation in any field, scaling a solution to a larger population is where the real challenges emerge. Payment models require a clear definition of who is eligible for a benefit, and when. For serious illness care, timing may be everything.

Identifying both which patients within a larger population should receive serious illness care, and when, is critical for payers like Medicare. Existing models being tested at CMMI that serve some of the patients that might be targeted by an serious illness care model (such as Independence at Home and the Medicare Care Choices Model) include eligibility criteria such as specific diagnoses and disability levels (the “who”) and  require specific types of healthcare use within a specific period (the “when”).

New models recently proposed to CMMI attempt to identify the eligible population. However, the proposed eligibility criteria require more patient interaction to determine patient eligibility:

• The Advanced Care Model (ACM) would target chronically ill patients who meet a mix of criteria including acute healthcare use, an irreversible new disability, declining body weight, or specific scores on a choice of palliative scales.
• Patient and Caregiver Support for Serious Illness (PACSSI) targets patients with designated serious illnesses (or three or more less serious chronic conditions), functional decline (defined using palliative scales, disabilities, or specific medical equipment), and acute healthcare use.

As diseases progress, a patient’s status can change quickly. As highlighted by these proposals, any eligibility requirement that targets patients at some point in their disease trajectory might rely on frequent assessments and clinical judgment. These can be both challenging to standardize and require more frequent interaction with patients (and thus cost).  

The research community echoes how difficult it can be to identify patients who could benefit from serious illness care even with rich data sources. For example, Kelley and Bollens-Lund attempted to define “high need” patients using traditional criteria like diagnoses and healthcare use. They enhanced this with survey data, including measures of disability, nutritional decline, cognitive impairment, caregiver strain, and symptoms such as pain.

But even this method showed only marginal success. The researchers found that these factors had limited ability to predict total Medicare costs, hospitalizations, and mortality within a 1-year timeframe—they were able to predict who wouldn’t have these outcomes, but not who would have them. Thus, even with rich survey data, it’s difficult to prospectively identify patients who have burdensome serious illness.

Moving Forward: What data sources can we use?           

To make integrated palliative care more widely available, we need data sources and criteria that can identify eligible patients. Yet data collection can be expensive and burdensome to providers. This leaves us with the conundrum of identifying rich data sources that are still practical.

Given the limited success even with rich survey data, new assessments might be helpful. CMS is currently supporting efforts to develop standardized physical and cognitive function assessment in rehabilitation care. Once standardized, these assessments could be extended to other care settings. But which health care professional should be conducting these assessments (nurses, physical therapists, etc.)?  And will these assessments be conducted in all care settings to ensure that as many eligible patients are captured as possible?

New technology is also worth considering. “Smart” devices and technologies could help collect real-time data from patients or caregivers. Currently, insurance companies offer reduced insurance premiums for patients who submit fitness tracker data, and researchers are using data from wearable sensors to anticipate falls. Innovators could develop applications  for smart phones and other smart home technology (e.g., Amazon’s Alexa) to assess other aspects of “high need” patients such as caregiver assessments and nutritional, cognitive, and symptom assessments. As with all potential solutions, technology presents its own set of challenges including consumer buy-in, privacy concerns, and data analysis and reporting difficulties.

In the meantime, there are formidable challenges to identifying the eligible population. Successfully identifying these patients could allow Medicare and other payers to scale any proven models. Most importantly, this could change the experience of serious illness for the next generation.  The burden of serious illness is great, and people like Betsy deserve a system better designed to relieve their suffering and support their families as they navigate a serious illness.