Breaking the Link Between Enrollment in Hospice and Discontinuation of Dialysis

Since the introduction of the Medicare Hospice Benefit in 1983, the use of hospice at the end of life has increased substantially.¹ However, not all terminal diagnoses provide patients the same level of access to the high-quality, patient-centered, and symptom-focused care that hospice aims to achieve. This is especially true for patients with end-stage kidney disease who are receiving dialysis. Although half of all Medicare beneficiaries receive hospice services at the end of life, hospice is used by only a fifth of those receiving maintenance dialysis. Moreover, those patients receiving dialysis who do enroll in hospice have shorter hospice lengths of stay than the general Medicare population (8 days vs 20 days),¹ and almost half of them are not enrolled in hospice until the last 3 days of life.²

Most patients with end-stage kidney disease discontinue treatment with dialysis before initiating hospice. Because nearly all patients who are dependent on dialysis die within a week or 2 of discontinuing the treatment, the decision to initiate hospice for such patients is uniquely consequential compared with the decision to initiate hospice for patients with most other terminal illnesses. Thus, the relatively low use of hospice for patients receiving maintenance dialysis is not surprising. However, the broader question is why do most patients discontinue dialysis before entering hospice?

Although factors such as patient preference and practice patterns contribute, Medicare policy has indirectly created and maintained the association between hospice enrollment and discontinuing dialysis. Medicare does not preclude hospice patients from continuing to receive disease-directed interventions, including dialysis. However, 3 policies impede access.

The first roadblock centers on payment policy: within the Medicare Hospice Benefit, when an individual selects hospice care, Medicare will no longer separately pay for any disease-directed therapies associated with their primary hospice diagnosis (the term used for the illness judged to be their terminal diagnosis). Rather, the hospice agency becomes financially responsible for all expenses associated with the patient’s primary hospice diagnosis. Thus, when a patient enrols in hospice for end-stage kidney disease, the hospice agency must cover the cost of any dialysis provided with their per diem rate, which is typically about $200 to $300. For most hospices, this approach is not financially viable. However, the situation can be complex: if a patient with end-stage kidney disease enrolls in hospice for a different primary diagnosis, such as cancer, Medicare coverage for dialysis and hospice is not precluded.

A Medicare-financed concurrent care payment model, in which Medicare would pay for hospice and dialysis simultaneously, has been proposed to address this payment roadblock. This type of financing innovation has been operationalized by the US Veterans Health Administration (VA), which, in 2009, rolled out a national policy that enables veterans to receive VA-financed disease-directed therapies while they are enrolled in hospice.³ Practically, this means that hospice agencies are no longer financially responsible for expensive disease-directed therapies, even when those expenses are directly associated with the patient’s terminal diagnosis.

The outcomes of the VA’s concurrent care model are promising for patients with cancer, who can receive concurrent hospice and chemotherapy or radiation therapy. An observational study of veterans with stage IV non–small cell lung cancer found that VAs that had high use of concurrent care had reduced rates of aggressive care at the end of life and demonstrated cost savings.⁴ However, to our knowledge, the concurrent care model has been less examined for other groups of patients. Thus, the degree to which findings from oncology are applicable to concurrent hospice and dialysis among veterans with end-stage kidney disease is uncertain. A 2022 study found that about 40% of veterans with end-stage kidney disease who were receiving maintenance dialysis and enrolled in hospice through the VA received concurrent hospice and dialysis.⁵ This relatively high rate of concurrent care challenges the premise that low rates of concurrent care primarily reflect patient preferences. The findings of the study suggest that when hospice payment policy and regulations are more flexible than in the Medicare program, many patients elect concurrent care.

The Center for Medicare and Medicaid Innovation is currently testing the Kidney Care Choices Model. Within this model, nephrologists and dialysis facilities partner to create kidney contracting entities to serve as accountable care organizations. The model includes a provision that enables the contracting entities to waive the Medicare requirement that beneficiaries who elect hospice care cannot receive dialysis that is financed by Medicare. The cost of the dialysis is included in the performance measure for total cost of care that affects Medicare’s payments to kidney contracting entities. The Kidney Care Choices Model could allow for an examination of the quality and cost of concurrent care in entities that choose to offer it. The US Centers for Medicare & Medicaid Services should use findings from kidney contracting entities and studies that examine similar outcomes among veterans with end-stage kidney disease who receive concurrent care within the VA to inform decision-making.

The second roadblock is regulatory: a key prerequisite for hospice eligibility is a predicted life expectancy of 6 months or less. Medicare has established disease-specific clinical guidance for assessing prognostic eligibility for hospice. To avoid the real or perceived risk of Medicare audits or other financial consequences, hospice agencies are wary of prognosticating beyond these guidelines, of which there are none specific to patients with end-stage kidney disease who are receiving dialysis. To address this issue, Medicare should provide clinical guidance for estimating 6-month prognosis for patients with end-stage kidney disease who are receiving dialysis. Such guidance can be used to assess patient eligibility for hospice. The guidance should clearly state that dialysis use does not preclude patients from being eligible for hospice, provided that they have a predicted life expectancy of 6 months or less even with ongoing receipt of dialysis. Clinical guidance could be based on existing prognostic tools that estimate 6-month prognosis for dialysis patients, such as a tool that incorporates clinical data (age, albumin, comorbid peripheral vascular disease, and comorbid dementia) and whether the patient’s physician reports that they would be surprised if the patient died within 6 months.⁶

The third roadblock involves the quality metrics that Medicare’s End-Stage Kidney Disease Quality Incentive Program uses to evaluate a dialysis facility’s performance. Scores on these quality metrics are directly associated with a portion of the facility’s Medicare reimbursement. Meeting performance benchmarks on these quality measures, which include dialysis adequacy, is achieved through a rigid dialysis prescription, which is usually hemodialysis 3 times a week for 4 hours per session. However, these metrics are not the key measures of care quality for a patient receiving dialysis who is expected to live less than 6 months and whose goals of care have shifted from a primary focus on prolonging life toward a primary focus on quality of life. For such patients, performance metrics should be modified to focus more on patient-centered measures, such as the Kidney Disease Quality of Life scale. This approach would potentially facilitate palliative dialysis, in which dialysis frequency is guided by patients’ symptoms (eg, volume overload and uremia) and quality of life rather than being dictated by laboratory values.⁷

Even in the absence of hospice payment reform, a shift in Medicare end-stage kidney disease quality metrics could improve care and potentially increase access to concurrent hospice. The reason is that with the current Medicare hospice per diem rate, it might be financially sustainable for hospices to provide palliative dialysis. A model to provide palliative dialysis care could potentially be adapted from that used by many hospices to provide blood transfusions: symptom-based, less frequent, and with no laboratory draws between transfusions.

Individuals at the end of life often seek a balance between interventions that may extend their life and those that promote comfort. Medicare can help patients with end-stage kidney disease achieve their preferred balance by breaking the link between eligibility for hospice and discontinuing dialysis. Through changes to payment models, regulatory barriers, and quality metrics, Medicare can potentially improve timely access to hospice care for patients receiving dialysis near the end of life.