This year, 90,000 Americans with end-stage renal disease (ESRD) will die and questions will legitimately be raised as to whether terminal treatment and location of death adequately represented their preferences. These concerns are linked by a failure on the part of patients and staff to discuss prognosis and share in end-of-life (EOL) planning. The rate of hospice use among patients dying with ESRD is half that of the national average and one-quarter the rate for patients with terminal cancer. In other patient populations when meaningful EOL conversation occurs this is associated with increased hospice referral and improved quality of the dying.

Patients receiving hemodialysis (HD) often desire but rarely communicate with staff about prognoses, know little about availability of community hospice resources, or how to complete advance directives. Nephrologists are not trained to have these conversations, and although accustomed to relying on interdisciplinary teams, they are unaccustomed to collaborating with community hospices. Our preliminary research began by using focus groups, created and validated the first clinically useful HD prognostic tool, and developed a prototype for Shared Decision Making and Renal Supportive Care (SDM-RSC). This is a novel multimodal intervention that familiarizes patients, families, and dialysis staff with community hospice resources, emphasizes dialysis social work support, conveys information about terminal care issues, and encourages advance care planning.

The proposed study tests the central hypothesis that EOL care can be improved by relying on patients and stakeholders to enhance SDM-RSC for HD patients who are most likely to die. It will test whether an intervention that targets communication deficiencies can alter EOL outcomes and achieve the goal of matching patient preferences with terminal treatments.

This year, 90,000 Americans with end-stage renal disease (ESRD) will die and questions will legitimately be raised as to whether terminal treatment and location of death adequately represented their preferences. These concerns are linked by a failure on the part of patients and staff to discuss prognosis and share in end-of-life (EOL) planning. The rate of hospice use among patients dying with ESRD is half that of the national average and one-quarter the rate for patients with terminal cancer. In other patient populations when meaningful EOL conversation occurs this is associated with increased hospice referral and improved quality of the dying.

Patients receiving hemodialysis (HD) often desire but rarely communicate with staff about prognoses, know little about availability of community hospice resources, or how to complete advance directives. Nephrologists are not trained to have these conversations, and although accustomed to relying on interdisciplinary teams, they are unaccustomed to collaborating with community hospices. Our preliminary research began by using focus groups, created and validated the first clinically useful HD prognostic tool, and developed a prototype for Shared Decision Making and Renal Supportive Care (SDM-RSC). This is a novel multimodal intervention that familiarizes patients, families, and dialysis staff with community hospice resources, emphasizes dialysis social work support, conveys information about terminal care issues, and encourages advance care planning.

The proposed study tests the central hypothesis that EOL care can be improved by relying on patients and stakeholders to enhance SDM-RSC for HD patients who are most likely to die. It will test whether an intervention that targets communication deficiencies can alter EOL outcomes and achieve the goal of matching patient preferences with terminal treatments.

Specific Aims:

1. To determine whether SDM-RSC impacts the use of hospice services, location of death, and EOL planning. The RSC intervention will be employed at the HD clinics with patients who are in the highest quintile of mortality.

   Hypothesis: SDM-RSC will increase a) hospice use, b) deaths at home (vs. in institutions), and c) completion of advance directives.

2. To determine the effect of SDM- RSC on quality of life/death and caregiver satisfaction with patient care in the last week of life. Validated instruments reflecting values and priorities of patient and caregiver will be collected by interviewers.

Hypothesis: SDM-RSC will significantly improve quality of life/death while also increasing caregiver satisfaction.

The proposed work addresses recommendations by the Renal Physician's Association for EOL care and the Institute of Medicine for patient-centered healthcare. The novel communication intervention is relevant to not only ESRD but also other high-mortality disorders.

Scope of the Problem:

A match between patient goals and the care received should be the gold standard for quality medical management. Meaningful end-of-life (EOL) conversations are associated with increased hospice referral, less aggressive and expensive medical treatment, and improved satisfaction on the part of families. However, a survey of ESRD patients reported less than 10% having a conversation about any EOL issues with their nephrologist in the previous year, and more than 90% endorsed that none of their physicians had ever discussed how much time they had to live. It is not known whether an intervention that targets deficiencies in communication and terminal care planning can improve EOL outcomes. This is an area that has been neglected in ESRD research, yet should have tremendous importance to patients, families, and policy-makers.

Over 600,000 Americans have ESRD at a cost to society of more than $42 billion a year, and patients over the age of 75 constitute the fastest growing segment of the ESRD population in the United States. Up to one-third of these elderly patients have four or more comorbid chronic health conditions, and hemodialysis (HD) does not always substantially prolong their lives. This year, Wong and associates reported that 49% of elderly long-term HD patients spent time in an ICU in their final month of life, compared with 24% of cancer patients, suggesting intensity of care at the EOL is substantially higher in ESRD than that received by other Medicare beneficiaries with life-limiting illnesses. According to analyses from the United States Renal Data System, these decisions may have important economic ramifications. This figure demonstrates that the costs of care in the last week of life for those using hospice and withdrawing from dialysis was approximately half the cost of those who did neither. In light of the substantial prevalence of symptoms, costs, and high mortality rates, it is time to reconsider the care, and especially the EOL care, offered to the ESRD population with advancing age.

Communication of prognosis is a key step in EOL planning, but occurs infrequently or extremely late in the dying process among patients with ESRD. Most ESRD patients want to learn about EOL issues, and according to Davison, the most valued components are being informed about prognosis, treatment options (including withdrawal from dialysis) and planning for death. In two studies, 95% and 97% of patients with ESRD preferred to be given life-expectancy information-even if their prognosis was poor-and patients specifically wanted their physician to disclose this information without prompting. Nevertheless, nephrologists and dialysis staff infrequently discuss EOL issues with patients and families, although they repeatedly face difficult determinations about treatment options, including whether to discontinue renal replacement therapies or to consider referral for hospice services. Despite national guidelines and an explicit recommendation by the Renal Physicians Association and American Society of Nephrology, patients and their family members or surrogate decision makers are often not provided with available information. This may help explain why only one-in-five dying dialysis patients currently receive hospice care-about half that of national figures for overall deaths in the United States and one-fourth that of people dying from cancer.

Both uncertainty regarding individual prognosis and a lack of training in communicating has limited EOL discussions between nephrologists and ESRD patients. Instruments to identify individuals undergoing maintenance HD who are at highest risk for death have only recently been shown to be reliable predictive models for clinical use. Integrated prognostic models now take into account laboratory values, comorbidities, changes in comorbidity score over time, functional status/fragility, quality of life (QOL), and sometimes either the patient's or clinician's prediction of survival. Drs. Cohen and Germain have developed a practical and reliable instrument that is available online (http://touchcalc.com/calculators/sq), as a free application for mobile devices and it is being increasingly cited for its potential value in the literature. In addition to the barriers of providing a reliable prognosis, communication has been highlighted as a barrier in a qualitative study of elderly ESRD patients and nephrologists.

Communication of prognosis may be an essential ingredient in the acceptance of life-limiting conditions, completion of advance directives, do-not-resuscitate orders, and the delineation of goals of care. However, nephrologists and dialysis staff have lacked training to meaningfully participate in these complex but much needed conversations. In recent years, guidelines have become available to assist medical personnel in communicating 'bad news' and introducing hospice and other EOL resources to patients and families in an honest and compassionate manner. Patients who talk with their physicians about approaching death are more satisfied with care than those who do not, and they also frequently request deaths take place at home. These results are consistent with other studies, suggesting that EOL communication between patients and physicians is associated with better outcomes, less expensive medical care, and decreased anxiety and depression on the part of the bereaved survivors.

We have identified a major barrier in the management of dialysis patients: the dearth of communication between staff and patients about prognosis and EOL considerations. Without these conversations, this population will continue to receive overly demanding and expensive treatments, infrequent referrals for hospice services, and a diminished quality of life and death. The extensive preliminary work preceding this study, which began with focus groups involving patients and families, has allowed us to develop and validate an accurate prognostic instrument for the identification of HD patients with poor short-term survival expectancy. The multimodal SDM-RSC intervention takes advantage of the existing interdisciplinary team approach that is the bedrock of HD practice, and it empowers the social worker to engage in EOL communication with the patients and their loved ones. The SDM-RSC intervention is designed to be practical, adaptable to existing staff, and easily replicated in dialysis facilities throughout the country. Additional involvement of patient-partners from two distinct sections of the country will greatly improve the protocol.

Renal supportive care holds the promise of becoming an increasingly important and desirable aspect of patient-centered care in ESRD. The principal elements of SDM-RSC include: identification of high-mortality patients, initiating discussions with them and families, sharing decision-making and advance care planning to establish agreed level of care, and when appropriate changing the emphasis from curative to palliative treatments, such as withdrawal of dialysis and referral for hospice services. This study intends to challenge the current paradigm of EOL care of ESRD patients and establish whether delivery of prognostic information, encouragement of advance directives, and sensitive discussion of terminal treatment options, e.g. hospice services, by nephrologists and dialysis social workers, can together meaningfully alter the last remaining portion of patient lives and the care they receive. The trial will confront the issue of overuse-when an appropriate service, e.g. HD, continues to be provided under circumstances in which the potential for harm may exceed the possible benefit. Furthermore, the proposed trial may significantly improve the QOL of patients and provide the opportunity for higher quality care.

Brief Overview of SDM-RSC Study Design:

The study utilizes a prospective intervention period and a retrospective control period at the same dialysis units, in order to balance efficacy, power, and accuracy. The intervention will be implemented at the unit level and is intended to impact not only the specific patients enrolled in the study but also other patients receiving care in these units. The design avoids the contamination effect and inadvertent cross-overs that would be present with an individual participant randomization design. Because we expect perceptions and practices to be changed among care providers other designs, e.g. cross-over designs, would also be unusable. Cluster randomization may also not be practical given the variation between units in practice patterns and logistics involved in conducting a study with a sufficient number of dialysis to maintain internal validity. The design may be open to potential selection bias and secular trends. Adjustments for these are discussed in the data analysis section.

This study will test whether the SDM-RSC intervention compared to usual care improves EOL care, including: increased hospice use, enhanced EOL planning, and improved patient and caregiver satisfaction. The multi-modal SDM-RSC intervention communicates prognosis in a face-to-face encounter, uses an algorithm-based social work intervention, and provides hospice outreach. This study will be recruiting subjects from 16 dialysis units associated with 2 clinical centers-Baystate-Tufts (8), University of New Mexico (8). The study sample will be selected using a validated prognostic tool to enrich the pool of patients who have a high risk of mortality and are most likely to benefit from the SDM-RSC intervention. Subjects will be recruited over 6 months and enrolled subjects will be treated and followed for at least 18 months.

Inclusion Criteria:

• High-risk patients, as defined by falling within the high-risk category of our validated prognostic instrument and have approximately 50% 18-month mortality.
• English- and Spanish-speaking patients will be included (estimated to be 95% of this population).
• Patients who receive hemodialysis at one of our 16 research dialysis sites during the data collection period.
• Patients must be willing and able to sign the consent form.
• Patients who are lack the capacity to meaningfully participate in medical decisions must have a surrogate who is willing to sign the informed consent.

Exclusion Criteria:

• Children 18 years of age. Children constitute 2% of the dialysis population, and our preliminary survey of the study sites found no children were active patients. In any case, the renal and other physical factors of children with ESRD are not directly comparable to those of adults.
• Does not belong to the population's high-mortality risk quintile according to our prognostic instrument
• Severe psychiatric disorders including schizophrenia, bipolar disorder which would interfere with participation in the study (severity determined by psychiatric hospitalization in the past month or actively suicidal)
• Active substance abuse (active abuse is defined as using alcohol or recreational drugs in the past 30 days in a way that interferes with their ability to function in daily life)
• Expectation of native kidney recovery
• History of poor adherence to thrice-weekly hemodialysis (poor adherence defined by missing 4 treatments in the past month)
• Unable to communicate in English or Spanish
• Scheduled for living donor kidney transplant, conversion to peritoneal dialysis, or plans to relocate to another hemodialysis unit
• Current pregnancy or actively planning to become pregnant
• Currently a prisoner
• Unable or unwilling to follow the study protocol
• Unable or unwilling to provide informed consent or sign Institutional Review Board (IRB)-approval or lack of a surrogate/proxy
• Exclusion by primary nephrologist or social worker due to risk of harm