Family Members’ Understanding of the End-of-Life Wishes of People Undergoing Maintenance Dialysis

Abstract

Visual Abstract

Background and objectives 

People receiving maintenance dialysis must often rely on family members and other close persons to make critical treatment decisions toward the end of life. Contemporary data on family members’ understanding of the end-of-life wishes of members of this population are lacking.

Design, setting, participants, & measurements 

Among 172 family members of people undergoing maintenance dialysis, we ascertained their level of involvement in the patient’s care and prior discussions about care preferences. We also compared patient and family member responses to questions about end-of-life care using percentage agreement and the κ-statistic.

Results 

The mean (SD) age of the 172 enrolled family members was 55 (±17) years, 136 (79%) were women, and 43 (25%) were Black individuals. Sixty-seven (39%) family members were spouses or partners of enrolled patients. A total of 137 (80%) family members had spoken with the patient about whom they would want to make medical decisions, 108 (63%) had spoken with the patient about their treatment preferences, 47 (27%) had spoken with the patient about stopping dialysis, and 56 (33%) had spoken with the patient about hospice. Agreement between patient and family member responses was highest for the question about whether the patient would want cardiopulmonary resuscitation (percentage agreement 83%, κ=0.31), and was substantially lower for questions about a range of other aspects of end-of-life care, including preference for mechanical ventilation (62%, 0.21), values around life prolongation versus comfort (45%, 0.13), preferred place of death (58%, 0.07), preferred decisional role (54%, 0.15), and prognostic expectations (38%, 0.15).

Conclusions 

Most surveyed family members reported they had spoken with the patient about their end-of-life preferences but not about stopping dialysis or hospice. Although family members had a fair understanding of patients’ cardiopulmonary resuscitation preferences, most lacked a detailed understanding of their perspectives on other aspects of end-of-life care.

Introduction

Toward the end of life, many people undergoing maintenance dialysis receive intensive, potentially burdensome disease-focused treatments that neither extend their life nor improve their quality of life and may not be aligned with their values, goals, and preferences (^1–3). Nearly 63% of US Medicare beneficiaries receiving maintenance dialysis are admitted to an intensive care unit during the last 3 months of life, while only 26% enroll in hospice (³). Available data among older Medicare beneficiaries suggest that many lack the cognitive capacity to make critical treatment decisions toward the end of life and need the help of a surrogate decision maker (⁴). In order to ensure that people undergoing maintenance dialysis receive end-of-life care that is aligned with their values and goals, both the National Academy of Medicine (⁵) and the Renal Physicians Association (⁶) have recommended that nephrologists engage patients with kidney disease and their surrogate decision makers in a process of advance care planning.

These recommendations are supported by published data suggesting that people undergoing dialysis want their family members to be involved in end-of-life planning (⁷) and that surrogate involvement in the process of advance care planning can help to promote goal-concordant care (⁸,⁹). Engaging surrogate decision makers in advance care planning can also help to reduce their emotional distress and increase their satisfaction with patients’ care (⁴,¹⁰,¹¹). However, there is no current mandate or incentive to involve surrogate decision makers in advance care planning for patients who have decisional capacity, which may limit their ability to learn about patients’ values, priorities, and treatment preferences (^12–16).

In 2001, Hines et al. (¹⁷) demonstrated a lack of congruence in the responses of patients undergoing dialysis and their surrogate decision makers across a range of scenarios intended to elicit the value patients placed on life prolongation. More recently, several studies have examined surrogate decision makers’ understanding of the treatment preferences of people undergoing maintenance dialysis, but most of these have focused more narrowly either on questions about values around life prolongation versus comfort (⁸) or on scenarios related to cardiopulmonary resuscitation (CPR) and dialysis discontinuation (¹⁸). To provide a more complete and contemporary understanding of family members’ perception of patients’ values and preferences pertaining to end-of-life care, we conducted a survey among people undergoing maintenance dialysis and their family members and close friends. Our goals were to understand (1) family members’ role in patients’ care; (2) their prior discussions with the patient about planning for serious illness; and (3) their understanding of the patient’s treatment preferences, values, preferred decisional role, and prognostic expectations.

Materials and Methods

As previously described (¹⁹,²⁰), we conducted a survey about end-of-life care (Supplemental Appendix) among 997 people receiving in-center maintenance dialysis at 31 nonprofit dialysis facilities in the greater Seattle and Nashville metropolitan areas from 2015 to 2018. Patients who provided written informed consent and completed the survey were asked to provide contact information for one or more family members or friends who could be invited to participate in the study and were instructed to list the person who was most involved in their care first. Throughout the article, we use the term “family” inclusively to refer to family members or friends. Patients who did not provide contact information for one or more family members were asked whether this was because they did not have anyone to ask or because they did not want the study team to contact their family members. Study staff attempted to contact all listed family members to discuss the study and invite them to participate. After providing verbal consent, family members participated in a phone survey to elicit their involvement in the patient’s care and elicit their involvement in the patient's care and understanding of their treatment preferences, values, preferred decisional role, and prognostic expectations. Although more than one family member could enroll in the study, these analyses describe only the responses of the enrolled family member who was most involved in the patient’s care. The study was approved by the institutional review board at the University of Washington.

Statistical Analyses

We used t tests and chi-squared tests as appropriate to compare the characteristics of patients with and without one or more family members enrolled in the study. We calculated the percentage agreement between patient and family member responses to survey questions about treatment preferences, values, preferred decisional role, and prognostic expectations. For each comparison, we also report the κ-statistic (values <0 indicate no agreement, values 0–0.20 indicate slight agreement, values 0.21–0.4 indicate fair agreement, values 0.41–0.60 indicate moderate agreement, values 0.61–0.80 indicate substantial agreement, and values 0.81–1.0 indicate near perfect agreement) (²¹). Although the patient and family member surveys included parallel questions, a few items did not include the same response options. In particular, for questions about CPR and mechanical ventilation, only family members were given the option to respond that they were not sure. In these instances, we calculated percentage agreement and κ-scores only among dyads where family members had provided a definitive response (i.e., had not responded that they were not sure). For questions about values around life prolongation and prognostic expectations, both patients and family members could respond that they were not sure. For these questions, we measured percentage agreement both across all responses—including those where one or both members of the dyad indicated they were not sure—and among dyads who provided definitive responses.

For items with ordered responses (i.e., preferences for CPR and mechanical ventilation, preferred decisional role, and definitive responses to the question about prognostic expectations), we also report the percentage of definitive responses that were maximally discordant between the patient and the family member.

We conducted sensitivity analyses among dyads that included (1) family members who indicated they were the patient’s surrogate decision maker, (2) family members who indicated that they were extremely confident in their knowledge about the care the patient would want to receive if seriously ill, (3) patients who were 75 years or older, and (4) patients who indicated that they were in fair or poor health (rather than excellent, very good, or good).

Results

Among 997 enrolled patients, 511 indicated that they did not want the study team to contact a family member or friend, 104 indicated that they did not have anyone to list, and 382 provided contact information for one or more family members. Listed family members of 210 patients either could not be contacted or chose not to participate, and 187 family members of 172 (17%) patients verbally consented to participate in the study and completed the survey (Supplemental Table 1).

Characteristics of Patients and Enrolled Family Members

The mean (SD) age of the 172 patients with an enrolled family member was 61 (±15) years, 82 (48%) were women, 95 (55%) were White individuals, 50 (29%) were Black individuals, 27 (16%) were of other race, and 12 (7%) were Hispanic individuals. The mean (SD) age of the 172 enrolled family members was 55±17 years, 136 (79%) were women, 93 (54%) were White individuals, 43 (25%) were Black individuals, 36 (21%) were of other race, and seven (4%) were Hispanic individuals (Table 1). A total of 67 (39%) family members were spouses or partners, 46 (27%) were children, 26 (15%) were siblings, 16 (9%) were parents, nine (5%) were other relatives, and eight (5%) were friends (Table 2). A total of 102 (59%) family members lived with the patient, and 154 (90%) indicated that they were the family member most involved with the patient's care. Most family members provided some care for the patient, including 56 (33%) who provided care “most of the time” and 57 (33%) who provided care “all of the time.” When asked how confident family members were about their knowledge of the care the patient would want to receive if seriously ill, 55 (32%) were “extremely confident,” 74 (43%) had “quite a bit” of confidence, 29 (17%) were “somewhat” confident, eight (5%) were “a little” confident, and six (3%) were “not at all” confident.

Table 1. - Characteristics of enrolled family members

Participant Characteristic	Enrolled Family Members Included in the Analytic Sample, n=172	Patients with Enrolled Family Members, n=172
Age (SD), yr	55 (17)	61 (15)
Women, n (%)	136 (79)	82 (48)
Race, n (%)
White	93 (54)	95 (55)
Black	43 (25)	50 (29)
Other a	36 (21)	27 (16)
Ethnicity, n (%)
Hispanic	7 (4)	12 (7)
Missing	0 (0)	3 (2)
Highest educational level, n (%)
Less than high school	9 (5)	18 (10)
Graduated from high school	65 (38)	42 (24)
Some college, community college, or trade school	30 (17)	40 (23)
Graduated from college, community college, or trade school	47 (27)	60 (35)
Graduate degree or some postgraduate education	21 (12)	12 (7)
Self-reported health status, n (%)
Excellent or very good	90 (52)	30 (17)
Good	52 (30)	60 (35)
Fair or poor	29 (17)	82 (48)
Missing	1 (1)	0 (0)
Years on dialysis, n (%)
<1		66 (38)
1–5		33 (19)
>5		72 (42)
Missing		1 (1)
No. of family members involved in caring for patient, n (%)
0		1 (1)
1		50 (29)
2		55 (32)
≥3		66 (38)

Percentages may not add up to 100% due to rounding.

^aAsian, Native Hawaiian or other Pacific Islander, American Indian or Alaskan Native, or belonged to more than one race.

Table 2. - Family members’ role in patient’s care (n [percentage])

Family Members' Characteristics and Responses to Survey Questions	n (%), n=172
Relationship to patient
Spouse/partner	67 (39)
Child	46 (27)
Sibling	26 (15)
Parent	16 (9)
Other relative	9 (5)
Friend	8 (5)
Living arrangement
With the patient	102 (59)
Within 40 miles	57 (33)
>40 miles	13 (8)
Frequency with which oversight or care is provided for the patient
None of the time	8 (5)
Some of the time	51 (30)
Most of the time	56 (33)
All of the time	57 (33)
Family member most involved in patient’s care	154 (90)
How confident are you that you know what kind of care the patient would want if he/she was very sick and unable to speak for himself/herself?
Extremely	55 (32)
Quite a bit	74 (43)
Somewhat	29 (17)
A little	8 (5)
Not at all	6 (3)
Has the patient spoken with you about whom he/she would want to make medical decisions if unable to speak for himself/herself?
Yes	137 (80)
No	35 (20)
Has the patient signed papers naming you as the person who would make medical decisions if unable to speak for himself/herself?
Yes	70 (41)
No	63 (37)
Not sure	39 (23)
Has the patient spoken with you about the kinds of situations that might reduce his/her quality of life so much that life would not be worth living?
Yes	91 (53)
No	81 (47)
Has the patient documented his/her wishes for care?
Yes	65 (38)
No	61 (35)
Not sure	46 (27)
Has the patient spoken with you about the kinds of treatments he/she would want and not want if he/she was to become very sick and unable to speak for himself/herself?
Yes	108 (63)
No	64 (37)
Has the patient ever had a conversation with you about stopping dialysis if he/she was to become sicker or if his/her goals changed?
Yes	47 (27)
No	125 (73)
Has the patient ever had a conversation with you about whether he/she would want to receive hospice care if he/she was to become sicker or if his/her goals changed?
Yes	56 (33)
No	116 (67)

Percentages may not add up to 100% due to rounding.

Engagement in Advance Care Planning

Most (80%, n=137) family members reported that the patient had spoken with them about whom they would want to make medical decisions on their behalf. Seventy family members (41%) indicated that they were the patient’s documented surrogate decision maker, while 39 (23%) were unsure about this. Ninety-one family members (53%) said that the patient had spoken with them about situations that might reduce their quality of life so much that life would not be worth living, and 108 (63%) reported that the patient had spoken with them about the types of treatments they would and would not want. Only 65 (38%) indicated that the patient had documented their wishes for care, and 46 (27%) were unsure about this. Overall, 47 (27%) family members indicated that the patient had spoken with them about stopping dialysis were they to become sicker or if their goals changed, and 56 (33%) said that the patient had spoken with them about whether they would want to receive hospice care under such circumstances.

Knowledge of Patient’s End-of-Life Wishes

Treatment Preferences.

Overall, ten (6%) family members were unsure about the patient’s CPR preferences (Table 3). After excluding these responses and those of one dyad for which the patient’s response was missing, there was 62% agreement between patient and family member responses to the question about CPR. The percentage agreement and κ-score were higher when responses were dichotomized as wanting versus not wanting CPR (83% agreement, κ=0.31). Among dyads where family members provided a definitive response to the question about CPR, only 3% offered maximally discordant responses.

Table 3. - Agreement between responses of family members and patients to survey questions about the patient, n (%)

Question, Variable Configuration, and Responses	Responses of Enrolled Family Members Included in the Analytic Sample, n=172	Responses of Patients, n=172	κ	Agreement, %	P Value
If the patient had to decide right now, do you think he/she would want CPR if his/her heart was to stop beating?
All responses
Definitely yes	110 (64)	109 (63)	0.26	62	<0.001
Probably yes	33 (19)	34 (20)
Probably no	5 (3)	15 (9)
Definitely no	14 (8)	13 (8)
Not sure	10 (6)	NA
Missing	0 (0)	1 (1)
Dichotomous response
Definitely or probably yes	143 (83)	143 (83)	0.31	83	<0.001
Definitely or probably no	19 (11)	28 (16)
Not sure	10 (6)	NA
Missing	0 (0)	1 (1)
If the patient had to decide right now, do you think he/she would want to be placed on a breathing machine if he/she was to become so sick he/she could not breathe on his/her own?
All responses
Definitely yes	42 (24)	60 (35)	0.12	35	0.006
Probably yes	42 (24)	45 (26)
Probably no	26 (15)	32 (19)
Definitely no	36 (21)	35 (20)
Not sure	26 (15)	NA
Dichotomous response
Definitely or probably yes	84 (49)	105 (61)	0.21	62	0.006
Definitely or probably no	62 (36)	67 (39)
Not sure	26 (15)	NA
Missing	0	0
What do you think the patient would value if he/she was seriously ill or dying?
All responses
Extending life	47 (27)	34 (20)	0.13	45	0.006
Relieving pain and discomfort	87 (51)	79 (46)
Not sure	38 (22)	59 (34)
Definitive responses
Extending life	47 (35)	34 (30)	0.21	67	0.02
Relieving pain and discomfort	87 (65)	79 (70)
If the patient had to decide right now, where do you think he/she would prefer to die if circumstances allowed him/her to choose?
At home or the home of a relative	111 (65)	108 (63)	0.07	58	0.20
Other response	40 (23)	62 (36)
Missing	21 (12)	2 (1)
What role do you think the patient would want to have in decision making?
Patient led	95 (55)	93 (54)	0.15	54	0.01
Shared	63 (37)	64 (37)
Physician led	8 (5)	13 (8)
Missing	6 (3)	2 (1)
How long would you guess people of similar age with similar health conditions to the patient usually live?
All responses, yr
<5	32 (19)	16 (9)	0.15	38	<0.001
5–10	39 (23)	29 (17)
>10	57 (33)	71 (41)
Not sure	42 (24)	56 (33)
Missing	2 (1)	0 (0)
Definitive responses, yr
<5	32 (25)	16 (14)	0.17	51	0.01
5–10	39 (31)	29 (25)
>10	57 (45)	71 (61)

CPR, cardiopulmonary resuscitation; NA, not applicable.

When asked about whether the patient would want to receive mechanical ventilation, 26 (15%) family members were unsure. After excluding these responses, there was 35% agreement between patient and family member responses to the question about mechanical ventilation (κ=0.12), with patients being more likely than family members to indicate that they would definitely want to be placed on a breathing machine. Percentage agreement was higher when responses were dichotomized as wanting versus not wanting mechanical ventilation (62%, κ=0.21). Among dyads in which family members provided a definitive response to the question about mechanical ventilation, 9% of responses were maximally discordant between the patient and the family member.

Values.

There was 45% agreement between patient and family member responses to the question about values around life prolongation versus relief of pain and discomfort (κ=0.13), with family members being more likely than patients to prioritize life prolongation and less likely to report that they were unsure about the patient’s values. After excluding patients and family members who indicated they were not sure, there was 67% agreement between family member and patient responses to this question (κ=0.21).

Overall, 21 (12%) family members did not respond to the question about the patient’s preferred place of death. Among dyads who responded to this question, there was 58% agreement on whether the patient would prefer to die at home or the home of a relative if circumstances allowed them to choose (κ=0.07), with similar percentages of patients and family members selecting this option.

Preferred Decisional Role.

There was 54% agreement between patients and family members about the patient’s preferred decisional role (κ=0.15), with family members being slightly more likely than the patient to indicate that the patient would prefer a patient-led or shared versus physician-led approach.

Prognostic Expectations.

There was 38% agreement between patient and family member responses to the question about prognostic expectations (κ=0.15). After excluding dyads in which either the patient or the family member responded that they were not sure, there was 51% agreement (κ=0.17), with patients having more optimistic expectations about prognosis than their family members. Overall, 15% of definitive responses to this question were maximally discordant.

The results of subgroup analyses were broadly consistent with the primary analysis (Supplemental Tables 2–5), with the exception that percentage agreement (50%, κ=0.3) for the question about prognosis was higher among dyads where the family member was extremely confident in their knowledge of the care the patient would want to receive if seriously ill.

Discussion

Among 172 dyads of patients undergoing dialysis and their family members, there was, at best, modest agreement between responses to a range of questions about end-of-life care. These findings are consistent with earlier work in a variety of different populations examining agreement between patients and their family members or surrogate decision makers around resuscitation preferences and the value placed on life prolongation (^12–16).

Our findings also align with those of several prior studies among people undergoing dialysis that have compared patient and family member responses to questions about values (⁸) around life prolongation, CPR, and dialysis discontinuation (¹⁷,¹⁸). In a 2001 study of 242 patients receiving dialysis and their surrogate decision makers, percentage agreement ranged from 3% to 49% across different scenarios intended to elicit the value placed on life prolongation (¹⁷). In a Japanese study of 398 patients on maintenance dialysis and their surrogate decision makers, family members’ predictions of patients’ wishes regarding CPR and dialysis discontinuation were no better than chance (¹⁸). In a study of 58 Black patients receiving dialysis and their surrogate decision makers, surrogates’ confidence in their knowledge of patients’ care preferences bore no relation to the accuracy of their knowledge of the patients’ goals of care, and only 34% of family members had an accurate understanding of these goals (²²).

In the context of this earlier work, our findings indicate that there has been limited progress over the last several decades in family members’ understanding of the values and preferences of people undergoing dialysis and highlight some of the challenges involved in supporting the advance care planning process (¹⁷,^23–25). Qualitative studies among family members of people both with (²⁴,²⁶) and without kidney disease (²⁷,²⁸) speak to the complex dynamics of engaging family members in patients' care. Similar to other chronic conditions (²⁷,²⁹), it is not unusual for the family members of people with advanced kidney disease to become involved in their care relatively late in the course of illness in situations of crisis or increasing care needs, at which time their involvement is often expected and prompted by the health system (¹⁷,²⁴). In their study of patients undergoing dialysis, Aasen et al. (²⁶) described how family members often had to struggle both with the patient and the health system to be involved in their care earlier in the course of illness.

It is perhaps instructive that when we invited patients enrolled in our study to identify one or more family members knowledgeable about their care, a majority indicated that they did not want the study team to contact their family members. Approximately 10% indicated that they did not have a family member to contact, and fewer than half of the remaining patients had one or more family members enroll in the study. Perhaps not surprisingly, those family members who did enroll had a relatively high level of involvement in patients’ care. With this context in mind, the low frequency of prior discussion about hospice and dialysis discontinuation and the generally poor agreement between patient and family member responses to survey questions are especially striking.

Our findings underscore the challenges and potential pitfalls of surrogate decision making (¹⁵) for people undergoing dialysis, including that some may lack friends and family who could serve in this role (³⁰). They also highlight the potential value of supporting and helping family members to prepare for their future role of upholding patients’ goals at the end of life and suggest that there may be opportunities to better engage them in planning for serious illness (⁹,²⁴,²⁷,^31–33). These efforts should ideally be contextualized within a person-centered paradigm that is grounded in an appreciation of the importance of learning about and upholding personhood and of conceptualizing advance care planning as a process of behavioral change (⁹,³⁴). Ideally, advance care planning should also be informed by an understanding of the inherent tensions around surrogate decision making, complexities of family relationships, and the individual and dynamic nature of patients’ readiness (³⁴) and ability to involve family members in their care (¹⁵,^35–38). Available evidence, both among people undergoing dialysis and in other populations, suggests that proactive engagement of patients’ family members holds promise not only for aligning care with patients’ values and goals but also for improving the experiences of both patients and their surrogate decision makers (⁸,¹¹).

For people with advanced kidney disease, decision making about dialysis initiation and discontinuation is often complex and intertwined with decisions about hospice enrollment. Limited discussion about dialysis discontinuation and hospice between patients and family members enrolled in our study highlights a potentially missed opportunity to prepare for these eventualities (³⁹,⁴⁰). Lack of prognostic awareness among patients (¹⁹,⁴¹) and their family members and the relatively high frequency of maximally discrepant responses to the question about expected prognosis are also striking. Both of these findings highlight the need for more work to improve communication between patients, family members, and clinicians and to understand what drives patients’ and family members’ prognostic expectations (¹,²,¹⁹,²⁴,⁴²). Even when there is substantial uncertainty about future illness trajectories, sharing information about prognosis can help to ensure that patients and family members have realistic expectations about the future course of illness and allow them to plan accordingly (^43–46). These discussions can also help to build trusting relationships with members of the health care team (⁴⁷), limit unwanted care, and improve access to palliative care and hospice services toward the end of life (⁴⁴,⁴⁸). Given the very limited life expectancy of many patients undergoing dialysis and the complexity of their care, a multidisciplinary family-centered approach will likely be needed in which clinicians are able to engage patients and family members in meaningful discussions about serious illness preferences and prognosis (³¹).

Our results should be interpreted with the following limitations in mind. First, our cohort was composed mainly of patients receiving in-center hemodialysis in two US metropolitan areas, and findings may not be generalizable to patients who are hospitalized, those receiving home dialysis, those living in other parts of the country, or members of other groups not well represented in our study. Second, although survey questions were developed and adapted on the basis of the published literature (²,⁴¹,⁴⁹,⁵⁰), most have not been subjected to psychometric testing and thus may yield different results in other patient samples. Third, only a minority of enrolled patients had a family member who agreed to participate in our study, which may limit the generalizability of our findings to the overall cohort and to other populations. On the other hand, because we enrolled a select group of family members who were relatively involved in patients’ care, percentage agreement probably exceeds what we might expect for the overall cohort. Fourth, the relatively small sample size may limit the precision of our estimates of agreement between patients and family members, particularly in subgroup analyses. Fifth, the structure of some questions differed for patients and family members, potentially reducing opportunities for agreement. Finally, we did not ask about the patient’s transplant eligibility status, which could shape how they responded to some questions, nor do our results offer insight into the circumstances under which patients might consider discontinuing dialysis.

In conclusion, family members of people undergoing maintenance dialysis enrolled in our study lacked a detailed understanding of patients’ end-of-life wishes. Although most patient-family member dyads had discussed broad treatment preferences, a minority had spoken about dialysis discontinuation or hospice. Although family members had a fair understanding of patients’ CPR preferences, there was less agreement in responses to other questions about end-of-life care. Our findings suggest that there are significant deficiencies in our current approach to planning for serious illness among people undergoing dialysis and highlight the potential value of a more family-centered approach.

Disclosures

C.R. Butler reports employment with the Veterans Affairs (VA) Puget Sound Health Care System. J.R. Curtis reports other interests/relationships with Cambia Health Foundation and the National Institutes of Health (NIH). P.L. Hebert reports employment with the VA Puget Sound Health Care System. M. Kurella Tamura reports employment with the VA Palo Alto Health Care System; receiving honoraria from the American Federation for Aging Research; serving as an associate editor of CJASN; and serving as a scientific advisor or member of the Beeson External Advisory Committee and the Clin-Star Advisory Board. A.M. O’Hare reports employment with the VA Puget Sound Health Care System; consultancy agreements with the Pathways Project (supported by a grant from the Moore Foundation to the Coalition for the Supportive Care of Kidney Patients); receiving research funding from the Centers for Disease Control and Prevention, NIH, and VA HSR&D; and receiving honoraria from Chugai Pharmaceutical Co. Ltd., the Coalition for the Supportive Care of Kidney Patients, Dialysis Clinics Inc., Fondation Devenir, Fresenius Medical Care, Hammersmith Hospital, the Japanese Society of Dialysis Therapy, Kaiser Permanente Southern California, the University of California San Francisco, the University of Pennsylvania, and UpToDate. A.M. O'Hare reports serving as an associate editor of American Journal of Kidney Diseases; on the editorial boards of American Journal of Kidney Diseases, CJASN, JASN, and JAMA Internal Medicine; on the external advisory panel to the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) for the Chronic Renal Insufficiency Cohort; and as a member of the United States Renal Data System Steering Committee. F. Saeed reports receiving research funding from an American Society of Nephrology career development award, NIDDK K-23, and the Renal Research Institute. E.K. Vig reports employment with the Veterans Health Administration and serving on the editorial boards of American Journal of Hospice and Palliative Medicine and Journal of Pain and Symptom Management. All remaining authors have nothing to disclose.

Funding

A.M. O’Hare reports that this work has been supported by National Institute of Diabetes and Digestive and Kidney Diseases grant U01DK102150 and National Heart, Lung, and Blood Institute grant T32HL125195-04. F. Saeed is funded by National Institute of Diabetes and Digestive and Kidney Diseases award K23DK121939 and a Renal Research Institute grant and is a recipient of the Carl W. Gottschalk Research Scholar Grant.