Preferences, Knowledge, Communication and Patient-Physician Discussion of Living Kidney Transplantation in African American Families

Abstract

It is unknown whether patient-physician discussion about live kidney transplantation (LT) among African Americans (AA) is affected by preferences, knowledge and family discussions regarding LT. We recruited 182 AA dialysis patients and their families and assessed the relation of preferences, knowledge and family discussions regarding LT to the occurrence of patient-physician discussion using multivariable logistic regression. Most patients (76%) desired a transplant, and most patients (93%), spouses (91%) and children (88%) had knowledge of LT. Nearly half of the families discussed transplantation. Only 68% of patients and less than half of their spouses (41%) and children (31%) had discussed transplantation with physicians. Patient-physician discussion was more common among patients: whose spouses acknowledged their interest in transplantation (adjusted odds ratio (AOR) (95% CI):3.5 (1.61–7.8)); who discussed transplantation with spouses (AOR(95% CI):5.25 (2.22–12.41)); whose spouses agreed that they discussed transplantation with patients (AOR (95%CI):5.20 (1.76–15.37)) and whose children discussed transplantation with patients' physicians (AOR (95%CI):7.4 (1.3–40.0)). Universal patient-physician discussion of LT does not occur despite patient preferences. Rates of family-physician discussions are low, and rates of family discussions vary. Early family-physician discussion of LT, use of allied health professionals to promote family discussion of LT and the institution of culturally appropriate programs to enhance discussion of LT in AA families could improve rates of discussion and enhance patients' access to LT.

Introduction

While the public health burden of chronic kidney disease (CKD) is substantial for the entire U.S. population (more than 10 million adults having some kidney damage (serum creatinine levels ≥1.5 mg/dL) and more than 400 000 persons with end-stage renal disease (ESRD)), certain ‘high-risk’ groups of persons bear the burden of the CKD more heavily than others (1–4). Over the past 20 years, African Americans have consistently had CKD and ESRD incidence and prevalence rates 4–6 times greater than those of their White counterparts after adjustment for age and gender (5).

Ethnic/racial (terms describing both cultural and biological heritage; henceforth used interchangeably) (6,7) disparities in the burden of CKD are exacerbated by ethnic/race disparities in the receipt of kidney transplantation in the United States, with persistent trends over the past 10 years demonstrating African Americans are less likely to receive kidney transplants than Whites (8,9). Lower rates of kidney transplantation for African Americans compared to Whites have been attributed to a variety of reasons, including immunological incompatibility of deceased donor kidneys, lower rates of referral of African Americans for transplantation, less access to health care and less desire on the part of African Americans for kidney transplantation (10–18). Disproportionately high rates of ESRD among African Americans and improved outcomes for persons receiving kidney transplantation magnify the need to identify potentially modifiable barriers to the receipt of transplantation for African Americans (19,20).

Live kidney transplantation (LT) not only offers improved outcomes over deceased donor transplantation but may also bypass certain barriers to transplantation for African Americans on dialysis, including immunological incompatibility issues and longer waiting times typically experienced for African Americans on the deceased donor kidney waiting list (21,22). Despite technological improvements in the live donor transplant procedure and widely publicized successes of LT, African Americans with ESRD continue to trail behind their Caucasian American counterparts with respect to the receipt of LT (16,20,23–27).

Patient-physician discussion about transplantation and LT as a treatment option for dialysis patients, represents a key step in the process of obtaining LT. While not all patients on dialysis are clinically suitable for or desire transplantation, universal discussion of transplantation provides a mechanism through which patients and their physicians can address patients' treatment preferences and suitability openly. When patients are clinically suitable and desire transplantation, discussion might also facilitate access to LT. In many dialysis networks, health care providers' documentation of universal discussion of clinical options for transplantation, including LT, is strongly encouraged (28,29).

While prior studies have shown that African Americans are less likely to discuss transplantation with physicians when compared to their White counterparts, little is known about how patient and family knowledge of LT and patterns of patient, physician and family communication regarding LT influence patient-physician discussions of LT (14). Because the receipt of LT depends not only on patients' desires for LT but also on the identification of willing potential donors (especially family members), improved understanding of factors which may affect patient, family and physician involvement in discussions regarding LT, may shed light on ways in which LT may be increased in African Americans.

In a study of African American dialysis patients and their families, we assessed the relation of patient preferences, patient and family knowledge of LT and patient-family discussion of LT with patient-physician discussion of LT.

Methods

Study participants and data collection

As part of a family investigation of nephropathy genes in African American dialysis patients (the Family Investigation of Nephropathy and Diabetes study), we performed a cross-sectional study of 182 African American prevalent dialysis patients with no prior history of transplantation, their spouses/partners (n = 182) and their related children (n = 81) in Maryland, Pennsylvania and Tennessee from August 2001 to December 2003. Patients were recruited from dialysis units during their dialysis sessions by study staff. Patients who agreed to participate in the study identified their spouses/partners and children who might also be interested in participation. Patients' spouses/partners (henceforth referred to as ‘spouses’) were defined as patients' current significant others who might or might not be living with patients at the time of the survey. Children were defined as patients' adult children (age 18 years or older) from current or past marriages/partnerships.

Using a standardized questionnaire, we asked patients and their families questions about: (i) patients' preferences and expressed interest regarding LT, (ii) patients' and their families' knowledge of LT as a treatment option for dialysis patients, (iii) the occurrence of discussions regarding LT among family members and (iv) patients' and families' discussions regarding LT with patients' physicians or health care providers (henceforth termed ‘physicians’) (Table 1). These questions were based on previously validated questions posed in a regional study of racial differences in preferences regarding transplantation (14). We also asked study participants to report their demographic characteristics (age, gender, education, employment) and whether they had been told by a doctor that they had any of the following conditions: hypertension, diabetes, congestive heart failure, high cholesterol, myocardial infarction, balloon angioplasty or bypass surgery and stroke or transient ischemic attack. Questionnaires were administered separately to patients, spouses and children via telephone or in-person interviews, and they were completed by study participants independently from other family members involved in the study. The study protocol and questionnaire were approved by the Johns Hopkins Medicine Institutional Review Board.

Table 1.  Questions posed to patients, spouses and children in separate standardized questionnaires* regarding patient preference for transplantation, patient-family knowledge and patient-family discussions regarding transplantation

Question asked [answers possible]	Study participant
	Patient	Spouse	Child
*All questionnares completed independent from other fmily members. •†Indicates question posed to study participant in their questionnaire.
Patient preference and expressed interest for transplantation
Do you wish to have a kidney transplant?[yes/no]	•†
Has your spouse/parent ever expressed an interest in receiving a kidney transplant?[yes/no]		•	•
Knowledge regarding living related transplantation
Do you think a living person can donate a kidney to a person on dialysis [yes/no]?	•	•	•
Discussion regarding living-related transplantation among family members
Have you ever discussed options for a living-related kidney transplant with your spouse?[yes/no]	•
Have you ever discussed options for a living-related kidney transplant with your child?[yes/no]	•
Have you ever discussed living-related kidney donation with your partner/spouse?[yes/no]		•
Have you ever discussed the possibility of donating your kidney to your parent?[yes/no]			•
Discussion regarding living-related transplantation with patient's physician or health care provider
Have you ever discussed options for a living-related kidney transplant with your physician or health care provider?[yes/no]	•
Have you discussed living-related kidney donation with your spouse's/parent's physician or health care provider?[yes/no]		•	•

Statistical analysis

We used descriptive statistics to quantify patients' and their families' reports of patients' preferences and expressed interest regarding transplantation, patients' and their families' knowledge regarding LT and the occurrence of patient-family, patient-physician or family-physician (for spouses and children) discussion regarding LT. Using bivariate (chi square) analysis and multivariable logistic regression models (controlling for age, gender, education and employment status of patients, spouses and their children), we assessed the unadjusted and independent relation of patient preferences, patient/family knowledge regarding LT, occurrence of patient-family discussion regarding LT and occurrence of family-physician discussion regarding LT with the occurrence of patient-physician discussions regarding LT. In these models, predictor variables (e.g. patient preferences, patient knowledge of LT, family knowledge of LT, patient-family discussion of LT) were each included in separate models, which included potential confounders (patient and family age, gender, education, employment status). To ascertain whether the presence of medical conditions in patients' spouses and children (which could preclude their abilities to become live kidney donors) affected rates of patient-physician discussion of LT, we performed a sensitivity analysis in which we used both stratified logistic regression models and interaction terms to ascertain whether rates of discussion differed in families where medical conditions of spouses and children were present versus families where no medical conditions were present in spouses or children. Analyses were performed using STATA Statistical Software: Release 8.0 (Stata Corporation, College Station, TX).

Results

Participant characteristics

Study participants included 182 African American patients, 182 spouses and 81 children. Patients were similar in age, gender and employment status when compared to national data on African American dialysis patients (20). A majority of patients were male (63%), had at least high school education (60%), were disabled or unemployed (65%) and were receiving hemodialysis (91%) (Table 2). The median (interquartile range) time on dialysis for patients was 3 years (1–5). Most patients (92%) reported that they had been told by a doctor or other physicians, that they have hypertension, over a third (39%) had been told that they have diabetes and nearly a third (38%) had been told that they have high cholesterol. Less than a quarter of patients reported having had or been told that they had congestive heart failure, myocardial infarction, balloon angioplasty or bypass surgery or stroke or transient ischemic attack. Most spouses were female (63%), had at least high school education (70%) and nearly half were employed full time (45%). Nearly half (46%) of spouses had been told they have hypertension, and nearly a third (30%) had been told that they have high cholesterol, but few reported other chronic illnesses. Most children were female (62%), had at least high school education and were employed full time (59%). Less than a quarter (22%) of children had been told that they have hypertension and few reported other chronic illnesses (Table 2).

Table 2.  Participant demographic and medical characteristics

Characteristic		Study participants
		Patient (N = 182)	Spouse (N = 182)	Child (N = 81)
*Percentages may not add up to 100% due to missing values. na = not applicable
Age, years [mean (SD)]		53.8 (13.1)	52.4 (13.7)	31.9 (10.9)
Gender, n (%)*	Male	114 (63)	68 (37)	30 (37)
	Female	68 (37)	114 (63)	50 (62)
Education, n (%)*	Less than high school education	72 (40)	55 (30)	14 (17)
	At least high school education	110 (60)	127 (70)	67 (83)
Employment, n (%)*	Disabled or unemployed	118 (65)	36 (20)	14 (17)
	Employed full time	17 (9)	81 (45)	48 (59)
	Employed part time, retired or home maker	46 (25)	51 (28)	8 (10)
Type of dialysis, n (%)*	Hemodialysis	165 (91)	na	na
	Peritoneal dialysis	17(9)	na	na
Smoking status, n (%)*	Current smokers	65 (36)	65 (36)	21 (26)
Informed by doctor they have had
the following medical conditions, n (%)*	Hypertension	168 (92)	84 (46)	18 (22)
	Diabetes	71 (39)	19 (10)	7 (9)
	Congestive heart failure	49 (27)	3 (2)	0 (0)
	High cholesterol	69 (38)	55 (30)	8 (10)
	Myocardial infarction	22 (12)	5 (3)	0 (0)
	Balloon angioplasty or bypass surgery	22 (12)	4 (2)	0 (0)
	Stroke or transient ischemic attack	23 (13)	6 (3)	0 (0)

Patient preferences/expressed interest, family knowledge and family discussion regarding LT

Of all patients, 139 (76%) stated they desired a kidney transplant and the majority (n = 118, 86%) of these were at least 75% certain of their desire. More than half of patients' spouses (62%) and children (53%) reported that the patient had previously expressed interest in transplantation to them (Table 3). Most patients (93%), spouses (91%) and children (88%) had knowledge of LT as a treatment option for dialysis patients. Patients reported that they had discussed transplantation with their spouses and children in approximately 45% and 53% of families, respectively. Similarly, patients' spouses and children reported that they had discussed transplantation with patients in approximately 60% and 62% of families, respectively (Table 3). Within families, patients and spouses both agreed they had discussed transplantation with each other in 57 (31%) families and patients and children both agreed they had discussed transplantation with each other in 33 (41%) families with children. In bivariate analysis, patients aged 18–34 years, 35–49 years and age 50–64 years were more likely than persons aged 65 years or older to report that they desired a kidney transplant (n (%): 11 (92%), 52 (84%), 49 (82%) and 13 (45%), respectively, p < 0.001). Patients who were employed full time were most likely to report they desired a transplant when compared to those who were disabled/unemployed or those who classified themselves as employed part time, retired, homemakers or other (n (%): 15 (88.2%) vs 95 (81.2%) and 28 (62.2%), respectively, p = 0.012). Desire for transplant was not associated with patient knowledge of LT as a treatment option, patient education or patients' reports that they had discussed LT with their spouses or their children.

Table 3.  Patient preferences and expressed interest regarding transplantation, family knowledge regarding living-related kidney transplantation and family discussion about living-related kidney transplantation

	Response	Study participant
		Patient (N = 182) n (%)*	Spouse (N = 182) n (%)*	Child (N = 81) n (%)*
*Percentages may not add up to 100% due to missing values. ^†Greater missing values with this question may be present because not all patients enrolled in the study had chldren. Of those with children enrolled in this study, 84% of patients responded. na = not applicable.
Patient preference and expressed interest regarding transplantation
Do you wish to have a kidney transplant?	Yes	139 (76)	na	na
	No	41 (23)	na	na
Has your spouse/parent ever expressed an interest in receiving a kidney transplant?	Yes	na	112 (62)	43 (53)
	No	na	59 (32)	33 (41)
Knowledge regarding living related transplantation
Do you think a living person can donate a kidney to a person on dialysis?	Yes	169 (93)	166 (91)	71 (88)
	No	10 (6)	7 (4)	4 (5)
Discussion regarding living-related transplantation among family members
Have you ever discussed options for a living-related kidney transplant with your spouse?	Yes	83 (46)	na	na
	No	65 (36)	na	na
Have you ever discussed options for a living-related kidney transplant with your child?^†	Yes	76 (41)	na	na
	No	69 (38)	na	na
Have you ever discussed living-related kidney donation with your partner/spouse?	Yes	na	109 (60)	na
	No	na	64 (35)	na
Have you ever discussed the possibility of donating your kidney to your parent?	Yes	na	na	50 (62)
	No	na	na	26 (32)

Prevalence and predictors of patient-physician, family-physician and family discussion of LT

Nearly two thirds (68%) of patients reported that they had discussed LT with their physicians, but less than half of spouses (41%) and children (31%) reported discussing LT with patients' physicians. There was no statistically significant difference in rates of patient-physician or family-physician discussion of LT between families where patients reported that they desired transplantation versus families where patients reported that they did not desire transplantation (Figure 1).

Figure 1. Percentage of family members who report having discussion regarding living-related kidney transplantation with patients' physicians or healthcare providers according to whether patients desire transplantation.

Patient-physician discussion of LT was associated with family members' reports that patients expressed interest in receiving a kidney transplant, family discussion of LT and family-physician discussion regarding LT (Table 4). Patients whose spouses reported that their partners, had previously expressed interest in receiving a kidney transplant were statistically significantly more likely (3-fold greater odds) to report patient-physician discussion of LT when compared to patients whose spouses did not report their partners were interested in receiving a kidney transplant (adjusted percentage (95% CI): 79 (71–87) vs 55 (41–68), respectively, p < 0.01). Patients who reported that they had discussed LT with their spouses were statistically significantly more likely (5-fold greater odds) to report patient-physician discussion of LT when compared to patients who had not discussed LT with their spouses (adjusted percentage (95% CI): 85 (75–92) vs 52 (38–65), respectively, p < 0.01). Similarly, when patients and their spouses both acknowledged that they had discussed transplantation with each other, patients were statistically significantly more likely (5-fold greater odds) to report patient-physician discussion of LT, when compared to patients in families where patients and their spouses did not dually acknowledge that they had discussed transplantation or families where patients and their spouses both acknowledged that they had not discussed LT with each other (adjusted percentage (95% CI): 84 (72–92) vs 68 (53–80) and 49 (31–67), respectively, p < 0.01). Finally, patients whose children reported that they had previously discussed LT with their parent's physician were statistically significantly more likely (7-fold greater odds) to report patient-physician discussion of LT when compared to patients whose children had not discussed LT with their parent's physician (adjusted percentage (95% CI): 90 (69–98) vs 53 (35–69), p < 0.01, respectively). Patient demographic characteristics and patient and family knowledge about transplantation as a treatment option were not associated with patients-physician discussion of LT (Table 4).

Table 4.  Relation between different factors and patient-physician discussion of transplantation

Factors		Total N	Patient-physician discussion of live transplantation
			n**	Unadjusted percentage	Adjusted* percentage (95% CI)	Adjusted odds ratio (95% CI)
^aSingle model incorporating all patient demographics simultaneously. ^bModel adjusted for patient demographics (age, gender, education and employment status). ^cModel adjusted for patient demographics (age, gender, education and employment status) plus demographics of spouse (in models of variables reported by patients' spouses). ^dModel adjusted for patient demographics (age, gender, education and employment status) plus demographic. characteristics of child (in models of variables reported by patients' children). *Adjusted percentage represents frequency of patient-physician discussion in patients/families with and without factor as if the demographics of persons with and without factor were similar. ^†Statistically significant trend (p < 0.01). ^††Statistically significant trend (p = 0.02). **n = number of patients in each category who reported having discussed live transplantation with their physicians. ***Not able to calculate due to no reference group. ^¶Includes persons employed part time, retired and home makers. –No participants in this category.
Patient demographics^a
Age	18–34 years	12	10	75	82 (49–95)	1.0 (reference)
	35–49 years	61	41	69	67 (55–78)	0.45 (0.09–2.32)
	50–64 years	67	43	70	73 (62–83)	0.56 (0.11–2.95)
	65 years and greater	32	20	59	62 (45–77)	0.30 (0.05–1.79)
Gender^a	Male	112	78	70	71 (61–79)	1.0 (reference)
	Female	68	46	68	69 (57–79)	0.85 (0.43–1.69)
Education^a	Less than high school	72	46	64	65 (53–76)	1.0 (reference)
	At least high school	108	78	72	73 (63–81)	1.41 (0.72–2.77)
Employment^a	Disabled or unemployed	117	78	67	67 (57–75)	1.0 (reference)
	Employed full time	17	14	82	93 (65–99)	6.44 (0.81–51.5)
	Other^¶	45	31	69	69 (54–81)	1.25 (0.54–2.91)
Patient preference/interest for transplant
Patient desire for transplant^b	Does not desire transplant	41	25	61	63 (47–76)	1.0 (reference)
	Desires transplant	136	99	71	72 (64–79)	1.51 (0.67–3.37)
Spouse reports patient expressed interest^c	Partner did not express interest	58	30	52	55 (41–68)	1.0 (reference)
	Partner expressed interest	111	87	78^†	79 (71–87)^†	3.5 (1.61–7.8)^†
Child reports patient expressed interest^d	Parent did not express interest	33	19	58	64 (42–81)	1.0 (reference)
	Parent expressed interest	46	31	67	69 (51–84)	1.89 (0.48–7.51)
Family knowledge of LT
Patient knowledge^b	Not aware	10	5	50	50 (22–78)	1.0 (reference)
	Aware	169	118	70	71 (63–77)	2.02 (0.54–7.63)
Spouse knowledge^c	Not aware	7	5	71	72 (32–93)	1.0 (reference)
	Aware	164	113	69	71 (63–78)	0.63 (0.10–3.96)
Child knowledge^d	Not aware	5	5	100	***	***
	Aware	73	44	60	63 (48–77)	***
Family discussion regarding LT
Patient discussion with spouse^c	Did not discuss	64	31	48	52 (38–65)	1.0 (reference)
	Discussed	82	69	84^†	85 (75–92)^†	5.25 (2.22–12.41)^†
Patient discussion with child^d	Did not discuss	24	16	67	77 (48–92)	1.0 (reference)
	Discussed	43	27	63	67 (46–84)	0.59 (0.12–2.86)
Spouse discussion with patient^c	Did not discuss	64	40	63	63 (51–74)	1.0 (reference)
	Discussed	107	78	73	76 (66–84)	1.72 (0.83–3.59)
Child discussion with patient^d	Did not discuss	26	15	58	65 (41–83)	1.0 (reference)
	Discussed	53	35	66	68 (50–82)	1.22 (0.31–4.83)
Patient and spouse agreement regarding discussion with each other^c	Agree they did not discuss LT with each other	31	15	48	49 (31–67)	1.0 (reference)
	Do not agree	51	32	63	68 (53–80)	2.03 (0.74–5.57)
	Agree they both discussed LT with each other	56	47	84^†	84 (72–92)^†	5.20 (1.76–15.37)^†
Patient and child agreement regarding discussion with each other^d	Agree they did not discuss LT with each other	37	22	59	68 (46–84)	1.0 (reference)
	Do not agree	0	0	0	–	–
	Agree they both discussed LT with each other	33	21	64	68 (45–84)	0.99 (0.26–3.75)
Family-physician discussion regarding LT
Spouse-physician discussion^c	Discussion did not occur	98	64	64	66 (55–75)	1.0 (reference)
	Discussion occurred	73	54	74	78 (66–86)	1.76 (0.83–3.72)
Child-physician discussion^d	Discussion did not occur	52	27	52	53 (35–69)	1.0 (reference)
	Discussion occurred	27	23	85^†	90 (69–98)^††	7.41 (1.34–40.06)^††

Effect of spouses' and children's medical conditions on patient-physician discussion of LT

In sensitivity analyses investigating the effect of medical conditions which might affect patients' spouses' and children's abilities to donate a live kidney, patients reported having discussed LT with their physicians in 67% of families where their spouses and children reported having no medical conditions that could preclude them from becoming live kidney donors. All other results were similar to main findings, indicating the presence of medical conditions in family members did not impact rates of patient-physician discussion of LT.

Discussion

Given disparities in African Americans' access to deceased donor kidney transplants, one might assume patient-physician discussion of LT as a treatment option would be universal for African American dialysis patients. In this study of 182 African American dialysis patients and their families, we found that, although most patients desire transplantation and although knowledge of LT among patients and their families is common, patient-physician discussion of LT is not universal, rates of family-physician discussion regarding LT are low and rates of family discussion regarding LT vary among family members. Family recognition of patients' interests in receiving a kidney transplant, patient-family discussions regarding LT and family-physician discussions regarding LT were associated with greater rates of patient-physician discussion of LT.

Our data indicate that even when African American patients report that they desire transplantation, patient-physician discussion does not occur in nearly one-third of cases. These results are consistent with previous findings of less than the optimal rates of discussion of transplantation between African American patients and their physicians in a regional study of dialysis patients (14). Suboptimal rates of patient-physician communication represent a potentially modifiable barrier to disparities in the receipt of transplantation for Africans Americans. Reasons for suboptimal patient-physician discussion could include patients' distrust of physicians or fear of surgical procedures and patients' subsequent reluctance to ask about new therapeutic options, physicians' assumptions that patients are not interested in transplantation and patients' or physicians' perceptions that patients are not medically suitable for LT (13,24,30–32). In a recent regional study of nephrologists caring for African American and White dialysis patients, over one-third of nephrologists perceived difficulties in patient-physician communication or trust as a potential contributor to ethnic/race disparities in the receipt of transplantation (33). The magnitude of potential improvement in LT rates for African Americans with greater patient-physician discussion is unclear. However, there is mounting evidence that improvements in patients' trust of health providers, patient-physician communication and participatory decision making in the patient-physician relationship can enhance patients' understanding of risks and benefits of potential therapies (34). Such improvements could help patients make informed treatment decisions that are most compatible with their preferences, leading to greater satisfaction with outcomes (35–37).

Our results confirm the importance of family participation in discussions regarding LT and reflect the complex nature of patient decisions pertaining to LT. Because LT involves the identification of eligible donors for transplantation as well as the decision to undergo a major surgical procedure, discussion of LT with close family members (particularly spouses, who may not only serve as potential donors, but may also provide important social and emotional support for patients) may not only help patients finalize decisions regarding LT and facilitate the transplantation process, but may also help with improvement of clinical outcomes following transplantation (38–41). However, for some families, the ethical complexities of LT may interfere with discussion of LT. Prior studies have demonstrated that family discussions about LT can have psychological consequences, including depression and family conflict (42).

While patterns of family discussion regarding LT are likely to be heterogeneous, in some cases, patients, their family members and patients' physicians may shy away from family discussion of LT to avoid perceived coercion of potential donors (43). In addition, some patients, families and their physicians may be hesitant to discuss live transplantation because of perceived increased medical risks to family members who might already have medical conditions that could increase the risks of donation. However, our subgroup analyses (restricted to families where spouses and children did not report having such medical conditions) did not support this latter possibility. Some clinicians might argue family discussion of transplantation could be helpful, regardless of perceptions of patient or family donor willingness to consider live transplantation. One study on experiences of donors, recipients and third parties after transplantation suggested that open discussion and better education about donation might provide emotional support and reduce psychological distress for both potential donors and potential recipients (44). Nonetheless, physicians encouraging family discussion of LT should be aware of potential psychological risks associated with discussion of LT and should be prepared to help families obtain needed psychosocial support throughout the LT decision-making process (45). At a minimum, some renal dialysis networks strongly encourage patient-physician discussion of LT be included in their discussions of patients' treatment options, irrespective of family willingness to participate in such discussions (28,29).

Avenues through which improved patient-physician and family discussion of LT could be achieved include promotion of early patient-physician and family discussions of LT, the utilization of allied health professionals to assist families in the decision-making process and implementation of culturally appropriate structured programs designed to enhance African American patient and family education regarding LT. Early family-physician meetings to discuss LT (even in late stages of CKD prior to the need for dialysis) could serve to increase patient and family awareness of LT and provide an opportunity for the identification of medically eligible family donors. Such discussions could enhance patients' access to live transplantation and increase the possibility of preemptive transplantation, demonstrated to improve clinical outcomes and avoid costs and psychological impacts of dialysis therapy (46–48). Early discussions would depend on the timely referral of patients with advanced CKD to a nephrologist or transplant surgeon (47,48). Health professionals (e.g. social workers or counselors) could aid families' decision making, facilitate open discussion between patients and families regarding concerns related to live donation and provide much needed psychosocial support to patients and families as they face potentially challenging ethical dilemmas related to LT (49–51). Structured educational programs have been demonstrated to inform patients and their families about treatment options (including LT), encourage family discussion and enhance recruitment of relatives for donation (52,53). To be most effective, programs should be culturally tailored to address specific concerns African Americans may have, regarding the transplantation process, including mistrust of health care professionals, knowledge of the need for transplantation in the African American community and potential religious/spiritual concerns (54–56). Culturally tailored programs have previously been successful in raising African American's awareness of the need for donation and in addressing issues such as mistrust of physicians (57).

Limitations of this study deserve mention. First, our study population was recruited from three states, limiting the generalizability of our results to African American dialysis patients and their families living in other areas of the United States. Second, because we did not compare African American families to White families, we cannot ascertain whether patterns of family communication differ between these two groups, thus limiting our ability to draw conclusions regarding racial differences in patterns of communication as a potential explanation for ethnic/race disparities in rates of LT. Nonetheless, our data regarding rates of patient-physician discussion of LT in African Americans are similar to previously published data demonstrating ethnic/race disparities in rates of patient-physician discussion of LT and our study lends insight into potential mechanisms for suboptimal rates of patient-physician communication (14). Third, the cross-sectional nature of our study limits our ability to infer the causal nature of associations we observe between family communication and patient-physician communication about LT. It is unclear whether family discussion of LT leads to patient-physician discussion or vice versa. Fourth, patient, physician and family member reports of discussions regarding LT are subject to recall bias and may differ. Although we did not contact patients' physicians to ascertain their perceptions regarding the frequency of patient-physician discussion of LT, there was consistency between patient and family member reports of patient-family discussions regarding LT, adding validity to these reports. Our study did not collect information on patients' physicians' race. It is possible that patients' recollections regarding their discussions with physicians could be affected by their perceived racial/ethnic similarity with their physicians, as other studies have demonstrated race concordance between patients and their physicians is associated with patient satisfaction with care (58). In addition, the design of the original study dictated that only one child be recruited for the study and in over half of families, no child participated. Children who participated in the study may have better relationships with their parents than children who did not participate, and may be more likely to discuss transplantation with parents than non-participating children. Fifth, although patients provided self-reports of medical conditions, lack of clinical data (such as chart reviews) limited our ability to determine which patients were eligible for transplantation. Some might argue that rates of patient-physician discussion could be affected by patients' clinical eligibility for transplantation. However, open discussion of treatment options, regardless of clinical eligibility, provides a key mechanism through which patients and physicians can share in the clinical decision-making process (38,59,60). Finally, while we ascertained frequencies of discussions regarding LT, we did not collect information which could provide context for why discussion did or did not occur. In addition, although our questions were based on previously validated questions regarding patient preferences for transplantation and patient-physician discussion regarding transplantation, few studies have been performed to understand the interplay of patient, family and physician communication; results may differ in patients and families of different ethnic/racial and cultural backgrounds (14). Further studies are needed to more clearly understand the contexts within which patient, family and physician discussion of LT might occur. Notwithstanding these limitations, our study represents one of the largest systematic assessments of patterns of discussion of LT in African American families and provides important insight to potentially modifiable barriers to improving ethnic/race disparities in LT.

In conclusion, most African American dialysis patients desire transplantation and most African American families have the knowledge of LT, but universal patient-physician discussion of LT does not occur, rates of family-physician discussion of LT are low and rates of family discussions regarding LT vary among family members. Family discussion regarding LT is related to patient-physician discussion of LT. Enhanced early family-physician discussion of LT, the employment of allied health professionals to promote family discussion of LT and the institution of culturally appropriate programs to enhance discussion of LT in African American families could serve to improve suboptimal rates of discussion and enhance African American patients' access to LT.

Acknowledgment

This study was supported by U01 DK57304 [National Institute of Diabetes & Digestive & Kidney Diseases (NIDDK)] (Drs. Klag, Parekh, Kao, Ms. Fink, Ms. Meoni); The Harold Amos Medical Faculty Development Program [Robert Wood Johnson Foundation] (Dr. Boulware); grants RO1DK59616-02S1 [NIDDK] (Dr. Boulware), K24-DK02856[NIDDK] (Dr. Klag), Carl W. Gottschalk Research Scholar Grant [American Society of Nephrology] and 1K23DK02872-01A1 [NIDDK] (Dr. Parekh) and K24-DK02643[NIDDK] (Dr. Powe).