- Email: [email protected]
Nonreferral and nonacceptance to dialysis by primary care physicians and nephrologists in Canada and the United States
Nonreferral and Nonacceptance to Dialysis by Primary Care Physicians and Nephrologists in Canada and the United States Mohamed Sekkarie, MD, Mihaela Cosma, MD, and David Mendelssohn, MD ● Research from Canada and the United States suggests that not offering dialysis to patients who might benefit still occurs. This study was conducted to investigate nonreferral and nonacceptance to dialysis by primary care physicians (PCPs) and nephrologists in these countries. We surveyed a random sample of Canadian and US PCPs and nephrologists concerning their attitudes toward and experience with withholding dialysis in patients with advanced chronic renal failure. In response to a question about whether the physician believes there should be an age beyond which dialysis should not be offered, 12% of Canadian PCPs, 20% of US PCPs, 4% of Canadian nephrologists, and 9% of US nephrologists answered yes. When asked about their recommendations concerning dialysis initiation in 10 vignettes of patients with impending end-stage renal disease (ESRD), the responses of Canadian and US physicians were similar. PCPs compared with nephrologists were less likely to recommend dialysis in cases with physical illnesses and more likely to recommend it in cases with neuropsychiatric impairments. Over a 3-year period, 13% of Canadian PCPs and 19% of US PCPs reported nonreferral to dialysis at least once. Withholding rates were 25% for Canadian PCPs, 16% for US PCPs, 13% for Canadian nephrologists, and 17% for US nephrologists. We conclude that although nonreferral of patients who might benefit from dialysis still occurs, it does not seem to be common, and the attitudes of Canadian and US physicians toward this issue are similar and could not entirely account for the much greater incidence of treated ESRD in the United States. PCPs and nephrologists should continue to be educated about the modern criteria for patient selection for dialysis. © 2001 by the National Kidney Foundation, Inc. INDEX WORDS: Dialysis; referral; withholding; primary care physicians (PCPs).
I
N ALL PARTS OF the world, incidence rates of end-stage renal disease (ESRD) are calculated based on the number of patients started on renal replacement therapy. These calculations do not include those individuals in whom dialysis therapy was withheld (ie, never started). Different withholding patterns may explain, at least in part, the large variation in acceptance rates reported in different parts of the world1 and may also be a factor contributing to varying mortality rates of dialysis patients among countries.2 Studying withholding of dialysis is important from an ethical perspective. Not offering renal replacement therapy to patients who might benefit from it because of their age, a misconception that they are “not candidates for dialysis,” or From the Department of Medicine, Division of Nephrology, West Virginia University, Morgantown, WV; the Department of Medicine, Bluefield Regional Medical Center, Bluefield, WV; and the Department of Medicine, Division of Nephrology, University of Toronto, Canada. Received September 25, 2000; accepted in revised form January 12, 2001. Presented in part as an abstract at the 32nd Annual Meeting of the American Society of Nephrology, Miami Beach, FL, 1999. Address reprint requests to Mohamed Sekkarie, MD, 510 Cherry St, Bluefield, WV 24701. E-mail: [email protected] © 2001 by the National Kidney Foundation, Inc. 0272-6386/01/3801-0006$35.00/0 doi:10.1053/ajkd.2001.25179 36
other factors will lead to their premature death. The recently published Renal Physicians Association and American Society of Nephrology clinical practice guidelines on shared decision making in the appropriate initiation and withdrawal from dialysis3 lists very few conditions in which not offering dialysis to a patient (or their proxy decision maker) who wants it is believed to be appropriate. These conditions include patients who have irreversible profound neurological impairment, patients in whom dialysis is not feasible technically (inability to establish a dialysis access, inability to cooperate with the dialysis procedure), and patients with estimated survival less than 6 months. Even in the last group, it is recommended that the renal team should be sensitive to patients’ goals and individual circumstances, such as wanting to live to attend a special family event.3 Recent research from Canada and the United States suggests that not offering dialysis to patients who might benefit still occurs, and that this decision is often made at the level of primary care physicians (PCPs) without the involvement of a nephrologist.4-6 This literature from the early and mid-1990s suggests that contrary to ethical principles, old age seemed to be a reason for withholding dialysis in some cases. The difference in the incidence of treated ESRD between these two countries is large and
American Journal of Kidney Diseases, Vol 38, No 1 (July), 2001: pp 36-41
NONREFERRAL TO DIALYSIS
difficult to explain, even after taking into account the greater proportion of blacks in the US population. In 1997, the incidence rate in the United States was 296 per million population (229 per million for whites), whereas it was only 116 per million in Canada.7 This study was conducted to investigate nonreferral and nonacceptance to dialysis by PCPs and nephrologists in both Canada and the United States. We tried to answer the following questions: (1) Is dialysis being withheld from patients who might benefit from it? and (2) Do differences in withholding dialysis between Canada and the United States explain the large difference in acceptance rates to dialysis programs in these countries. METHODS We surveyed a random sample of Canadian and US PCPs and nephrologists concerning their attitudes toward and experience with withholding dialysis in patients with advanced chronic renal failure (CRF). Names of physicians were obtained from the College of Family Practice of Canada and the American Academy of Family Practice for family physicians, the American Society of Nephrology and the Canadian Society of Nephrology for nephrologists, and a commercial vendor that contracts with the American Medical Association for the names of US primary care internists. Random PCP lists were generated by computers of the providing agencies. All Canadian nephrologists were surveyed, and US nephrologists were chosen from the membership directory of the American Society of Nephrology using a random number table. Internists in Canada were not surveyed because they do not function as PCPs. Questionnaires were sent to 520 Canadian family physicians, 500 US PCPs (252 family physicians and 248 primary care internists), 299 Canadian nephrologists, and 250 US nephrologists. The surveys, which were mailed in 1999, asked questions about physicians’ demographics, practice type, and whether they believe that there should be an age cutoff beyond which dialysis should not be offered. In addition, physicians were given brief descriptions of 10 potential patients with ESRD and asked if they would recommend dialysis (Table 1). Nephrologists were asked to estimate how many patients with advanced CRF they personally treated who died of renal failure without initiation of dialysis therapy and how many they started on dialysis therapy in 1998. PCPs were asked how many patients with advanced CRF they personally treated who died of renal failure without nephrology referral and how many of their patients with ESRD were started on dialysis therapy during the previous 3 years. The survey instrument is available on request from the authors. Responses to the survey were kept anonymous and confidential. Forms were numbered to track nonresponders. Those who did not respond to the first survey were sent a second
37
mailing. Because of the lower response for US internists, a third mailing was sent to this group. Withholding rates for each group were calculated by dividing the number of patients withheld from dialysis therapy by the sum of the patients withheld from therapy and patients started on dialysis therapy in that group. Responses of the groups were compared by Fisher’s exact tests, and P less than 0.05 is considered significant. Data were analyzed using the statistical software SPSS (SSPS Inc, Chicago, IL). Some data were analyzed using Stata (version 6; Stata Corp, College Station, TX) and Excel (Microsoft Corp, Redmond, CA). Unless otherwise indicated, data presented in the result section pertain to responders.
RESULTS
After excluding such ineligible responses as those from physicians who retired or changed specialty or questionnaires that were returned because of the wrong address, the overall response rate was 49%. Canadian physicians had a better response rate than their US counterparts. Response rates and demographics of the respondents are listed in Table 2. Responses of US family physicians and primary care internists were similar; therefore, their data are presented together as US PCPs. Of the respondents, approximately 60% of Canadian and US PCPs and US nephrologists were in community-based group practices, and 75% of Canadian nephrologists were in academic medicine. Approximately 95% of PCPs indicated that they were board certified. Most PCPs had a nephrologist practicing within a 25-km radius of their practices. In response to the question about whether the physician believes there should be an age beyond which dialysis should not be offered, 12% of Canadian family practitioners, 20% of US PCPs, 4% of Canadian nephrologists, and 9% of US nephrologists answered yes. The differences were statistically significant for Canadian versus US PCPs (P ⫽ 0.03), nephrologists versus PCPs in both countries (P ⫽ 0.005 for Canada, P ⫽ 0.02 for the United States), but not Canadian versus US nephrologists. Those who answered yes suggested cutoff values that ranged between the ages of 65 to 100 years (median, 80 years); practice type and proximity to a nephrologist were not associated with the physicians’ position on this issue. The proportions of physicians recommending dialysis for each of the 10 vignettes are listed in Table 1 as percentages of physicians who recom-
38
SEKKARIE, COSMA, AND MENDELSSOHN Table 1.
The Vignettes and Proportion of Physicians Recommending Dialysis
Case
Cases
a
An 80-year-old patient with diabetes with an ejection fraction of 25% caused by ischemic cardiomyopathy, lives with a sitter and takes care of herself, normal mental status. A 73-year-old patient with metastatic prostate cancer, no pain, unable to walk because of cord compression, lives with family, normal mental status. A 90-year-old patient with hypertensive nephrosclerosis, lives alone, takes care of herself, seen by a visiting nurse weekly to supervise medication intake. A 45-year-old patient with persistent vegetative state for the last 3 months after a cardiac arrest, prognosis for neurological recovery is poor. A 35-year-old schizophrenic patient, very combative despite medications, not a competent decision maker, institutionalized. A 62-year-old patient with multi-infarct dementia, incontinent, feeds self, bed and chair confined, does not recognize family. A 70-year-old patient on chemotherapy for multiple myeloma, ambulatory, normal mental status. A 29-year-old paraplegic, permanently ventilator dependent, alert and coherent. A 68-year-old patient with advanced chronic obstructive, pulmonary disease caused by smoking, chronically on oxygen, becomes dyspneic after walking ⱖ5 feet. A 30-year-old blind patient with diabetes.
b
c
d
e f g h i
j
Canadian PCP (%)
US PCP (%)
Canadian NEP (%)
US NEP (%)
74
70
81
85
50
54
59
60
76
71
80
88
5
3
1
1
59
54
32
36
9
12
1
5
88
91
98
100
90
87
79
78
63
62
61
73
99
99
99
100
NOTE. Cases are of patients potentially close to needing dialysis. Physicians were instructed to assume that the patient and/or proxy decision maker is ambivalent about whether to accept chronic dialysis therapy and requests the physician’s recommendation. Based on the presented data, physicians were asked to indicate whether they would recommend dialysis. The percentage of physicians who recommended dialysis is shown. Canadian PCP versus US PCP, P ⬍ 0.05 for case e only; Canadian NEP versus US NEP, P ⬍ 0.05 for case f only; Canadian PCP versus Canadian NEP, P ⬍ 0.05 for cases d, e, g, and h; and US PCP versus US NEP, P ⬍ 0.05 for cases a, c, and g. Abbreviation: NEP, nephrologist.
mended dialysis therapy. When the responses of each country’s physicians were combined, in none of the cases did the difference between Canadian and US physicians reach statistical Table 2.
Demographics and Response Rates of Physicians
Eligible Responded Response rate (%) Men (%) Mean age (y)
Canadian PCP
US PCP
Canadian NEP
US NEP
490 263 54 55 41
452 176 38 77 46
258 166 64 80 48
231 93 40 88 47
NOTE. Differences in response rates: Canadian PCP versus US PCP, P ⫽ 0.001; Canadian PCP versus Canadian NEP, P ⫽ 0.005; US PCP versus US NEP, P ⫽ not significant; Canadian NEP versus US NEP, P ⫽ 0.001. Abbreviation: NEP, nephrologist.
significance. Conversely, when results of all PCPs were compared with those of all nephrologists, PCPs were less likely to recommend dialysis in cases a, b, c, and g and more likely to recommend dialysis in cases d, e, f, and h. Responses for cases i and j did not differ. Over a 3-year period, 13% of Canadian PCPs and 19% of US PCPs reported nonreferral to dialysis at least once. Withholding rates were 25% for Canadian family practitioners, 15% for US PCPs, 13% for Canadian nephrologists, and 16% for US nephrologists. The differences between Canadian PCPs versus US PCPs, Canadian nephrologists versus US nephrologists, and Canadian PCPs versus Canadian nephrologists were statistically significant. Table 3 lists the median number of patients referred and not referred for dialysis therapy per PCP per 3 years. Table 4 lists the median number
NONREFERRAL TO DIALYSIS Table 3.
39
Median Number of Patients Referred and Not Referred Per PCP Per 3 Years
Referred Not referred
Canadian PCP
US PCP
1 0
2 0
of patients started on and withheld from dialysis therapy per nephrologist per year. PCPs who withheld dialysis listed on the average three reasons per nonreferred patient (n ⫽ 67). Patient refusal, the most common cause, was cited as one of the reasons for nonreferral in 60% of the cases, followed by end-stage heart, lung, or liver disease (48%); old age (46%); overall frail condition (39%); and then family refusal, dementia, and terminal cancer (⬃31% each). This distribution of causes did not differ between the two countries. Old age was cited as the sole cause only once. DISCUSSION
To the best of our knowledge, this study is the first that compares the attitudes and experiences of Canadian and US physicians, both PCPs and nephrologists, toward the issue of withholding dialysis therapy. Contrary to a widely held belief,1 our findings do not support the theory that the difference in incidence rates of ESRD in these countries is primarily caused by more liberal attitudes of physicians concerning acceptance criteria in the United States. Canadian physicians were less likely to consider age a reason to withhold dialysis, and their attitudes toward recommending dialysis in the hypothetical cases did not differ from those of US physicians. PCPs in both countries appear to understand modern acceptance criteria reasonably well. Patient refusal was the most commonly cited cause for nonreferral to dialysis by PCPs. It is likely, at least in some cases, that these nonreferred patients were not adequately informed about the dialysis procedure and its potential benefits. The design of our study did not allow us to test this hypothesis. There is some suggestion in the literature that Canadian nephrologists might be less liberal in accepting patients to dialysis than US nephrologists. In a 1998 survey of Canadian, British, and US nephrologists, McKenzie et al,8 using case
vignettes, found that American nephrologists were slightly more likely to offer dialysis than their Canadian and British counterparts. Hirsch et al9 from Nova Scotia, Canada, reported that in 1992, one quarter of their referred patients with ESRD were not offered dialysis, which contrasts with a 7% withholding rate by nephrologists of West Virginia in 1995.6 Our study showed a less liberal attitude by Canadian nephrologists only in the case of the patient with multi-infarct dementia. Although random error and different sensitivities of the research instruments may have had a role in these small discrepancies, we believe that because our survey is more recent and is a national rather than a provincial experience, it is more reflective of the current attitudes of Canadian nephrologists. If the attitudes of PCPs and nephrologists are not different in Canada and the United States, then what might explain the large difference in incidence rates between the two countries? It is theoretically possible that more patients with ESRD die without being diagnosed in Canada, but this seems unlikely. Under universal health insurance, access to PCPs should be easier in Canada than in the United States. Other potential explanations for the greater acceptance rates in the United States need to be considered, including a genuinely greater incidence of chronic renal insufficiency in the United States (it is estimated from the Third Nutrition and Health Examination Survey and the 1990 US census that 0.4% of the total US population aged ⱖ12 years have a serum creatinine value ⬎ 2.0 mg/ dL10; we are not aware of similar data from Canada), lower mortality from competing diseases in the pre-ESRD period in the United States,11 slower rate of progression of kidney disease in Canada secondary to easier access to predialysis care, and/or later initiation of dialysis in Canada. Finally, access to a nephrologist in Table 4.
Started Withheld
Median Number of Patients Started on and Withheld From Dialysis Therapy Per Nephrologist Per Year Canadian NEP
US NEP
15 2
20 3
Abbreviation: NEP, nephrologist.
40
Canada is only through referral from another physician. Conversely, in the United States, direct access through self-referral to a nephrologist is possible and may allow for easier access to renal replacement therapy. Data to support or rule out these potential explanations are lacking. In both the United States and Canada, concerns have been raised about the issue of depriving certain patients, especially the elderly, from potentially beneficial dialysis therapy.5,6,12 In a survey of PCPs in Ontario in 1994, 14.2% of these physicians reported withholding dialysis on their own at least once over a period of 3 years. In the same study, these physicians were less likely to refer older patients.5 Similar findings were found in a survey of a group of PCPs in West Virginia in 1995 in which 60% of the surveyed PCPs cited age as a major factor in dialysis decision making.6 Our study, performed at a later date, also shows that this phenomenon continues to exist; 12% of Canadian and 20% of US PCPs believe there should be an age cutoff beyond which dialysis should not be offered. Although these percentages are smaller for nephrologists, they are still a cause for concern. The Institute of Medicine Committee for the Study of Medicare End-Stage Renal Disease Program explicitly rejected age as a criterion for patient acceptance,13 and many elderly patients have a good quality of life and reasonable life expectancy on dialysis.14 Educating physicians about this issue should be a priority. Consistent with our findings on the age cutoff issue, the responses of US and Canadian physicians to the hypothetical cases were also similar. However, small but statistically significant differences between PCPs’ and nephrologists’ recommendations were seen. In the cases of patients with ischemic cardiomyopathy (case a), prostate cancer (case b), nephrosclerosis (case c), and myeloma (case g), PCPs were less likely to recommend dialysis. The opposite was true for cases with neurological or psychiatric impairment (cases d, e, f, and h) in which the patients were incidentally younger. The opinions of the nephrologists on many of these cases also differ. Our questions regarding the hypothetical cases assumed that the patient and/or proxy decision maker was ambivalent about accepting renal replacement therapy and were thus designed to
SEKKARIE, COSMA, AND MENDELSSOHN
study the attitudes of physicians toward whether to recommend dialysis without the influence of patient autonomy. This is important because patients and proxy decision makers are generally strongly influenced by the recommendations of their physicians.15 Although finding a “correct” answer for many of our cases might be difficult, the available literature provides guidance on many of the scenarios. Barrett et al16 showed that predicting which patient will do poorly on dialysis at the time of the initiation of dialysis therapy is difficult. These findings argue for a more liberal approach toward accepting marginal candidates and are more consistent with current recommendations.3 Patients with renal failure and systolic cardiac dysfunction can develop significant improvement of their symptoms and have reasonable survival rates on dialysis therapy.17 Dialysis of patients with ESRD caused by myeloma kidney has been shown to be beneficial.18 Volume overload could be a contributing factor to the dyspnea of patients with chronic obstructive pulmonary disease and ESRD, and controlling the extracellular fluid volume–related component with ultrafiltration may relieve the shortness of breath. Conversely, dialyzing combative patients is technically difficult and rather dangerous for patients and staff and is thus considered a reason to forgo dialysis. Patients who cannot communicate with others for reasons other than uremia are also recommended not to undergo dialysis.19 Like many studies of similar design, our investigation has several limitations. First, we cannot exclude the possibility that there may be a bias such that those physicians who responded to the survey might differ from those who did not. The lowest response rate belonged to the US internists, the only group chosen through a commercial vendor. This list may not have been current and might have included subspecialists more likely to discard the survey. Second, the reported attitudes and experiences of participating physicians might be different from their actual practices. This concern is most pertinent to our data regarding nonreferral rates, which are based on physician’s memory, raising potentials for recall bias. Third, the sample size did not allow for an extensive subgroup analysis and possibly identification of risk factors associated with nonrefer-
NONREFERRAL TO DIALYSIS
ral. Despite these limitations, we believe our findings regarding reported nonreferral by PCPs of patients who might benefit from dialysis and the suggestion that differences in the incidence of ESRD between the United States and Canada are not caused by different physicians’ attitudes toward withholding dialysis remain valid. Largescale cohort and cross-sectional studies that attempt to answer these questions are needed to confirm our findings and shed more light on these important issues. In the meantime, it is imperative to dedicate more effort toward educating PCPs and nephrologists about modern criteria for patient selection for dialysis, especially not using older age as a reason to withhold this therapy. Canadian guidelines for referral of patients with elevated creatinine levels that address these issues have recently been published.20 Similarly, awareness is needed that when there is doubt about the outcome of renal replacement therapy, willing patients should be given a trial of dialysis therapy to assess the benefits and burdens, with reassessment of the question of whether to continue ongoing chronic dialysis therapy a month or 2 after initiation. Last, it is important to emphasize that the best approach to dealing with these difficult cases is to create an atmosphere of cooperation between patients and their families, as well as among the care team, including the PCP, nephrologist, dialysis nurses, social workers, and dietitians. REFERENCES 1. Port FK: Worldwide demographics and future trends in end-stage renal disease. Kidney Int 41:S4-S7, 1993 (suppl) 2. Friedman EA: End-stage renal disease therapy: An American success story. JAMA 14:1118-1122, 1996 3. Renal Physicians Association and American Society of Nephrology: Clinical Practice Guidelines on Shared Decision-Making in the Appropriate Initiation and Withdrawal from Dialysis. Rockville, MD, Renal Physicians Association, 2000, pp 30-31 4. Moulton LH, Port FK, Wolfe RA, Foxman B, Guire KE: Patterns of low incidence of treated end-stage renal disease among the elderly. Am J Kidney Dis 1:55-62, 1992 5. Mendelssohn DC, Kua BT, Singer PA: Referral for dialysis in Ontario. Arch Intern Med 22:2473-2478, 1995 6. Sekkarie MA, Moss AH: Withholding and withdrawing dialysis: The role of physician specialty and education
41
and patient functional status. Am J Kidney Dis 3:464-472, 1998 7. US Renal Data System: USRDS 1999 Annual Data Report: International Comparison of ESRD therapy. The National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD, 1999 (CD ROM) 8. McKenzie JK, Moss AH, Feest TG, Stocking CB, Siegler M: Dialysis decision making in Canada, the United Kingdom, and the United States. Am J Kidney Dis 1:12-18, 1998 9. Hirsch DJ, West ML, Cohen AD, Jindal KK: Experience with not offering dialysis to patients with a poor prognosis. Am J Kidney Dis 3:463-466, 1994 10. Healthy People 2010 Objectives. Chronic Kidney Disease, Draft Chapter. Available at: www.niddk.nih.gov/ federal/kuhdic/kidsub/2010.htm. Accessed: September 1, 2000. 11. Young EW: An improved understanding of the causes of end-stage renal disease. Semin Nephrol 3:170-175, 1997 12. Hamel MB, Teno JM, Goldman L, Lynn J, Davis RB, Galanos AN, Desbiens N, Connors AF Jr, Wenger N, Phillips RS: Patient age and decisions to withhold life-sustaining treatments from seriously ill, hospitalized adults. SUPPORT Investigators. Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatment. Ann Intern Med 2:116-125, 1999 13. Cassel CK, Moss AH, Rettig RA, Levinsky NG: Ethical issues, in Rettig RA, Levinsky NG (eds): Kidney Failure and the Federal Government. Washington, DC, National Academy Press, 1991, pp 53-56 14. Moss AH: Dialysis decisions and the elderly. Clin Geriatr Med 3:463-473, 1994 15. Bradley JG, Zia MJ, Hamilton N: Patient preferences for control in medical decision making: A scenario-based approach. Fam Med 7:496-501, 1996 16. Barrett BJ, Parfrey PS, Morgan J, Barre P, Fine A, Goldstein MB, Handa SP, Jindal KK, Kjellstrand CM, Levin A, Mandin H, Muirhead N, Richardson RM: Prediction of early death in end-stage renal disease patients starting dialysis. Am J Kidney Dis 2:214-222, 1997 17. Hebert MJ, Falardeau M, Pichette V, Houde M, Nolin L, Cardinal J, Ouimet D: Continuous ambulatory peritoneal dialysis for patients with severe left ventricular systolic dysfunction and end-stage renal disease. Am J Kidney Dis 5:761-768, 1995 18. Clark AD, Shetty A, Soutar R: Renal failure and multiple myeloma: Pathogenesis and treatment of renal failure and management of underlying myeloma. Blood Rev 2:79-90, 1999 19. Moss AH: To use dialysis appropriately: The emerging consensus on patient selection guidelines. Adv Ren Replace Ther 2:175-183, 1995 20. Mendelssohn DC, Barrett BJ, Brownscombe LM, Ethier J, Greenberg DE, Kanani SD, Levin A, Toffelmire EB: Elevated levels of serum creatinine: Recommendations for management and referral. CMAJ 161:413-417, 1999
I
N ALL PARTS OF the world, incidence rates of end-stage renal disease (ESRD) are calculated based on the number of patients started on renal replacement therapy. These calculations do not include those individuals in whom dialysis therapy was withheld (ie, never started). Different withholding patterns may explain, at least in part, the large variation in acceptance rates reported in different parts of the world1 and may also be a factor contributing to varying mortality rates of dialysis patients among countries.2 Studying withholding of dialysis is important from an ethical perspective. Not offering renal replacement therapy to patients who might benefit from it because of their age, a misconception that they are “not candidates for dialysis,” or From the Department of Medicine, Division of Nephrology, West Virginia University, Morgantown, WV; the Department of Medicine, Bluefield Regional Medical Center, Bluefield, WV; and the Department of Medicine, Division of Nephrology, University of Toronto, Canada. Received September 25, 2000; accepted in revised form January 12, 2001. Presented in part as an abstract at the 32nd Annual Meeting of the American Society of Nephrology, Miami Beach, FL, 1999. Address reprint requests to Mohamed Sekkarie, MD, 510 Cherry St, Bluefield, WV 24701. E-mail: [email protected] © 2001 by the National Kidney Foundation, Inc. 0272-6386/01/3801-0006$35.00/0 doi:10.1053/ajkd.2001.25179 36
other factors will lead to their premature death. The recently published Renal Physicians Association and American Society of Nephrology clinical practice guidelines on shared decision making in the appropriate initiation and withdrawal from dialysis3 lists very few conditions in which not offering dialysis to a patient (or their proxy decision maker) who wants it is believed to be appropriate. These conditions include patients who have irreversible profound neurological impairment, patients in whom dialysis is not feasible technically (inability to establish a dialysis access, inability to cooperate with the dialysis procedure), and patients with estimated survival less than 6 months. Even in the last group, it is recommended that the renal team should be sensitive to patients’ goals and individual circumstances, such as wanting to live to attend a special family event.3 Recent research from Canada and the United States suggests that not offering dialysis to patients who might benefit still occurs, and that this decision is often made at the level of primary care physicians (PCPs) without the involvement of a nephrologist.4-6 This literature from the early and mid-1990s suggests that contrary to ethical principles, old age seemed to be a reason for withholding dialysis in some cases. The difference in the incidence of treated ESRD between these two countries is large and
American Journal of Kidney Diseases, Vol 38, No 1 (July), 2001: pp 36-41
NONREFERRAL TO DIALYSIS
difficult to explain, even after taking into account the greater proportion of blacks in the US population. In 1997, the incidence rate in the United States was 296 per million population (229 per million for whites), whereas it was only 116 per million in Canada.7 This study was conducted to investigate nonreferral and nonacceptance to dialysis by PCPs and nephrologists in both Canada and the United States. We tried to answer the following questions: (1) Is dialysis being withheld from patients who might benefit from it? and (2) Do differences in withholding dialysis between Canada and the United States explain the large difference in acceptance rates to dialysis programs in these countries. METHODS We surveyed a random sample of Canadian and US PCPs and nephrologists concerning their attitudes toward and experience with withholding dialysis in patients with advanced chronic renal failure (CRF). Names of physicians were obtained from the College of Family Practice of Canada and the American Academy of Family Practice for family physicians, the American Society of Nephrology and the Canadian Society of Nephrology for nephrologists, and a commercial vendor that contracts with the American Medical Association for the names of US primary care internists. Random PCP lists were generated by computers of the providing agencies. All Canadian nephrologists were surveyed, and US nephrologists were chosen from the membership directory of the American Society of Nephrology using a random number table. Internists in Canada were not surveyed because they do not function as PCPs. Questionnaires were sent to 520 Canadian family physicians, 500 US PCPs (252 family physicians and 248 primary care internists), 299 Canadian nephrologists, and 250 US nephrologists. The surveys, which were mailed in 1999, asked questions about physicians’ demographics, practice type, and whether they believe that there should be an age cutoff beyond which dialysis should not be offered. In addition, physicians were given brief descriptions of 10 potential patients with ESRD and asked if they would recommend dialysis (Table 1). Nephrologists were asked to estimate how many patients with advanced CRF they personally treated who died of renal failure without initiation of dialysis therapy and how many they started on dialysis therapy in 1998. PCPs were asked how many patients with advanced CRF they personally treated who died of renal failure without nephrology referral and how many of their patients with ESRD were started on dialysis therapy during the previous 3 years. The survey instrument is available on request from the authors. Responses to the survey were kept anonymous and confidential. Forms were numbered to track nonresponders. Those who did not respond to the first survey were sent a second
37
mailing. Because of the lower response for US internists, a third mailing was sent to this group. Withholding rates for each group were calculated by dividing the number of patients withheld from dialysis therapy by the sum of the patients withheld from therapy and patients started on dialysis therapy in that group. Responses of the groups were compared by Fisher’s exact tests, and P less than 0.05 is considered significant. Data were analyzed using the statistical software SPSS (SSPS Inc, Chicago, IL). Some data were analyzed using Stata (version 6; Stata Corp, College Station, TX) and Excel (Microsoft Corp, Redmond, CA). Unless otherwise indicated, data presented in the result section pertain to responders.
RESULTS
After excluding such ineligible responses as those from physicians who retired or changed specialty or questionnaires that were returned because of the wrong address, the overall response rate was 49%. Canadian physicians had a better response rate than their US counterparts. Response rates and demographics of the respondents are listed in Table 2. Responses of US family physicians and primary care internists were similar; therefore, their data are presented together as US PCPs. Of the respondents, approximately 60% of Canadian and US PCPs and US nephrologists were in community-based group practices, and 75% of Canadian nephrologists were in academic medicine. Approximately 95% of PCPs indicated that they were board certified. Most PCPs had a nephrologist practicing within a 25-km radius of their practices. In response to the question about whether the physician believes there should be an age beyond which dialysis should not be offered, 12% of Canadian family practitioners, 20% of US PCPs, 4% of Canadian nephrologists, and 9% of US nephrologists answered yes. The differences were statistically significant for Canadian versus US PCPs (P ⫽ 0.03), nephrologists versus PCPs in both countries (P ⫽ 0.005 for Canada, P ⫽ 0.02 for the United States), but not Canadian versus US nephrologists. Those who answered yes suggested cutoff values that ranged between the ages of 65 to 100 years (median, 80 years); practice type and proximity to a nephrologist were not associated with the physicians’ position on this issue. The proportions of physicians recommending dialysis for each of the 10 vignettes are listed in Table 1 as percentages of physicians who recom-
38
SEKKARIE, COSMA, AND MENDELSSOHN Table 1.
The Vignettes and Proportion of Physicians Recommending Dialysis
Case
Cases
a
An 80-year-old patient with diabetes with an ejection fraction of 25% caused by ischemic cardiomyopathy, lives with a sitter and takes care of herself, normal mental status. A 73-year-old patient with metastatic prostate cancer, no pain, unable to walk because of cord compression, lives with family, normal mental status. A 90-year-old patient with hypertensive nephrosclerosis, lives alone, takes care of herself, seen by a visiting nurse weekly to supervise medication intake. A 45-year-old patient with persistent vegetative state for the last 3 months after a cardiac arrest, prognosis for neurological recovery is poor. A 35-year-old schizophrenic patient, very combative despite medications, not a competent decision maker, institutionalized. A 62-year-old patient with multi-infarct dementia, incontinent, feeds self, bed and chair confined, does not recognize family. A 70-year-old patient on chemotherapy for multiple myeloma, ambulatory, normal mental status. A 29-year-old paraplegic, permanently ventilator dependent, alert and coherent. A 68-year-old patient with advanced chronic obstructive, pulmonary disease caused by smoking, chronically on oxygen, becomes dyspneic after walking ⱖ5 feet. A 30-year-old blind patient with diabetes.
b
c
d
e f g h i
j
Canadian PCP (%)
US PCP (%)
Canadian NEP (%)
US NEP (%)
74
70
81
85
50
54
59
60
76
71
80
88
5
3
1
1
59
54
32
36
9
12
1
5
88
91
98
100
90
87
79
78
63
62
61
73
99
99
99
100
NOTE. Cases are of patients potentially close to needing dialysis. Physicians were instructed to assume that the patient and/or proxy decision maker is ambivalent about whether to accept chronic dialysis therapy and requests the physician’s recommendation. Based on the presented data, physicians were asked to indicate whether they would recommend dialysis. The percentage of physicians who recommended dialysis is shown. Canadian PCP versus US PCP, P ⬍ 0.05 for case e only; Canadian NEP versus US NEP, P ⬍ 0.05 for case f only; Canadian PCP versus Canadian NEP, P ⬍ 0.05 for cases d, e, g, and h; and US PCP versus US NEP, P ⬍ 0.05 for cases a, c, and g. Abbreviation: NEP, nephrologist.
mended dialysis therapy. When the responses of each country’s physicians were combined, in none of the cases did the difference between Canadian and US physicians reach statistical Table 2.
Demographics and Response Rates of Physicians
Eligible Responded Response rate (%) Men (%) Mean age (y)
Canadian PCP
US PCP
Canadian NEP
US NEP
490 263 54 55 41
452 176 38 77 46
258 166 64 80 48
231 93 40 88 47
NOTE. Differences in response rates: Canadian PCP versus US PCP, P ⫽ 0.001; Canadian PCP versus Canadian NEP, P ⫽ 0.005; US PCP versus US NEP, P ⫽ not significant; Canadian NEP versus US NEP, P ⫽ 0.001. Abbreviation: NEP, nephrologist.
significance. Conversely, when results of all PCPs were compared with those of all nephrologists, PCPs were less likely to recommend dialysis in cases a, b, c, and g and more likely to recommend dialysis in cases d, e, f, and h. Responses for cases i and j did not differ. Over a 3-year period, 13% of Canadian PCPs and 19% of US PCPs reported nonreferral to dialysis at least once. Withholding rates were 25% for Canadian family practitioners, 15% for US PCPs, 13% for Canadian nephrologists, and 16% for US nephrologists. The differences between Canadian PCPs versus US PCPs, Canadian nephrologists versus US nephrologists, and Canadian PCPs versus Canadian nephrologists were statistically significant. Table 3 lists the median number of patients referred and not referred for dialysis therapy per PCP per 3 years. Table 4 lists the median number
NONREFERRAL TO DIALYSIS Table 3.
39
Median Number of Patients Referred and Not Referred Per PCP Per 3 Years
Referred Not referred
Canadian PCP
US PCP
1 0
2 0
of patients started on and withheld from dialysis therapy per nephrologist per year. PCPs who withheld dialysis listed on the average three reasons per nonreferred patient (n ⫽ 67). Patient refusal, the most common cause, was cited as one of the reasons for nonreferral in 60% of the cases, followed by end-stage heart, lung, or liver disease (48%); old age (46%); overall frail condition (39%); and then family refusal, dementia, and terminal cancer (⬃31% each). This distribution of causes did not differ between the two countries. Old age was cited as the sole cause only once. DISCUSSION
To the best of our knowledge, this study is the first that compares the attitudes and experiences of Canadian and US physicians, both PCPs and nephrologists, toward the issue of withholding dialysis therapy. Contrary to a widely held belief,1 our findings do not support the theory that the difference in incidence rates of ESRD in these countries is primarily caused by more liberal attitudes of physicians concerning acceptance criteria in the United States. Canadian physicians were less likely to consider age a reason to withhold dialysis, and their attitudes toward recommending dialysis in the hypothetical cases did not differ from those of US physicians. PCPs in both countries appear to understand modern acceptance criteria reasonably well. Patient refusal was the most commonly cited cause for nonreferral to dialysis by PCPs. It is likely, at least in some cases, that these nonreferred patients were not adequately informed about the dialysis procedure and its potential benefits. The design of our study did not allow us to test this hypothesis. There is some suggestion in the literature that Canadian nephrologists might be less liberal in accepting patients to dialysis than US nephrologists. In a 1998 survey of Canadian, British, and US nephrologists, McKenzie et al,8 using case
vignettes, found that American nephrologists were slightly more likely to offer dialysis than their Canadian and British counterparts. Hirsch et al9 from Nova Scotia, Canada, reported that in 1992, one quarter of their referred patients with ESRD were not offered dialysis, which contrasts with a 7% withholding rate by nephrologists of West Virginia in 1995.6 Our study showed a less liberal attitude by Canadian nephrologists only in the case of the patient with multi-infarct dementia. Although random error and different sensitivities of the research instruments may have had a role in these small discrepancies, we believe that because our survey is more recent and is a national rather than a provincial experience, it is more reflective of the current attitudes of Canadian nephrologists. If the attitudes of PCPs and nephrologists are not different in Canada and the United States, then what might explain the large difference in incidence rates between the two countries? It is theoretically possible that more patients with ESRD die without being diagnosed in Canada, but this seems unlikely. Under universal health insurance, access to PCPs should be easier in Canada than in the United States. Other potential explanations for the greater acceptance rates in the United States need to be considered, including a genuinely greater incidence of chronic renal insufficiency in the United States (it is estimated from the Third Nutrition and Health Examination Survey and the 1990 US census that 0.4% of the total US population aged ⱖ12 years have a serum creatinine value ⬎ 2.0 mg/ dL10; we are not aware of similar data from Canada), lower mortality from competing diseases in the pre-ESRD period in the United States,11 slower rate of progression of kidney disease in Canada secondary to easier access to predialysis care, and/or later initiation of dialysis in Canada. Finally, access to a nephrologist in Table 4.
Started Withheld
Median Number of Patients Started on and Withheld From Dialysis Therapy Per Nephrologist Per Year Canadian NEP
US NEP
15 2
20 3
Abbreviation: NEP, nephrologist.
40
Canada is only through referral from another physician. Conversely, in the United States, direct access through self-referral to a nephrologist is possible and may allow for easier access to renal replacement therapy. Data to support or rule out these potential explanations are lacking. In both the United States and Canada, concerns have been raised about the issue of depriving certain patients, especially the elderly, from potentially beneficial dialysis therapy.5,6,12 In a survey of PCPs in Ontario in 1994, 14.2% of these physicians reported withholding dialysis on their own at least once over a period of 3 years. In the same study, these physicians were less likely to refer older patients.5 Similar findings were found in a survey of a group of PCPs in West Virginia in 1995 in which 60% of the surveyed PCPs cited age as a major factor in dialysis decision making.6 Our study, performed at a later date, also shows that this phenomenon continues to exist; 12% of Canadian and 20% of US PCPs believe there should be an age cutoff beyond which dialysis should not be offered. Although these percentages are smaller for nephrologists, they are still a cause for concern. The Institute of Medicine Committee for the Study of Medicare End-Stage Renal Disease Program explicitly rejected age as a criterion for patient acceptance,13 and many elderly patients have a good quality of life and reasonable life expectancy on dialysis.14 Educating physicians about this issue should be a priority. Consistent with our findings on the age cutoff issue, the responses of US and Canadian physicians to the hypothetical cases were also similar. However, small but statistically significant differences between PCPs’ and nephrologists’ recommendations were seen. In the cases of patients with ischemic cardiomyopathy (case a), prostate cancer (case b), nephrosclerosis (case c), and myeloma (case g), PCPs were less likely to recommend dialysis. The opposite was true for cases with neurological or psychiatric impairment (cases d, e, f, and h) in which the patients were incidentally younger. The opinions of the nephrologists on many of these cases also differ. Our questions regarding the hypothetical cases assumed that the patient and/or proxy decision maker was ambivalent about accepting renal replacement therapy and were thus designed to
SEKKARIE, COSMA, AND MENDELSSOHN
study the attitudes of physicians toward whether to recommend dialysis without the influence of patient autonomy. This is important because patients and proxy decision makers are generally strongly influenced by the recommendations of their physicians.15 Although finding a “correct” answer for many of our cases might be difficult, the available literature provides guidance on many of the scenarios. Barrett et al16 showed that predicting which patient will do poorly on dialysis at the time of the initiation of dialysis therapy is difficult. These findings argue for a more liberal approach toward accepting marginal candidates and are more consistent with current recommendations.3 Patients with renal failure and systolic cardiac dysfunction can develop significant improvement of their symptoms and have reasonable survival rates on dialysis therapy.17 Dialysis of patients with ESRD caused by myeloma kidney has been shown to be beneficial.18 Volume overload could be a contributing factor to the dyspnea of patients with chronic obstructive pulmonary disease and ESRD, and controlling the extracellular fluid volume–related component with ultrafiltration may relieve the shortness of breath. Conversely, dialyzing combative patients is technically difficult and rather dangerous for patients and staff and is thus considered a reason to forgo dialysis. Patients who cannot communicate with others for reasons other than uremia are also recommended not to undergo dialysis.19 Like many studies of similar design, our investigation has several limitations. First, we cannot exclude the possibility that there may be a bias such that those physicians who responded to the survey might differ from those who did not. The lowest response rate belonged to the US internists, the only group chosen through a commercial vendor. This list may not have been current and might have included subspecialists more likely to discard the survey. Second, the reported attitudes and experiences of participating physicians might be different from their actual practices. This concern is most pertinent to our data regarding nonreferral rates, which are based on physician’s memory, raising potentials for recall bias. Third, the sample size did not allow for an extensive subgroup analysis and possibly identification of risk factors associated with nonrefer-
NONREFERRAL TO DIALYSIS
ral. Despite these limitations, we believe our findings regarding reported nonreferral by PCPs of patients who might benefit from dialysis and the suggestion that differences in the incidence of ESRD between the United States and Canada are not caused by different physicians’ attitudes toward withholding dialysis remain valid. Largescale cohort and cross-sectional studies that attempt to answer these questions are needed to confirm our findings and shed more light on these important issues. In the meantime, it is imperative to dedicate more effort toward educating PCPs and nephrologists about modern criteria for patient selection for dialysis, especially not using older age as a reason to withhold this therapy. Canadian guidelines for referral of patients with elevated creatinine levels that address these issues have recently been published.20 Similarly, awareness is needed that when there is doubt about the outcome of renal replacement therapy, willing patients should be given a trial of dialysis therapy to assess the benefits and burdens, with reassessment of the question of whether to continue ongoing chronic dialysis therapy a month or 2 after initiation. Last, it is important to emphasize that the best approach to dealing with these difficult cases is to create an atmosphere of cooperation between patients and their families, as well as among the care team, including the PCP, nephrologist, dialysis nurses, social workers, and dietitians. REFERENCES 1. Port FK: Worldwide demographics and future trends in end-stage renal disease. Kidney Int 41:S4-S7, 1993 (suppl) 2. Friedman EA: End-stage renal disease therapy: An American success story. JAMA 14:1118-1122, 1996 3. Renal Physicians Association and American Society of Nephrology: Clinical Practice Guidelines on Shared Decision-Making in the Appropriate Initiation and Withdrawal from Dialysis. Rockville, MD, Renal Physicians Association, 2000, pp 30-31 4. Moulton LH, Port FK, Wolfe RA, Foxman B, Guire KE: Patterns of low incidence of treated end-stage renal disease among the elderly. Am J Kidney Dis 1:55-62, 1992 5. Mendelssohn DC, Kua BT, Singer PA: Referral for dialysis in Ontario. Arch Intern Med 22:2473-2478, 1995 6. Sekkarie MA, Moss AH: Withholding and withdrawing dialysis: The role of physician specialty and education
41
and patient functional status. Am J Kidney Dis 3:464-472, 1998 7. US Renal Data System: USRDS 1999 Annual Data Report: International Comparison of ESRD therapy. The National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD, 1999 (CD ROM) 8. McKenzie JK, Moss AH, Feest TG, Stocking CB, Siegler M: Dialysis decision making in Canada, the United Kingdom, and the United States. Am J Kidney Dis 1:12-18, 1998 9. Hirsch DJ, West ML, Cohen AD, Jindal KK: Experience with not offering dialysis to patients with a poor prognosis. Am J Kidney Dis 3:463-466, 1994 10. Healthy People 2010 Objectives. Chronic Kidney Disease, Draft Chapter. Available at: www.niddk.nih.gov/ federal/kuhdic/kidsub/2010.htm. Accessed: September 1, 2000. 11. Young EW: An improved understanding of the causes of end-stage renal disease. Semin Nephrol 3:170-175, 1997 12. Hamel MB, Teno JM, Goldman L, Lynn J, Davis RB, Galanos AN, Desbiens N, Connors AF Jr, Wenger N, Phillips RS: Patient age and decisions to withhold life-sustaining treatments from seriously ill, hospitalized adults. SUPPORT Investigators. Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatment. Ann Intern Med 2:116-125, 1999 13. Cassel CK, Moss AH, Rettig RA, Levinsky NG: Ethical issues, in Rettig RA, Levinsky NG (eds): Kidney Failure and the Federal Government. Washington, DC, National Academy Press, 1991, pp 53-56 14. Moss AH: Dialysis decisions and the elderly. Clin Geriatr Med 3:463-473, 1994 15. Bradley JG, Zia MJ, Hamilton N: Patient preferences for control in medical decision making: A scenario-based approach. Fam Med 7:496-501, 1996 16. Barrett BJ, Parfrey PS, Morgan J, Barre P, Fine A, Goldstein MB, Handa SP, Jindal KK, Kjellstrand CM, Levin A, Mandin H, Muirhead N, Richardson RM: Prediction of early death in end-stage renal disease patients starting dialysis. Am J Kidney Dis 2:214-222, 1997 17. Hebert MJ, Falardeau M, Pichette V, Houde M, Nolin L, Cardinal J, Ouimet D: Continuous ambulatory peritoneal dialysis for patients with severe left ventricular systolic dysfunction and end-stage renal disease. Am J Kidney Dis 5:761-768, 1995 18. Clark AD, Shetty A, Soutar R: Renal failure and multiple myeloma: Pathogenesis and treatment of renal failure and management of underlying myeloma. Blood Rev 2:79-90, 1999 19. Moss AH: To use dialysis appropriately: The emerging consensus on patient selection guidelines. Adv Ren Replace Ther 2:175-183, 1995 20. Mendelssohn DC, Barrett BJ, Brownscombe LM, Ethier J, Greenberg DE, Kanani SD, Levin A, Toffelmire EB: Elevated levels of serum creatinine: Recommendations for management and referral. CMAJ 161:413-417, 1999