Nephrologist Care and Mortality in Patients With Chronic ...
Upcoming SlideShare
Loading in...5

Nephrologist Care and Mortality in Patients With Chronic ...






Total Views
Views on SlideShare
Embed Views



0 Embeds 0

No embeds



Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
Post Comment
Edit your comment

Nephrologist Care and Mortality in Patients With Chronic ... Nephrologist Care and Mortality in Patients With Chronic ... Document Transcript

  • ORIGINAL INVESTIGATION Nephrologist Care and Mortality in Patients With Chronic Renal Insufficiency Jerry Avorn, MD; Rhonda L. Bohn, ScD; Elliott Levy, MD, MPH; Raisa Levin, MS; William F. Owen, Jr, MD; Wolfgang C. Winkelmayer, MD, ScD, MPH; Robert J. Glynn, ScD Background: For patients with chronic renal insuffi- Results: From multivariate analyses, we found that pa- ciency, rates of referral to nephrologists are highly vari- tients who did not see a nephrologist until 90 days or less able, and little is known about the effect of such consul- before initiation of dialysis had a 37% higher likelihood tation on clinical outcomes. We sought to determine of death in the first year of dialysis compared with pa- whether early or frequent access to nephrologist care tients with earlier referral (95% confidence interval, 1.22- prior to the initiation of dialysis was associated with a 1.52; P .001). Similarly, those who saw a nephrologist difference in mortality rates in the first year after dialysis on fewer than 5 occasions in the year prior to dialysis had began. a 15% higher mortality rate in the first year of dialysis com- pared with those who had had 5 or more nephrologist Methods: We identified all patients in the New Jersey visits (95% confidence interval, 1.03-1.28; P=.01). Medicaid and Medicare programs who began mainte- nance dialysis during a 6-year period and who had been Conclusions: For patients with long-standing renal dis- diagnosed with renal disease more than 12 months prior ease, earlier consultation with a nephrologist and more fre- to dialysis. Use of nephrologist services was docu- quent specialist encounters is associated with lower mor- mented during this 1-year period, along with other clini- tality in the first year of dialysis. These findings need to cal and sociodemographic variables. The outcome mea- be confirmed in younger and less indigent patients as well. sure of our analysis was mortality in the first year after initiation of dialysis. Arch Intern Med. 2002;162:2002-2006 W ITH INCREASING penditures by the Medicare program, re- pressures on all as- gardless of the age of the patient. In 1998, pects of health care the total cost of care for the numerically expenditures, and small ESRD program was approximately growing concern $12 billion, rendering it the most costly about the cost-effectiveness of medical in- single Medicare program.1 Since the late terventions, attention has turned to the role 1960s, the care of ESRD patients has been of the specialist in various domains of pa- a lightning rod for societal questions about From the Division tient care. Many payers and health care sys- health care access, equity, and cost. of Pharmacoepidemiology tems have placed renewed emphasis on the While dialysis care of ESRD patients and Pharmacoeconomics role of the primary care physician as both is provided almost uniformly by nephrolo- (Drs Avorn, Bohn, Winkelmayer, and Glynn and coordinator of services and as “gate- gists, the predialysis management of Ms Levin), Division of keeper” in rationing access to costly in- chronic renal insufficiency is a domain in Nephrology (Drs Levy and terventions. When done optimally, this can which issues of specialist vs generalist care Owen), Department of reduce unnecessary expenditures and im- loom large. Little is known about the ap- Medicine, Brigham and prove the coordination of care. However, propriate place of nephrologist care for pa- Women’s Hospital, Harvard excessive application of such pressures can tients with chronic renal insufficiency not Medical School, Boston, Mass. result in denial of access to more costly ser- requiring renal replacement therapy, a Dr Levy is now with vices, even when these are clinically ap- phase that often lasts for several years. Al- Bristol-Myers Squibb propriate and may be cost-effective. though extensive data are available through Pharmaceutical Research Chronic renal disease is one arena in the Medicare ESRD program once pa- Institute, Princeton, NJ, and Dr Owen is with Duke Institute which these issues have been debated in re- tients receive chronic renal replacement of Renal Outcomes Research cent years. End-stage renal disease (ESRD) therapy (ie, maintenance hemodialysis, and Health Policy, Duke is the only clinical condition for which the peritoneal dialysis, or renal transplanta- University Medical Center, development of a given diagnosis automati- tion), it is much more difficult to as- Durham, NC. cally results in coverage of health care ex- semble data during the period prior to their (REPRINTED) ARCH INTERN MED/ VOL 162, SEP 23, 2002 WWW.ARCHINTERNMED.COM 2002 Downloaded from on October 20, 2010 ©2002 American Medical Association. All rights reserved.
  • entry into this program. However, this is precisely the grams, patients were required to have had at least 1 medical en- period during which it is most critical to study the effect counter of any kind in the year prior to their first maintenance of care on the subsequent clinical course of such pa- dialysis procedure. In addition, we required that the first diag- tients. nosis of renal disease have occurred more than 1 year prior to the initiation of dialysis, to exclude patients with new-onset re- Some have argued that early involvement of a ne- nal insufficiency who may not have had the opportunity for many phrologist in such care can result in better clinical status encounters with a nephrologist prior to dialysis because of the at the time of initiation of dialysis, which in turn could time course of their condition. result in an improved clinical outcome once dialysis has Patients were excluded if they had no second dialysis fol- begun. Such preparation can include the timely and pro- lowing the index procedure but survived more than 1 month, active development of vascular access required for hemo- if they had more than 2 months between 2 consecutive dialy- dialysis, and optimization of hematologic, endocrine, nu- sis procedures, or if their health care providers could not be tritional, metabolic, and hemodynamic function in the face identified. All patient identifiers were transformed into ano- of progressive renal failure. In 1997, 55% of patients with nymized untraceable study numbers in all analyses to protect chronic renal insufficiency were first seen only within 12 confidentiality. For each patient, we then defined age, sex, and race on months prior to initiating dialysis, and 33% were first seen the date of initiation of maintenance dialysis. We also roughly within 3 months of initiating dialysis.2,3 Several previous characterized socioeconomic status at this point by defining studies have demonstrated substantial heterogeneity in the whether the patient had been enrolled in Medicaid or the New management of patients with chronic renal insufficiency, Jersey Pharmaceutical Assistance to the Aged and Disabled perhaps related to the variability of timing and intensity (PAAD) program, a state-specific program that reimburses drug of the interaction with nephrologists. The annual mortal- expenditures for state residents of modest income who are not ity rate for ESRD patients is approximately 20% per year, indigent enough to qualify for Medicaid. We then defined the and half these deaths are attributed to cardiovascular com- frequency and timing of visits with a nephrologist in the 12 plications.4 This mortality translates into a life expec- months prior to initiation of maintenance dialysis. Physician tancy that is only 16% to 37% that of the age-, sex-, and specialty was identified by Medicare and Medicaid specialty codes as well as by Unique Physician Identification Numbers as- race-matched general population.2 The highest death risk signed to all practicing physicians. For each physician encoun- occurs during the incident year of renal replacement ter, provider numbers were searched for the specialty code for therapy, and is subsequently lower.1,5 nephrologist. The effect of delayed referral to a nephrologist on mor- We also extracted information on all hospitalizations, phy- bidity and mortality has been examined in a limited man- sician visits, procedures, and nursing home care received by ner.6 In a small study,7 patients receiving care from a ne- these patients during this period. This made it possible to iden- phrologist had shorter hospitalizations to initiate tify all diagnoses assigned to these patients by all clinicians who hemodialysis than did patients cared for by nonnephrolo- cared for them, including specific renal diagnoses as well as co- gists, or those who received no medical care at all. Simi- morbidities such as hypertension, diabetes mellitus, conges- larly, Hakim et al8 reported that hospital stays were longer tive heart failure, ischemic heart disease, and other relevant con- ditions (Table 1). Death rates were measured in the year after and more costly for patients with chronic renal insuffi- maintenance dialysis was begun, using data from Medicare and ciency referred to a nephrologist relatively late in the course Medicaid enrollment files. of their disease. Other studies have examined the associa- Referral patterns to a nephrologist were studied using 2 tion of late nephrologist referral and mortality in new ESRD definitions. Late referral was defined as a patient having a first patients, but have produced conflicting results.9-12 These nephrologist encounter 90 days or less before the initiation of studies used different definitions of late vs early referral, maintenance dialysis. A second definition, that of frequency of and were limited to small and highly selected samples. nephrologist care, was defined as the number of encounters with The present study is the most comprehensive to date a nephrologist during the 12 months prior to initiation of main- to examine this question in a very large population of pa- tenance dialysis. This variable was dichotomized: 5 or more vis- tients cared for in typical settings. It was designed to ex- its vs less than 5 visits. We developed a proportional hazards regression model to amine the association between the utilization of nephrolo- predict the likelihood of death in the 12 months following ini- gist services by patients with chronic renal insufficiency tiation of dialysis. Age, sex, race, PAAD/Medicaid member- during the year before the initiation of chronic dialysis, ship as proxy for socioeconomic status, and underlying renal and mortality during the first year of renal replacement diagnosis (categorical) were forced into the model. Then we therapy. added the number of predialysis nephrologist visits and the vari- able for late referral, respectively, into 2 separate models. Fi- PATIENTS AND METHODS nally, we tested all other covariates (categories of comorbid- ity, Charlson comorbidity score, physician characteristics) for We identified all patients who began maintenance dialysis be- individual significance or confounding. We also tested for tween 1991 and mid-1996, and had been active participants in significance of potential effect modifiers, most importantly the either the Medicare or Medicaid programs of the state of New interaction between age and timing of referral.13 Jersey for at least the prior 12 months. All personal identifiers were transformed into coded study numbers to protect confi- dentiality. Maintenance dialysis was identified by the Interna- RESULTS tional Classification of Diseases, Ninth Revision (ICD-9) and Cur- rent Procedural Terminology codes for hemodialysis, peritoneal We identified 17 884 patients who underwent mainte- dialysis, renal transplantation, and ESRD. The first record of main- nance dialysis at some point during the period 1991 to tenance dialysis during this period was referred to as the index 1996; of these patients, 12557 had adequate baseline data claim. To ensure eligibility in the Medicare or Medicaid pro- for a full year prior to dialysis in Medicaid and/or Medi- (REPRINTED) ARCH INTERN MED/ VOL 162, SEP 23, 2002 WWW.ARCHINTERNMED.COM 2003 Downloaded from on October 20, 2010 ©2002 American Medical Association. All rights reserved.
  • As seen in Table 1, the study population included Table 1. Characteristics of Study Population large numbers of older patients, with 1288 (42.7%) be- tween the ages of 65 and 74 years and 1063 (35.3%) aged No. (%) of Patients* 75 to 84 years. There were slightly more men than women Early Late (56.2% vs 43.8%), and 25.7% were nonwhite. The most Referral Referral common renal diagnoses specified were hypertensive ne- Characteristic (n = 1975) (n = 1039) phropathy and diabetic nephropathy. Approximately one Age, y third of patients (34.5%) did not see a nephrologist un- 45 51 (2.6) 32 (3.1) til 90 days or less before their first renal replacement 45-54 79 (4.0) 58 (5.6) therapy. Half of the patients (50.5%) had fewer than 5 55-64 174 (8.8) 87 (8.4) nephrologist consultations in the year prior to renal re- 65-74 881 (44.6) 407 (39.2) 75-84 691 (35.0) 372 (35.8) placement therapy. 85 99 (5.0) 83 (8.0) As expected, age was a strong predictor of the like- Sex lihood of mortality, with the risk ratio for death (com- Female 890 (45.1) 430 (41.4) pared with a referent group aged 45 years) increasing Male 1085 (54.9) 609 (58.6) systematically from 2.97 for patients aged 45 to 54 years, Race to 11.53 for patients 85 years and older (Table 2). Af- White 1484 (75.1) 756 (72.8) Black 391 (19.8) 195 (18.8) ter controlling for clinical diagnoses (see below), we found Other 100 (4.1) 88 (8.5) no association between sex and mortality. However, blacks Entitlement program/socioeconomic were found to have a lower risk of mortality, as reported status in other studies of ESRD,14 with a risk ratio of 0.77 Non-PAAD,† non-Medicaid 1298 (65.7) 712 (68.5) (P .001). Individuals of race other than white or black PAAD or Medicaid 677 (34.3) 327 (31.5) were at even lower risk of death in their first year on re- Renal diagnoses‡ Chronic renal failure with acute onset 81 (4.1) 24 (2.3) nal replacement therapy (risk ratio, 0.65; P=.002). Being Polycystic kidney disease 105 (5.3) 52 (5.0) enrolled in either the New Jersey Medicaid or PAAD pro- Diabetic nephropathy 635 (32.2) 240 (23.1) grams, which served as a proxy for lower socioeco- Hypertensive nephropathy (malignant) 138 (7.0) 50 (4.8) nomic status, was not significantly predictive of higher Hypertensive nephropathy (benign) 1324 (67.0) 527 (50.7) mortality (risk ratio, 0.97; P =.57). Pyelonephritis 79 (4.0) 39 (3.75) After controlling for age, race, socioeconomic sta- Obstructive nephropathy 100 (5.1) 71 (6.8) tus, and the presence of all renal diagnoses recorded in the Renovascular disease 93 (4.7) 38 (3.7) Miscellaneous 966 (48.9) 389 (37.4) year before dialysis (Table 2), we examined the risk of death Unknown 35 (1.8) 51 (4.9) independently associated with late referral to a nephrolo- Comorbidities gist as defined above, using proportional hazards regres- Diabetes 1011 (51.2) 469 (45.1) sion analysis. Patients with late referral (first nephrolo- Hypertension 1638 (82.9) 741 (71.3) gist consultation 90 days before initiation of dialysis) had Coronary artery disease 1424 (72.1) 712 (68.5) a 37% increase in risk of death in the first year of dialysis Congestive heart failure 1265 (64.1) 654 (63.0) Malignancy 436 (22.1) 228 (21.9) compared with patients whose first nephrologist encoun- No. of nephrologist consultations ter occurred more than 90 days before the start of dialysis 5 1341 (67.9) 151 (14.5) (95% confidence interval, 22%-52%; P .001). 5 634 (32.1) 888 (85.5) We next replaced this variable in the model with a variable defining the frequency of nephrologist visits dur- *Early referral indicates first nephrologist visit 90 days before dialysis; ing the 12 months prior to initiation of dialysis. Here again, late referral, first nephrologist visit 90 days before dialysis. †PAAD indicates Pharmaceutical Assistance to the Aged and Disabled. after controlling for all other variables studied, patients ‡Not mutually exclusive. who had seen a nephrologist fewer than 5 times in the year prior to initiation of dialysis had a 15% higher risk of death in the subsequent year, compared with those who care to permit further study. In this population, 5242 pa- saw a nephrologist more frequently during that period tients had their first renal diagnosis at least 1 year prior (P =.01). Adjusting for the Charlson comorbidity score to the initiation of dialysis and were therefore eligible for and for other specific nonrenal clinical conditions (eg, further study. Most of the patients without a previous di- diabetes, ischemic heart disease) yielded essentially iden- agnosis of kidney disease who were eliminated in this step tical results (risk ratiolate referral, 1.35; P .001). In using can be assumed to have undergone one-time or short- interaction terms to study whether the effect of late re- term renal replacement therapy as a consequence of acute ferral differed by age group, we found that the effect was renal failure. Six hundred twenty-six patients in this group not significantly changed by the age of the patient. of 5242 patients had less than 30 days of renal replace- ment therapy, indicating that they had acute renal fail- COMMENT ure. Five hundred ninety-nine patients had more than 2 months without claims for renal replacement therapy and These findings are based on what is, to our knowledge, survived without additional dialysis care, and 1003 lacked the largest study to date of predialysis patients, and raise adequate data describing their health care providers. This important questions about access to and utilization of left a study population of 3014 patients. Of these, 1430 nephrologist services in this vulnerable population. The (47.4%) died in the first year of dialysis. excess mortality seen in patients with chronic renal (REPRINTED) ARCH INTERN MED/ VOL 162, SEP 23, 2002 WWW.ARCHINTERNMED.COM 2004 Downloaded from on October 20, 2010 ©2002 American Medical Association. All rights reserved.
  • insufficiency who have late or infrequent consultation with nephrologists does not appear to be an artifact Table 2. Proportional Hazards Regression Model caused by differences in patients’ ages, sex, cause of their of Risk Factors for Mortality in the First Year of Maintenance Dialysis* renal disease, socioeconomic status, or other comorbid conditions. Risk Ratio The present study may help reconcile the conflicting (95% Confidence P results among previous, smaller studies. While Ifudu et al9 Variable Interval) Value (United States), Innes et al10 (United Kingdom), and Sesso Age, y et al11 (Brazil) had found an increase in mortality in pa- 45 1.0 (Referent) ... tients with delayed nephrological care, Roubicek et al12 45-54 2.97 (1.44-6.14) .003 (France) did not find an increase in 1-year or 5-year mor- 55-64 3.75 (1.89-7.45) .001 tality in patients who first saw a nephrologist less than 4 65-74 4.97 (2.57-9.65) .001 75-84 6.57 (3.39-12.73) .001 months prior to onset of dialysis. However, differences in 85 11.53 (5.85-22.74) .001 methodology and populations under study probably con- Sex founded these previous results.15 The largest population in Female 1.0 (Referent) ... which this question was studied previously contained only Male 1.02 (0.92-1.14) .72 270 subjects,12 compared with the 3014 in the present re- Race port. The failure to find a difference in survival between White 1.0 (Referent) ... Black 0.77 (0.66-0.89) .001 patients referred to a nephrologist late vs early in previous Other 0.65 (0.49-0.86) .002 studies is probably attributable to selection bias, or prob- Entitlement program/socioeconomic lems in generalizability from one population to another, status particularly when cross-national health system issues are Non-PAAD,† non-Medicaid 1.0 (Referent) ... considered. Friedman15 has pointed out that the frail or PAAD or Medicaid 0.97 (0.86-1.09) .58 elderly patients most vulnerable to delayed referral may Renal diagnoses‡ Chronic renal failure nonspecified 1.0 (Referent) ... be less likely to receive renal replacement therapy in Chronic renal failure with acute onset 1.24 (0.95-1.62) .11 France, while such patients with high comorbid health Polycystic kidney disease 0.70 (0.55-0.91) .007 status and age, who may benefit most from timely ne- Diabetic nephropathy 1.02 (0.91-1.16) .70 phrologist consultation, are likelier to receive treatment Hypertensive nephropathy (malignant) 1.06 (0.86-1.31) .61 in the United States. Hypertensive nephropathy (benign) 0.84 (0.75-0.94) .003 Although extensive outcome data are available Pyelonephritis 0.95 (0.72-1.25) .72 Obstructive nephropathy 1.06 (0.85-1.31) .63 through Medicare and its associated ESRD registries, it Renovascular disease 1.08 (0.84-1.39) .55 is more difficult to assemble data describing patients prior Miscellaneous 0.98 (0.88-1.10) .77 to their renal replacement therapy. However, this pe- Unknown 0.91 (0.65-1.28) .60 riod is critical to study the effect of care on the subse- Timing of first nephrologist visit quent clinical course of such patients. In evaluating these 90 d before dialysis 1.0 (Referent) ... findings, some limitations of the study’s reliance on Medi- 90 d before dialysis 1.37 (1.22-1.52) .001 care and Medicaid data must be kept in mind. For pa- No. of nephrologist consultations 5 1.0 (Referent) ... tients younger than 65 years, the only source of predi- 5 1.15 (1.03-1.28) .01 alysis information was care received through the Medicaid program, raising concerns about whether data from this *The variables for timing and frequency of nephrologist consultations group of patients can be generalized to younger, nonin- were entered separately into the same regression model. The parameter digent patients with these conditions. However, for all estimates are from the model including timing of first nephrologist visit. When replacing this variable with number of nephrologist consultations, the patients who began dialysis after age 66, predialysis en- effect estimates and P values for all other variables in the model remained rollment in Medicare was virtually universal, and we found essentially unchanged. no difference in the effect of nephrologist referral in this †PAAD indicates Pharmaceutical Assistance to the Aged and Disabled. age group compared with younger patients (P=.21). Sec- ‡Not mutually exclusive. ond, because the diagnoses were based in part on data from inpatient and outpatient encounter claims from the ords, the attribution of ESRD to a particular clinical cause Medicare and Medicaid programs, there is a possibility is often difficult and often suspect, because most pa- of misclassification with regard to the cause of ESRD or tients with chronic renal insufficiency do not undergo other clinical conditions. Assessment of renal diagnosis diagnostic renal biopsy; it is not unusual for patients to was approached differently in the present study com- receive the default diagnosis of “hypertensive nephropa- pared with United States Renal Data System (USRDS) data. thy” or “diabetic nephropathy” in such primary sources. In the USRDS data set, renal diagnosis is drawn from the We also considered possible confounding by non- HCFA [Health Care Financing Administration] 2728 renal clinical conditions, such as diabetes, hyperten- form, where providers of renal replacement therapy en- sion, malignancy, congestive heart failure, or coronary ter the most likely renal diagnosis for a given patient. In- heart disease. As expected, there was high collinearity be- stead, we screened for all renal diagnoses assigned in the tween some comorbidities (eg, diabetes) and their re- year prior to onset of maintenance dialysis by all physi- lated renal diagnoses (eg, diabetic nephropathy). When cians and hospitals caring for the study subjects. Hence, adding comorbid conditions to the model, or when re- renal diagnoses are not mutually exclusive in our study. placing renal comorbidities with nonrenal comorbid con- Nonetheless, even in studies using primary medical rec- ditions, the observed findings were virtually identical (risk (REPRINTED) ARCH INTERN MED/ VOL 162, SEP 23, 2002 WWW.ARCHINTERNMED.COM 2005 Downloaded from on October 20, 2010 ©2002 American Medical Association. All rights reserved.
  • ratiolate referral, 1.35; P .001). It is possible that the sud- tions,20 as well as health policy and resource allocation den onset of rapidly progressive renal failure might have decisions.21 been associated both with a higher death rate on dialy- sis as well as with a lower rate of nephrologist visits in Accepted for publication February 6, 2002. the preceding year. However, this possibility was ad- This work was supported by grant R0-1-HS09398 from dressed by requiring that all patients studied have a di- the Agency for Healthcare Research and Quality. Addi- agnosis of chronic renal failure extending at least a full tional support was provided by the Health Care Financing year prior to the initiation of dialysis. Administration. Our findings cannot at this point resolve the ques- Corresponding author and reprints: Jerry Avorn, MD, tion of whether the utilization of nephrologist services Division of Pharmacoepidemiology and Pharmacoeconom- may also serve as a marker for other characteristics that ics, Brigham and Women’s Hospital, 221 Longwood Ave, BLI/ may also be important risk factors for mortality in ESRD. 341, Boston, MA 02115. Socioeconomic status does not appear to play a role, since we did not find an increased adjusted risk of death in Med- REFERENCES icaid or PAAD patients compared with those who were less indigent. However, use of specialist services may cor- 1. United States Renal Data System. USRDS 2000 Annual Data Report. Bethesda, relate with other issues of access and quality of care, which Md: National Institutes of Health, National Institute of Diabetes and Digestive and in themselves may play an important role in outcomes. Kidney Diseases; 2000. 2. United States Renal Data System. USRDS 1997 Annual Data Report. Bethesda, For example, patients who are referred to a nephrolo- Md: National Institutes of Health, National Institute of Diabetes and Digestive and gist may have a primary care physician who is also more Kidney Diseases; 1997. conscientious at managing other aspects of their care, such 3. Stehman-Breen CO, Sherrard DJ, Gillen D, Caps M. Determinants of type and timing of initial permanent hemodialysis vascular access. Kidney Int. 2000;57: as hypertension or nutrition. Or, such patients may them- 639-645. selves be more compliant with recommendations for their 4. United States Renal Data System. USRDS 1999 Annual Data Report. Bethesda, care, or more adept at navigating the health care system. Md: National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases; 1999. Alternatively, the nephrologist involvement may make 5. Chertow GM, Johansen KL, Lew N, Lazarus JM, Lowrie EG. Vintage, nutritional referred patients more likely to be treated with erythro- status, and survival in hemodialysis patients. Kidney Int. 2000;57:1176-1181. 6. Obrador GT, Pereira BJG. Early referral to the nephrologist and timely initiation poietin for anemia, possibly lowering the risk of death, or of renal replacement therapy: a paradigm shift in the management of patients to have a permanent vascular access created for mainte- with chronic renal failure. Am J Kidney Dis. 1998;31:398-417. nance dialysis, especially a primary fistula. This in turn 7. Ifudu O, Dawood M, Homel P, Friedman EA. Excess morbidity in patients start- ing uremia therapy without prior care by a nephrologist. Am J Kidney Dis. 1996; may reduce infectious risk during renal replacement 28:841-845. therapy and/or improve dialysis doses. Other testable hy- 8. Hakim RM, Breyer J, Ismail N, Schulman G. Effects of dose of dialysis on mor- potheses are that patients who see a nephrologist earlier bidity and mortality. Am J Kidney Dis. 1994;23:661-666. 9. Ifudu O, Dawood M, Iofel Y, Valcourt JS, Friedman EA. Delayed referral of black, and more often have better treatment of comorbid condi- Hispanic, and older patients with chronic renal failure. Am J Kidney Dis. 1999; tions associated with progressive renal failure, such as hy- 33:728-733. 10. Innes A, Rowe PA, Burden RP, Morgan AG. Early deaths on renal replacement pertension, malnutrition, and/or hyperphosphatemia, or therapy: the need for early nephrological referral. Nephrol Dial Transplant. 1992; may be better prepared psychologically for dialysis, re- 7:467-471. sulting in better compliance with dietary and/or fluid regi- 11. Sesso R, Belasco AG, Ajzen H. Late diagnosis of chronic renal failure. Braz J Med Biol Res. 1996;29:1473-1478. mens. If processes of care for pre-ESRD patients differ across 12. Roubicek C, Brunet P, Huiart L, et al. Timing of nephrology referral: influence on specialties and these differences can have a beneficial im- mortality and morbidity. Am J Kidney Dis. 2000;36:35-41. pact on ESRD mortality, follow-up studies will be neces- 13. Winkelmayer WC, Glynn RJ, Levin R, Owen W, Avorn J. Determinants of de- layed nephrologist referral in patients with chronic kidney disease. Am J Kidney sary to further define these and other possibilities. This Dis. 2001;38:1178-1184. may be of particular importance if manpower constraints 14. Owen WF Jr, Chertow GM, Lazarus M, Lowrie EG. Dose of hemodialysis and sur- vival: differences by race and sex. JAMA. 1998;280:1764-1768. in the coming years limit the number of nephrologists avail- 15. Friedman EA. Selection bias impacts outcome reports of uremia therapy [edito- able to provide such additional care; might there be some rial]. Am J Kidney Dis. 2000;36:208-210. particular interventions that could also be performed by 16. McClellan WM, Soucie M. Facility mortality rates for new end-stage renal dis- ease patients: implications for quality improvement. Am J Kidney Dis. 1994;24: a primary care physician or another consultant if nephrolo- 280-289. gist care were in short supply? 17. McClellan WM, Frederick PR, Helgerson SD, Hayes RP, Ballard DJ, McMullan If these findings are replicated in other settings, es- M. A data-driven approach to improving the care of in-center hemodialysis pa- tients. Health Care Financ Rev. 1995;16:129-140. pecially in younger and less indigent cohorts, it will be 18. Steinman TI. The dialysis facility of the future: the financial and social environ- important to further define the mechanism through which ment. Semin Nephrol. 1997;17:298-305. 19. Steinman TI. Nephrology workforce shortfall: solutions are needed. Am J Kid- early and/or more frequent nephrologist input appears ney Dis. 1999;33:798-800. to have a beneficial effect on patient outcomes. If con- 20. Ritz E, Koch M, Fliser D, Schwenger V. How can we improve prognosis in dia- firmed, such findings could have important implica- betic patients with end-stage renal disease? Diabetes Care. 1999;22:B80-B83. 21. Chandna SM, Schulz J, Lawrence C, Greenwood RN, Farrington K. Is there a ra- tions for quality improvement programs,16,17 manpower tionale for rationing chronic dialysis? a hospital based cohort study of factors projections,18,19 the care of particular high-risk popula- affecting survival and morbidity. BMJ. 1999;318:217-223. (REPRINTED) ARCH INTERN MED/ VOL 162, SEP 23, 2002 WWW.ARCHINTERNMED.COM 2006 Downloaded from on October 20, 2010 ©2002 American Medical Association. All rights reserved.