To the Editor: Chronic kidney disease (CKD) affects more than one-third of adults aged 70 and older.1 Referral to a nephrologist is recommended for all patients with Stage 4 CKD (glomerular filtration rate (GFR)<30 mL/min per 1.73 m2),2 but older patients are often not referred as recommended.3–5 Clinical vignettes were used to better understand how patient factors and physician specialty influence the decision to refer a patient to a nephrologist.
A cross-sectional survey was conducted of internists, geriatricians, and nephrologists at the University of Chicago in 2007, as approved by the Biological Sciences Division institutional review board.
Five clinical vignettes were created about patients with Stage 4 CKD (GFR 18–24 mL/min per 1.73 m2). The vignettes varied patient age, comorbidities, and degree of functional and cognitive impairment.
The main outcome measure was the physician's decision to refer hypothetical patients to a nephrologist. Respondents rated how strongly (0–10) each patient factor affected their recommendation to refer (0=did not affect at all and 10=very strongly affected).
Chi-square tests (χ2) were used to assess differences between physician demographic characteristics, rates of referral between physician specialties, and awareness of practice guidelines. Ratings of how strongly patient factors affected the referral decision were assessed using analysis of variance. All statistical analyses were performed using STATA, version 9.2 (Stata Corp., College Station, TX).
Sixteen internists, 14 geriatricians, and eight nephrologists completed the survey (55.1% response rate). Respondents had a mean age of 42 and were mostly female (63.2%) and white (71.0%).
Referral rates varied between vignettes from more than 80% of all physicians recommending referral in Vignette 1 to only 31.6% recommending referral in Vignette 3 (Table 1). Physician age, sex, race, and amount of time spent in direct patient care did not predict a decision to refer.
Referral Is Appropriate
|(1) 75-year-old man with diabetes mellitus and congestive heart failure with an ejection fraction of 25%. No cognitive impairment. Lives at home with his daughter and is independent in ambulating, dressing, grooming, and bathing. Creatinine 2.8 (GFR 24)||Minor||Absent||None||31 (81.6)||13 (81.3)||10 (76.9)||8 (100.0)||.32|
|(2) 91-year-old woman with well-controlled hypertension for 30 years, diabetes mellitus for 15 years, and rheumatoid arthritis. She is cognitively intact and lives alone but requires some assistance with shopping. She describes increasing fatigue. Creatinine 2.8 (GFR 20)||Minor||Absent||None||22 (57.9)||7 (46.7)||7 (50.0)||8 (100.0)||.03|
|(3) 80-year-old woman who is bed-bound since recent hip fracture. She has been a nursing home resident for the last 5 years with long-standing hypertension for 20 years and corticosteroid-dependent chronic obstructive pulmonary disease. She has mild cognitive impairment (MMSE score 24/30). Creatinine 2.8 (GFR 18)||Major||Present||Mild||12 (31.6)||3 (20.0)||1 (7.1)||7 (87.5)||<.001|
|(4) 84-year-old woman with poorly controlled hypertension for 30 years, history of cerebrovascular accident with residual right-side weakness and aphasia and moderate dementia (MMSE score 17/30). Lives at home with daughter; ambulates with cane; requires assistance with shopping, managing finances, and bathing. Creatinine 2.8 (GFR 21)||Major||Present||Moderate||18 (47.4)||7 (43.8)||4 (28.6)||7 (87.5)||.03|
|(5) 68-year-old man with metastatic prostate cancer. Experiences moderate pain that restricts activities outside the home. Lives with his wife, has moderate cognitive impairment (MMSE score 15/30) and depression. Creatinine 2.8 (GFR 24)||Major||Present||Moderate||13 (34.2)||6 (37.5)||1 (7.1)||6 (75.0)||.01|
There were significant differences between the physicians' recommendation to refer based on medical specialty in all scenarios except Vignette 1, with nephrologists most likely to recommend referral and geriatricians least likely (P=.005, Table 1). There was greater variation in the tendency to refer across vignettes for geriatricians (referral rate 7–77%) and internists (20–81%) than nephrologists (75–100%).
There were significant differences between specialists in how they rated the importance of patient characteristics in their decision to refer. Geriatricians rated GFR as less important (mean 6.2±1.9) than nephrologists (9.4±1.3) or internists (7.4±2.2) (P=.005). Geriatricians also rated patient age (4.4±2.1) as less important than did nephrologists (7.5±2.0) or internists (5.1±1.9) (P=.01).
Patients without functional or cognitive impairment were referred by 63.5% of geriatricians, 64.0% of internists, and 100% of nephrologists (P=.73). Patients with functional impairment were referred by 14.3%, 33.8%, and 88.3% (P=.04), respectively. Patients with moderate cognitive impairment were referred by 17.9%, 40.7%, and 81.3% (P=.23), respectively (Table 1).
Report of awareness of CKD practice guidelines was 100% for nephrologists, compared with 31.3% of internists and 57.1% of geriatricians (P=.01). Guideline awareness was associated with the referral decision in Vignettes 1 (P=.01) and 3 (P=.04).
Vignettes were created to meet guideline referral criteria, yet only 50% of the decisions were to refer to the nephrologist. These data are consistent with other studies showing that older patients with CKD are not referred to nephrology.3,5 The discrepancy between guideline recommendations and nephrology referral has been attributed to a lack of awareness of CKD guidelines,6,7 although results of this pilot study suggest that patient characteristics and physician specialty may also be important in this decision.
Significant differences were found between specialist ratings of how strongly patient factors affected the referral decision. For example, although nephrologists rated GFR as very strongly affecting their recommendations, geriatricians rated it much less strongly. The perceived importance of a patient characteristic may determine how much weight it is given in the decision and may, in part, explain why different groups of physicians treated patients with similar levels of severe kidney impairment differently.
Significant differences were also found between groups of physicians in their decision to refer patients. Fewer geriatricians and internists recommended referral for patients than nephrologists, as most dramatically illustrated in Vignette 3. The presence of functional and cognitive impairment in vignette patients may account for some of the discrepancy in referral decisions.
This small, single-center pilot study of physician decision-making suggests that patient characteristics and physician specialty, in addition to guideline awareness, may be important in the assessment of the benefit of referral for individual patients. The reasons for failure to refer older patients with advanced CKD to a nephrologist are probably more complex than previously postulated.
The authors thank Ann M. O'Hare, MD, and Carol Stocking, PhD, for their contributions to this letter.
Conflict of Interest: The editor in chief has reviewed the conflict of interest checklist provided by the authors and has determined that the authors have no financial or any other kind of personal conflicts with this paper.
Dr. Stankus is on the speaker's bureau for Abbott and Amgen. Dr. Sachs is a member of the Editorial Board at the Journal of the American Geriatrics Society and an American Geriatrics Society board member. Funded by the John A. Harford Foundation Center of Excellence in Geriatrics at the University of Chicago.
Author Contributions: Drs. Campbell, Dale, Hemmerich, Smith, and Sachs: study concept and design. Drs. Campbell, Smith, and Stankus: acquisition of subjects and data. Drs. Campbell, Hemmerich, Dale, Smith, and Sachs: analysis and interpretation of data. All authors: preparation of the letter.
Sponsor's Role: The sponsor had no role in the design, methods, subject recruitment, data collection, analysis, or preparation of this letter.