Complications of chronic kidney disease (CKD), such as anemia, metabolic acidosis, nutritional deficits, secondary hyperparathyroidism and hypertension occur well before renal replacement therapy is needed and significantly contribute to morbidity and mortality [1–5]. National treatment guidelines for CKD, and for the management of hypertension, dyslipidemia, bone disease, nutrition, anemia and cardiovascular disease in patients with CKD, have been published [6–12], and dedicated CKD clinics were established based on the conviction that such clinics would help implement Kidney Disease Outcomes Quality Initiative (KDOQI) goals and thus improve outcomes [13, 14]. This belief was supported by the observation that patients with already-established advanced CKD who were referred late to a nephrologist had worse outcomes than those referred earlier [15, 16], and several groups of researchers [17–20] have postulated that CKD care could be improved by early detection, early referral, and a structured approach to CKD care. An analysis based on the National Health and Nutrition Examination Surveys (NHANES) III and Medicare databases showed that CKD care is suboptimal [21], and recent data suggests that primary care providers may not be familiar will KDOQI guidelines [22]. However, given the vast number of patients with CKD and the limited number of nephrologists [23, 24], most patients with early CKD are not seen by a nephrologist. Therefore, this web-based survey was designed to ascertain perceptions and practice patterns regarding CKD care in family medicine and internal medicine trainees.

Delayed referral of CKD patients to a nephrologist has been identified as an important predictor of poor outcomes [15, 16]. However, in order to initiate timely referral of CKD patients, primary providers need to be aware of risk factors for CKD and co-morbidities associated with CKD, as well as clinical practice guidelines describing optimal CKD care, such as KDOQI guidelines, the recently-published recommendations from the American Diabetes Association (ADA) [27], or the clinical practice guidelines developed by the Renal Physicians Association [28], the latter being published as an executive summary with clear "hands-on" guidance as to the management of patients with an estimated GFR of less than 30 ml/min/1.73 m². Several educational efforts are currently under way to enhance awareness of CKD [29, 30]. However, the data presented in this report show that both family medicine and internal medicine trainees have important knowledge gaps when it comes to CKD care.

Similar to data recently published by Lea et al [22], almost all physicians identified diabetes and hypertension as strong risk factors, while minority status or family history did not receive the same recognition. This may be due to the absence of a unifying model predicting CKD risk in the general population [31]. Some risk factors, such as diabetes and hypertension, have clearly been established [32]. However, attempts to quantitate the contribution of others, such as race, ethnicity, and socio-economic status, has proven to be more challenging because of the varying prevalence of co-morbid diseases such as hypertension [33–35]. In addition, the prevalence of CKD among different racial and ethnic groups may vary according to the stage of CKD [36]. Thus, it may be necessary to perform and pool data from large, population-based studies to further delineate the contribution of individual CKD risk factors [37, 38], allowing primary providers to screen for CKD akin to cardiovascular risk assessment based on the Framingham Study [39].

The use of mathematical formulas to estimate GFR rather than a timed urine collection for creatinine clearance is recommended for most patients [40], and the examination of a random urine sample for albumin and creatinine is the preferred screening method for albuminuria in adults [7, 27]. However, despite this, more than half the physicians preferred a timed urine collection for creatinine, and in a hypothetical patient with diabetes and stage 3 CKD, only 70% of physicians elected to screen for microalbuminuria. Moreover, among those who did screen for microalbuminuria, only about a third would concomitantly obtain a urine creatinine concentration to normalize the albumin result (family medicine trainees: 19%, internal medicine trainees: 37%), as recommended by both the NKF and ADA [27, 41]. Thus, further emphasis may need to be placed on hands-on implementation of clinical practice guidelines to improve the detection of subjects with CKD.

The vast majority of respondents identified blood pressure goal and first line antihypertensive agents for patients with CKD as recommended by the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure [42]. However, practitioners, and in particular family medicine residents, were hesitant to use ACEi in the setting of stage 3 or 4 CKD. Data supporting the use of ACEi and ARB show that these agents are not only beneficial in patients with an abnormal serum creatinine, but also safe, as long as patients are carefully monitored and counseled [43, 44]. Thus, it may be necessary to not only disseminate information about the usefulness of a given intervention, but also hands-on information about monitoring, prevention, and treatment of potential complications.

Almost all physicians recognized worsening hypertension, anemia, hyperkalemia, volume overload, and metabolic acidosis as complications of CKD. However, secondary hyperparathyroidism, malnutrition, and lipid disorders were identified by a significantly smaller proportion of respondents. Moreover, even though about 90% of physicians recognized anemia as a complication of CKD, only about 40% indicated that they would obtain a CBC in a hypothetical patient with diabetes and an eGFR of 34 ml/min/1.73 m². Similarly, even among those who identified secondary hyperparathyroidism as a CKD complication, only a small fraction would order the appropriate screening test. This reflects the observation that the majority of patients with stage 4 CKD referred to a nephrologist were never screened for secondary hyperparathyroidism and had lower hemoglobin concentrations than those seen in a dedicated CKD clinic for at least six months [45]. In addition, the management of anemia and secondary hyperparathyroidism requires complex decision-making. These disorders have not only to be recognized as a complication of CKD, but in order to avoid future complications, screening needs to be initiated in stage 3 CKD and physicians also need to know about recommended treatment targets in order to either initiate therapy or consult an experienced specialist [9, 12]. The data from this survey show that no more than 10% of physicians had the skills to integrate data from all three areas correctly, allowing them to formulate appropriate management plans for CKD patients with anemia or secondary hyperparathyroidism. Thus, unless an educational effort aimed specifically at non-nephrologists fills these knowledge gaps, nephrologists will likely have to co-manage patients with stage 3 and 4 CKD.

Clinical practice guidelines may present treatment goals, but often lack the necessary guidance as to why and how to reach these goals. Recently, an approach to link evidence with practice in CKD care, which may be particularly useful in managed care settings, has been published [46]; however, none of the strategies to improve compliance tested thus far have proven fail-safe [47]. Deficits in compliance with clinical practice guidelines are not unique to nephrology, and the reasons for non-compliance are complex [48]. Recent data show that there is an acute lack of awareness of clinical practice guidelines for CKD patients among primary providers [22, 49]. Based on the data presented here, educational efforts aimed at improving CKD care will have to start during the training program, even though this may be a challenging undertaking given the requirements and restrictions placed upon US training programs [50]. Specifically, trainees will need to be instructed about prevalence of CKD, risk factors for CKD, screening methods for CKD and its complications, and treatment options to slow the progression of CKD and reduce morbidity and mortality [51–54]. In order to succeed, guidelines will have to be presented in a manner that fosters reflection, critically appraises clinical evidence, is practice-based, and is easy to understand for practitioners and patients alike [48, 55].

This study has several important limitations. Given that it is based on a voluntary, anonymous online survey, certain important variables, such as training program size and setting, or the presence or absence of formal teaching in CKD care, cannot be controlled. In addition, demographic characteristics of the study cohort cannot be verified. Finally we cannot be certain that the answers truly reflect the knowledge of the person completing the survey, given that the survey instrument has not been formally validated in this respect, although we did include nephrologists who might serve as a positive control group (see Additional file 2). However, given that we received a large number of responses from across the United States, we believe that the data presented here are representative and covey important information about practice patterns of future providers.

In conclusion, the results from this survey suggest that physicians currently in training have significant knowledge gaps in CKD management, and with few exceptions attendings' responses closely mimicked those of trainees. These finding suggests that education is needed to raise awareness of clinical practice guidelines and recommendations for patients with CKD among future practitioners. Surveys similar to ours may be useful in the planning of training programs to assess areas that need additional emphasis.