Primary care physicians’ familiarity, beliefs, and perceived barriers to practice guidelines in non-diabetic CKD: a survey study

Chronic kidney disease (CKD) affects millions of Americans [1] and estimates suggest that over 50% of adults in the United States (US) will develop CKD during their lifetime [2]. Primary care providers (PCPs) deliver the majority of non-dialysis dependent CKD care and this trend is expected to continue due to limitations in the nephrologist workforce [3].

Multiple studies have documented deficiencies in PCP delivery of CKD care. These include suboptimal screening/monitoring of patients with CKD risk factors [4, 5], infrequent discussions between providers and patients regarding CKD [6], suboptimal albuminuria testing in CKD patients [7, 8], suboptimal blood pressure control [9], and suboptimal renin-angiotensin blockade in CKD patients with proteinuria [10, 11]. In light of these deficiencies, studies have also demonstrated shortcomings in PCP knowledge of CKD risk factors [12, 13] and poor awareness of Kidney Disease Outcomes Quality Initiative (KDOQI) clinical practice guidelines [14‐17].

Recently, Kidney Disease Improving Global Outcomes (KDIGO) released updated CKD clinical practice guidelines [18]. For PCPs, perhaps the most salient modification is the emphasis on albuminuria as a risk stratification marker for poor outcomes. We employed the knowledge, attitudes, and behavior framework adapted by Cabana [19] to study reasons for PCP non-adherence to key components of the 2012 KDIGO CKD guidelines (i.e., assessing estimated glomerular filtration rate [eGFR] and albuminuria/proteinuria, definition of CKD, and use of angiotensin converting enzyme inhibitor [ACEi] or angiotensin II receptor blockers [ARBs] in overtly albuminuric/proteinuric patients) [18], focusing on non-diabetic CKD. We chose to concentrate on non-diabetic CKD because several studies have noted superior PCP performance in diabetic CKD (including albuminuria quantification and ACEi/ARB use) and high rates of familiarity with the American Diabetic Association’s clinical practice guidelines [7, 12, 20]. We also assessed interventions PCPs would find most acceptable to optimize CKD care.

We found that PCPs overwhelmingly agreed that eGFR values were helpful. Most also reported testing for albuminuria in their non-diabetic hypertensives regardless of eGFR; however, frequent barriers cited included a lack of effect on management, limited time, and the perceived absence of guidelines recommending albuminuria testing. While PCPs expressed very high agreement with the definition of CKD in patients with marked decrements in eGFR (eGFR < 45 ml/min/1.73 m²) or decrements in eGFR coupled with albuminuria, agreement was less robust when eGFR was > 60 ml/min/1.73 m² or in CKD stage 3a without albuminuria. Similarly, a substantial minority of PCPs were concerned with overdiagnosis in older adults with an eGFR of 45-59 ml/min/1.73 m². While most PCPs felt that ACEi/ARBs improved outcomes in CKD, even in the absence of albuminuria, agreement with the potential benefits of ACEi/ARBs was surprisingly highest with moderate rather than severe albuminuria. PCPs reported a lack of familiarity with KDOQI/KDIGO guidelines; however, most were receptive to systematic interventions beyond CME to improve their CKD care.

Few have previously examined PCPs’ beliefs, attitudes, and perceived barriers to eGFR reporting, albuminuria testing, CKD definitions, and ACEi/ARB use [30]. Since the publication of the KDOQI guidelines and the implementation of automated eGFR reporting, CKD has garnered increasing public and provider attention. Unfortunately, studies have documented limited improvements in CKD-related processes of care in recent years, including testing for proteinuria and use of ACEi/ARBs [7, 8, 38‐41]. Previous surveys suggested that knowledge deficiencies [12, 13, 16, 17, 42] may contribute to continued gaps in care. In contrast to prior studies where only serum creatinine and proteinuria were used [16, 42], we found that PCPs were relatively knowledgeable in identifying CKD based on eGFR and albuminuria categories. However, PCPs were less attuned to isolated albuminuria abnormalities (Figure 1). For example, while 99% of PCPs recognized that patients with a severe reduction in eGFR to 30–44 ml/min/1.73 m² with “normoalbuminuria” had CKD, only 84% of PCPs recognized that patients with severe albuminuria and relatively preserved eGFR had CKD (P < 0.001).

This distinction is clinically important because the renal benefits of ACEi/ARBs in non-diabetic CKD primarily accrue in severely albuminuric patients [43, 44] and the debated benefits of lower targeted blood pressure in CKD also primarily apply to albuminuric/proteinuric patients [45, 46]. The under appreciation of the significance of albuminuria may be due in part to the greater emphasis on eGFR in the 2002 KDOQI guidelines. As recent studies highlight the importance of albuminuria as a marker for poor outcomes [47‐49], initiatives to disseminate this information and help PCPs understand how to use albuminuria to modify their CKD care are needed. The most frequently cited barrier to testing for albuminuria was that it would have no impact on management, perhaps because responding PCPs presumed such patients would already be receiving ACEi/ARBs. However, recent data suggests that active ACEi/ARB use is unlikely to fully explain the low rates of albuminuria/proteinuria testing [7, 38, 50]. We suspect limited time and a lack of clarity on how to integrate information regarding albuminuria in developing non-diabetic CKD care goals is also contributing to low rates of testing.

Prior studies have documented suboptimal ACEi/ARB use in CKD patients with proteinuria [10, 11]. We found that most PCPs felt that non-diabetic patients with an eGFR < 60 ml/min/1.73 m² would benefit from an ACEi/ARB regardless of proteinuria. These patterns may be due in part to greater familiarity with recommendations from alternative guidelines [51] that favor ACEi/ARB use as first line treatment in patients with CKD and hypertension regardless of albuminuria. Of note, similar to prior studies [15‐17], we confirmed low familiarity with KDOQI/KDIGO CKD guidelines.

Our findings support efforts to disseminate the KDIGO CKD guidelines and to provide PCPs with guidance regarding the use of albuminuria and other factors to identify patients at high risk for poor outcomes [49] and most likely to benefit from ACEi/ARBs. The most commonly endorsed barrier to ACEi/ARB use was adverse effects. Studies are needed to determine whether targeting high-risk CKD patients who may be more likely to achieve benefits from these medications [52] could reduce adverse effects and improve outcomes in older adults with multiple medications (e.g., NSAIDs, diuretics) [53] compared to other less nuanced recommendations [51, 54]. Such an approach could also help PCPs identify a smaller cohort of ACEi/ARB intolerant patients with a higher likelihood to benefit from the use of these medications. In the setting of severe albuminuria, these individuals may benefit from nephrologist input to determine if alternative strategies can allow for ACEi/ARB use.

Although a previous survey of PCPs noted patient non-adherence and the cost of medications as major obstacles to appropriate care and improving patient outcomes [16], few respondents in our survey identified these as obstacles that prevent them from ordering urine albumin tests or ACEi/ARBs. This may reflect differences in the underlying populations the responding PCPs serve and changes in the cost of medications, including the greater availability of generic ACEi/ARBs.

Many of the barriers we identified are potentially addressable using systematic interventions. For example, decision support within an electronic health record could alert a provider as s/he orders a serum creatinine that a patient with possible CKD does not have a documented urine albumin test and is not receiving an ACEi/ARB. The alert could also provide references to guidelines and the studies that informed those guidelines. Similarly, collaborative practice agreements or audit and feedback could target high-risk CKD patients to ensure they receive optimal, evidence-based care. Combining several of these interventions while targeting patients who are most likely to benefit from ACEi/ARBs [52] may prove even more impactful. Most PCPs were receptive to at least one of these approaches and studies are needed to understand whether such strategies can improve care without substantially disrupting PCP workflow [55].

Our findings should be interpreted in light of several limitations. First, although we employed best practices for electronic surveys [37], the survey response rate was low, which may limit generalizability. Responders may have substantively differed from non-responders. For example, they may have had an interest in kidney disease and been more knowledgeable than the average PCP. Respondents were more likely to be internists and had graduated medical school more recently, characteristics that may suggest greater familiarity with CKD guidelines [13, 16, 42, 56]. Indeed, self-reported guideline familiarity was modestly higher than in prior PCP survey studies [12, 15, 16], although it remained less than 50%. Second, the majority of targeted PCPs never opened a study email. While we were able to exclude individuals with non-valid emails, we could not determine whether an unopened message was due to an unmonitored email account, automated filtering into an infrequently checked folder (e.g., junk), a choice not to open the email based on the subject matter or sender, or other reasons. Including individuals with unopened emails would lower the response rate. Third, PCPs may not have accurately recalled their practice patterns or may have attempted to provide what they perceived as the correct answer. To mitigate this risk, we framed survey questions as assessing beliefs rather than knowledge. Finally, there were additional questions that would have been helpful to examine such as the perceived utility of cystatin C [18]. Nevertheless, our findings provide novel information regarding PCPs’ beliefs and reported barriers to adoption of recent CKD practice guidelines. Future work to examine how local resources, healthcare environments, and patient characteristics influence providers’ beliefs, perceived barriers, and practice patterns is warranted.

Our survey found that PCPs value eGFR measures and are able to recognize CKD based on a decrement in eGFR. However, they were less familiar with albuminuria abnormalities and endorsed several barriers to testing for albuminuria in non-diabetics including lack of outcome expectancy, lack of guideline familiarity, and lack of time. While PCPs believed that ACEi/ARBs improve outcomes in various eGFR/albuminuria stages of CKD, future interventions will be needed to help operationalize how to identify patients at high risk for progression and who are more likely to benefit from ACEi/ARBs. Fortunately, most PCPs appear receptive to interventions to improve their CKD care.