Trends in Quality of Care for Patients with CKD in the United States

Abstract

Visual Abstract

Background and objectives 

Improving the quality of CKD care has important public health implications to delay disease progression and prevent ESKD. National trends of the quality of CKD care are not well established. Furthermore, it is unknown whether gaps in quality of care are due to lack of physician awareness of CKD status of patients or other factors.

Design, setting, participants, & measurements 

We performed a national, serial, cross-sectional study of visits to office-based ambulatory care practices for adults with diagnosed CKD from the years 2006 to 2014. We assessed the following quality indicators: (1) BP measurement, (2) uncontrolled hypertension, (3) uncontrolled diabetes, (4) angiotensin-converting enzyme inhibitor or angiotensin receptor blocker use among patients with hypertension, (5) statin use if age ≥50 years old, and (6) nonsteroidal anti-inflammatory drug use. Using multivariable linear regression and chi-squared analysis, we examined the change in quality performance over time.

Results 

Between 2006 and 2014, there were 7099 unweighted visits for patients with CKD representing 186,961,565 weighted visits. There was no difference in the prevalence of uncontrolled hypertension (>130/80 mm Hg) over time (46% in 2006–2008 versus 48% in 2012–2014; P=0.50). There was a high prevalence of uncontrolled diabetes in 2012–2014 (40% for hemoglobin A1c >7%). The prevalence of ACEi/ARB use decreased from 45% in 2006–2008 to 36% in 2012–2014, which did not reach statistical significance (P=0.07). Statin use in patients with CKD who were 50 years or older was low and remained unchanged from 29% in 2006–2008 to 31% in 2012–2014 (P=0.92).

Conclusions 

In a nationally representative dataset, we found that patients with CKD had a high prevalence of uncontrolled hypertension and diabetes and a low use of statins that did not improve over time and was not concordant with guidelines.

Introduction

CKD is highly prevalent, affecting over 30 million Americans (¹). CKD and ESKD are associated with high morbidity, mortality, and health care costs (²). Improving the quality of CKD care has important public health implications to delay disease progression, prevent ESKD, and improve health status.

There has been enormous investment in quality of care measurement in outpatient practices over the past decade. Mandatory or voluntary performance assessment is used for federal reporting, alternative payment models through the Centers for Medicare and Medicaid Services, and risk-based contracts with commercial payers (³). Quality measurement in CKD in particular poses specific challenges. First, physician awareness of CKD is extremely low, and therefore, it is unknown whether gaps in quality of care are due to lack of physician awareness of CKD status of patients or other factors (⁴). Second, national quality measures lack CKD-specific quality indicators, such as appropriate medication use and nephrotoxin avoidance (⁵,⁶).

The Centers for Disease Control (CDC) CKD Surveillance System and several other studies have found gaps in the quality of care in CKD, including BP control, medication use, and CKD detection (¹,⁷). However, some prior studies rely on data from cohorts, single centers, Veterans Affairs medical facilities, or international settings that may not be generalizable to the CKD population in the United States. Thus, the effect of an increased emphasis on performance measurement on the quality of CKD care remains unclear, and longitudinal trends in the quality of CKD care are not well established (Supplemental Table 1).

In this study, we assess temporal trends in the quality of CKD care among adults visiting office-based practices in the United States using a nationally representative dataset. We examine the quality of CKD care among patients with physician-diagnosed CKD as a proxy for physician CKD awareness.

Materials and Methods

Data Source

We performed a national, serial cross-sectional study using data from the National Ambulatory Medical Care Survey (NAMCS) from the years 2006–2014. The NAMCS is a federally funded national survey performed annually by the National Center for Health Statistics within the CDC. The NAMCS sample includes participants in the United States who attended an office-based ambulatory care visit excluding federally operated, military, institutional, or educational facilities. The NAMCS uses a visit-based sampling method by selecting a sample of practicing physicians from 112 primary sampling units. A random sample of patient visits is selected during 1 week. CDC field representatives abstract data for the sampled visits from patient medical records, including patient demographics, laboratory and imaging tests ordered, physician-reported diagnoses, and medications (prescription and over the counter). These data are entered into the NAMCS electronic Patient Record Form. In addition, practice characteristics of the facilities are obtained. Medications are coded using Lexicon Plus, a comprehensive database of all prescription and some nonprescription drug products available in the United States. The National Hospital Ambulatory Medical Care Survey was not available for the years 2012–2014, and therefore, visits to hospital outpatient departments were not included in this study.

Study Population

To be eligible for our analysis, visits had to made by adults 18 years old and older who had an International Classification of Diseases, Ninth Edition (ICD-9) code for CKD or a physician-reported diagnosis of “chronic renal failure” (2006–2013) or “chronic kidney disease” (2014–2015). Patients with ESKD were excluded. Patients were considered to have hypertension if they had an ICD-9 code consistent with hypertension, receipt of antihypertensive medications, or a physician-reported diagnosis of hypertension. Patients with diabetes were defined as having an ICD-9 code of diabetes, receipt of antidiabetic medications, or physician-reported diagnosis of diabetes. Relevant ICD-9 codes are listed in Supplemental Table 2.

Quality Indicator Outcomes

We systematically assembled quality indicators related to CKD on the basis of national quality measures, clinical practice guidelines, and the medical literature. Quality indicators were compiled from the following professional societies and organizations: Kidney Disease Improving Global Outcomes (KDIGO), Kidney Disease Outcomes Quality Initiative, the National Kidney Foundation, the Renal Physicians Association, the Healthcare Effectiveness Data and Information Set, the Physician Quality Reporting System, and the National Quality Forum. Quality indicators developed via consensus-based approaches were also reviewed (^8–11). Quality indicators contained in the 2017 American Heart Association/American College of Cardiology (AHA/ACC) guidelines were not included, because they were released after the data for this study were collected (¹²). Among the potential quality indicators identified, six had data elements available in the NAMCS and were included in our analysis: (1) BP measurement, (2) uncontrolled BP among patients with hypertension, (3) uncontrolled hemoglobin A1c (HbA1c) among patients with diabetes, (4) angiotensin-converting enzyme inhibitor (ACEi) or angiotensin receptor blocker (ARB) use among patients with hypertension, (5) statin use if age ≥50 years old, and (6) nonsteroidal anti-inflammatory drug (NSAID) use. Detailed information on the quality indicators is provided in Supplemental Table 1 and Table 1.

Table 1. - Definitions of CKD quality of care indicators

Quality Indicator	Numerator (Visits by Adults with)	Denominator (Visits by Adults with)	Quality Indicator Source	Strength of Recommendation
BP measured	BP recorded during visit	CKD (physician reported or ICD-9 code)	KDOQI (51)	Not graded
Uncontrolled hypertension	Systolic BP >140 or diastolic BP >90 or systolic BP >130 or diastolic BP >80	(1) CKD (physician reported or ICD-9 code) and (2) hypertension (physician reported or ICD-9 code or receipt of antihypertensive)	JNC 7 (13) (2003) and JNC 8 (14) (2014)	E a
Uncontrolled diabetes	HbA1c>7% or HbA1c>8%	(1) CKD (physician reported or ICD-9 code) and (2) diabetes (physician reported or ICD-9 code or receipt of antidiabetic agent)	KDOQI (16)	1A b
ACEi/ARB use	ACEi/ARB listed or dispensed during visit	CKD (physician reported or ICD-9 code)	KDIGO CKD (52,53)	1B, 2D c , d
Statin use	Statin listed or dispensed during visit	(1) CKD (physician reported or ICD-9 code) and (2) age ≥50 yr	KDIGO Lipid (19)	1A b , 1B c
NSAID use	NSAIDs listed or dispensed during visit	CKD (physician reported or ICD-9 code)	KDIGO CKD (52)	Not graded

ICD-9, International Classification of Diseases, Ninth Edition; KDOQI, Kidney Disease Outcomes Quality Initiative; JNC, Joint National Committee; HbA1c, hemoglobin A1c; ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; KDIGO, Kidney Disease Improving Global Outcomes; NSAID, nonsteroidal anti-inflammatory drug.

^aE indicates expert opinion.

^b1A: Level 1 “We recommend”—high quality of evidence.

^c1B: Level 1 “We recommend”—moderate quality of evidence.

^d2D: Level 2 “We suggest”—very low quality of evidence.

We assessed whether patients had their BP measured during the visit. We defined uncontrolled hypertension using two BP cutoffs on the basis of guidelines at the times of the sampling years: (1) >130/80 mm Hg recommended by Joint National Committee 7 (JNC 7) guidelines for patients with CKD and (2) >140/90 mm Hg recommended by JNC 8 guidelines (¹³,¹⁴). HbA1c values were available from 2012 to 2014. We defined uncontrolled diabetes in two ways on the basis of differing guideline recommendations: (1) HbA1c>7% or (2) HbA1c>8% (¹⁵,¹⁶).

Analyses

We divided our data into the years 2006–2008, 2009–2011, and 2012–2014 for comparisons. Chi-squared tests were used to compare the proportions of visits meeting quality indicators between the three time periods. Using multivariable linear regression, we examined the change in quality indicator performance among CKD visits over time after adjusting for patient age, sex, hypertension, diabetes, congestive heart failure, and cardiovascular disease status. We adjusted for these factors, because age, sex, and prevalence of comorbidities affect severity of chronic diseases and medication use. We then stratified quality indicator performance by primary care, medical specialty care, and surgical care and used chi-squared tests to compare the proportion of visits meeting quality indicators by specialty category. Primary care included general/family practice, internal medicine, pediatrics (visits for participants age ≥18 years old), and obstetrics and gynecology; medical specialty care included cardiovascular diseases, dermatology, psychiatry, neurology, oncology, and “other specialties.” Surgical care included general surgery, orthopedic surgery, urology, ophthalmology, otolaryngology, and “other specialties.” We used multivariable logistic regression to examine specialty category as a predictor of quality indicator performance adjusting for patient age, sex, hypertension, diabetes, congestive heart failure, and cardiovascular disease status. Weights for patient visits were used to produce national estimates. The analysis had 80% power to detect a 2.1-point change in BP and 5% difference in medication use at significance level of α=0.05.

We chose 2006 as a start date, because it was before the passage of the Medicare Improvements for Patients and Providers Act and the Affordable Care Act, which introduced numerous quality-related reforms, and it was the first year that the NAMCS used the Multum Lexicon Drug Database for coding medications. From 2006 to 2014, there were changes in survey administration in the total number of medications recorded. Up to eight medications were recorded from 2006 to 2011, up to ten were collected in 2012 and 2013, and up to 30 were collected in 2014. Our primary analysis analyzed only the first eight medications listed to prevent measurement bias on the basis of survey collection. We performed a sensitivity analysis using all available medications to assess robustness of our analysis to the inclusion of additional medications.

Results

Baseline Characteristics

Between 2006 and 2014, there were 7099 unweighted visits for patients with CKD representing 186,961,565 weighted visits. In weighted analysis, 2.7% of visits were for patients with diagnosed CKD. Overall, all patients with visits for CKD from 2006 to 2014 had an average age of 69 years old, were 53% male, 69% non-Hispanic white, 13% non-Hispanic black, and 11% Hispanic. Patients with CKD who had visits in 2012–2014 were older and seen fewer times in the past 12 months compared with those with visits in 2006–2008 and 2009–2011 (Table 2).

Table 2. - Baseline characteristics of United States office-based visits for patients with CKD (weighted)

Characteristic	NAMCS 2006–2008, n=49,996,391	NAMCS 2009–2011, n=75,026,445	NAMCS 2012–2014, n=61,820,283	P Value
Demographics
Age	67 [15]	69 [14]	70 [13]	<0.001
Sex, %
Men	55	52	48	0.33
Women	45	48	52
Race/ethnicity, %
Non-Hispanic white	70	70	68	0.67
Non-Hispanic black	11	12	15
Hispanic	11	10	12
Non-Hispanic other	8	8	5
Comorbidities, %
Congestive heart failure	11	14	11	0.28
Cardiovascular disease	14	13	15	0.53
Hypertension	75	87	83	0.13
Diabetes	41	44	42	0.67
Geographic region, %
Northeast	21	17	17	0.56
Midwest	17	19	17
South	42	34	41
West	21	30	25
Metropolitan area, %
Yes	93	90	92	0.48
No	7	10	8
Type of payment, %
Private insurance	27	24	25	0.22
Medicare	60	67	64
Medicaid	7	4	6
Other	6	5	6
Type of specialty, %
Primary care	38	38	49	0.16
Surgical care	9	10	9
Medical specialty care	54	52	42
Visit characteristics
Seen in this practice before, %	92	91	89	0.32
No. of times seen in past 12 mo	3.8 [4.9]	4.2 [7.1]	3.3 [6.2]	<0.001

Congestive heart failure and cardiovascular disease were on the basis of physician-reported diagnosis. Hypertension and diabetes were defined using International Classification of Diseases, Ninth Edition codes or by use of medications for hypertension and diabetes or physician-reported hypertension or diabetes. Continuous variables age and number of times seen in the past 12 months are listed as mean [SD]. Percentages may not add to 100% due to rounding. NAMCS, National Ambulatory Medical Care Survey.

Hypertension and Diabetes Measurement and Control

The percentage of visits in which BP was measured increased from 89% in 2006–2008 and 2009–2011 to 93% in 2012–2014 (P=0.03) (Table 3). The prevalence of uncontrolled hypertension >130/80 mm Hg and >140/90 mm Hg was high among visits for patients with CKD. There was no statistically significant difference in the prevalence of uncontrolled hypertension (>130/80 mm Hg) across visits in the three time periods (46% in 2006–2008 versus 50% in 2009–2011 versus 48% in 2012–2014; unadjusted P=0.50). Of patients with CKD and diabetes mellitus, 17% had an HbA1c result during the visit. There was a high prevalence of uncontrolled diabetes in 2012–2014 (40% for HbA1c>7% and 24% for HbA1c>8%). Medical specialty care had a higher prevalence of uncontrolled diabetes (51% for HbA1c>7%) in those with CKD and diabetes compared with primary (34%) and surgical care (18%), which was not attenuated by adjustment for demographics and comorbidities (adjusted P=0.04) (Table 4).

Table 3. - Performance on CKD quality indicators in the United States office-based ambulatory practice

Quality Indicator	n (Weighted)	2006–2008, %	n (Weighted)	2009–2011, %	n (Weighted)	2012–2014, %	P Value a (Unadjusted)	P Value for Trend b (Adjusted)
BP measured in patients with CKD	49,996,391	89	75,026,445	89	61,820,283	93	0.03	0.02
Uncontrolled hypertension (>130/80) in patients with CKD and hypertension	37,630,268	46	65,143,017	50	51,190,632	48	0.50	0.27
Uncontrolled hypertension (>140/90) in patients with CKD and hypertension		26		30		26	0.22	0.28
Uncontrolled diabetes with HbA1c>7% in patients with CKD and diabetes	—	—	—	—	26,129,977	40	—	—
Uncontrolled diabetes with HbA1c>8% in patients with CKD and diabetes	—	—	—	—		24	—	—
ACEi/ARB use in patients with CKD and hypertension	37,630,268	45	65,143,017	39	51,190,632	36	0.07	0.003
Statin use in patients with CKD age ≥50 yr	42,297,397	29	67,893,219	30	56,994,494	31	0.92	0.66
NSAID use in patients with CKD	49,996,391	2	75,026,445	3	61,820,283	4	0.01	0.01

Missing data: 1% of patients with CKD had missing medication lists. HbA1c, hemoglobin A1c; —, not collected in dataset; ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; NSAID, nonsteroidal anti-inflammatory drug.

^aChi-squared test and unadjusted analysis.

^bP value for trend from 2006 to 2014 adjusted for age, sex, race/ethnicity, hypertension, diabetes, congestive heart failure, and cardiovascular disease.

Table 4. - Performance on CKD quality indicators in the United States office-based ambulatory practices by specialty category, 2006–2014

Quality Indicators	n (Weighted)	Primary Care, %	n (Weighted)	Medical Specialty Care	n (Weighted)	Surgical Care, %	P Value a (Unadjusted)	P Value b (Adjusted)
BP measured in patients with CKD	109,118,226	96	103,049,528	95	20,191,325	45	<0.001	0.52
Uncontrolled hypertension (>130/80) in patients with CKD and hypertension	93,326,797	47	86,492,963	49	13,568,451	49	0.66	0.59
Uncontrolled hypertension (>140/90) in patients with CKD and hypertension		26		29		30	0.51	0.49
Uncontrolled diabetes with HbA1c>7% in patients with CKD and diabetes	49,299,430	34	43,709,834	51	8,144,313	18	0.008	0.04
Uncontrolled diabetes with HbA1c>8% in patients with CKD and diabetes		17		35		18	<0.001	0.002
ACEi/ARB use in patients with CKD and hypertension	93,326,797	43	86,492,963	41	13,568,451	23	<0.001	0.24
Statin use in patients with CKD age ≥50 yr	100,684,231	30	90,895,367	30	18,042,503	18	0.04	0.94
NSAID use in patients with CKD	109,118,226	7	103,049,528	2	20,191,325	5	<0.001	<0.001

Primary care includes general/family practice, internal medicine, pediatrics (visits for participants age ≥18 years old), and obstetrics and gynecology. Medical specialty care includes cardiovascular diseases, dermatology, psychiatry, neurology, oncology, and “other specialties.” Surgical care includes general surgery, orthopedic surgery, urology, ophthalmology, otolaryngology, and “other specialties.” HbA1c, hemoglobin A1c; ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; NSAID, nonsteroidal anti-inflammatory drug.

^aChi-squared test and unadjusted analysis

^bP value for logistic regression adjusted for age, sex, race/ethnicity, hypertension, diabetes, congestive heart failure, and cardiovascular disease.

Medication Use

The prevalence of ACEi/ARB prescriptions among visits for patients with CKD decreased from 45% in 2006–2008 to 36% in 2012–2014, which did not reach statistical significance in unadjusted analysis (P=0.07) (Table 3). Prevalence of appropriate statin use in patients with CKD who were 50 years or older remained unchanged from 29% in 2006–2008 to 31% in 2012–2014 (P=0.92). When all medications were analyzed, including up to 10 in 2012–2013 and up to 30 in 2014, prevalence of appropriate statin use increased from 29% in 2006–2008 to 36% in 2012–2014 but was not statistically significantly different (P=0.11) (Supplemental Table 3). Surgical care had a lower prevalence of ACEi/ARB use and statin use, compared with primary and medical specialty care, which was attenuated by adjustment for demographics and comorbidities (adjusted P=0.24) (Table 4).

The prevalence of recorded NSAID use was low among visits for patients with CKD averaging 3% during the entire study period (Figure 1). Recorded NSAID use increased from 2% in 2006–2008 to 4% in 2012–2014 (P=0.01) (Table 3).

Figure 1.: Quality indicator performance in office-based visits for patients with CKD did not improve over time. ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; NSAID, nonsteroidal anti-inflammatory drug.

Discussion

Our results show that, in a national, serial cross-sectional study of outpatient visits in the United States for patients with diagnosed CKD, there were substantial and persistent gaps in quality of care from 2006 to 2014. Patients with CKD had a high prevalence of uncontrolled hypertension, which did not decrease over time. Despite strong evidence for the efficacy of ACEi/ARB treatment for patients with CKD in reducing progression to ESKD, we found that ACEi/ARB use declined over time (¹⁷,¹⁸). We found that statins are extremely underused in patients with CKD, despite guideline recommendations and evidence that statins significantly reduce cardiovascular events and mortality in patients with CKD (^19–21). Recorded NSAID use among patients with CKD increased over time.

Our results are concordant with findings from other studies that there are substantial gaps in quality of care for patients with CKD. An analysis from baseline visits of the Chronic Renal Insufficiency Cohort from 2003 to 2007 found a similar prevalence of uncontrolled hypertension among patients with CKD, whereas a previous study of the National Health and Nutrition Examination Survey (NHANES) from 1999 to 2006 reported higher rates of uncontrolled hypertension (²²,²³). The percentages of uncontrolled hypertension in patients with CKD in a national cohort of veterans with advanced CKD and in a large primary care practice in Massachusetts were similar to our findings (²⁴,²⁵). In terms of medication use, an analysis using the NHANES data from 2001 to 2010 found that the use of ACEi/ARBs in patients with CKD was 34%, which is similar to our findings, whereas a single-center study in the state of New York found ACEi/ARB use to be 59% in patients with CKD (²⁶,²⁷). Our results are also consistent with findings that prevalence of statin use in patients with CKD is far below guideline recommendations. Statin use in the NHANES and other United States studies has been shown to be suboptimal (²⁸,²⁹). However, a study of veterans with CKD and diabetes found that use of ACEi/ARB was 66% and that use of statins was 85% in those with CKD stage 3; although still suboptimal, quality of care for CKD may be higher in the veteran population (³⁰). In terms of NSAID use, the CDC CKD Surveillance System team found that 5% of patients with moderate to severe CKD were using NSAIDs from 1999 to 2004 (³¹). Our study provides contemporary national estimates showing that NSAID use in patients with CKD is still prevalent, despite avoiding NSAIDs in patients with CKD being an American Board of Internal Medicine Choosing Wisely recommendation (¹⁶).

Despite the introduction of multiple national quality reporting programs and CKD-specific guidelines during our study period, we did not see improvement in quality of CKD care over time. Of note, our study describes the quality of CKD care nationally using diagnosis codes rather than laboratory values, capturing patients who were known to have CKD by their physicians and who may have later stage CKD. Sensitivity of ICD-9 diagnosis codes for CKD is low and highly variable, and therefore, our study reflects quality of care only in patients known to have CKD by their physicians, not in all patients with CKD (^32–35). In our study, 2.7% of visits were for patients with diagnosed CKD, whereas the national CKD prevalence is >14% (¹). Our study suggests that quality of CKD care remains suboptimal due to reasons other than lack of physician awareness of CKD status (³⁶).

One reason for lack of improvement in medication use may be due to a lack of dedicated CKD-specific quality metrics. There was no metric for ACEi/ARB use in CKD in previous national quality programs, such as the Physician Quality Reporting System (³⁷). The National Quality Forum endorsed a measure from the Renal Physicians Association for ACEi/ARB use in proteinuric CKD; however, the adoption of the measure in primary care and specialty settings is unknown (³⁸). Similarly, there is no CKD-specific quality measure for lipid management endorsed by the National Quality Forum or in the Renal Physicians Association Kidney Quality Improvement Registry. Currently, the Merit-Based Incentive Payment System does not include measures for ACEi/ARB use or statin use in CKD in the Nephrology Specialty Measures Set. Another contributing factor to the lack of improvement in quality of CKD care is insufficient knowledge of CKD-specific guidelines (³⁹,⁴⁰). For example, KDIGO recommends statins in all patients with CKD age ≥50 years old, which differs from prior AHA/ACC guidelines (¹⁹,⁴¹,⁴²). Lack of knowledge of the KDIGO guideline may explain low rates of statin use among patients with CKD. In 2017, ACC/AHA guidelines for hypertension were released (¹²). Although this occurred after the study period, the AHA/ACC BP goal for patients with CKD was similar to that in JNC 7.

There are other likely reasons for suboptimal quality of care for patients with CKD. Low rates of nephrology referral, even in patients with moderate to severe CKD, may further drive decreased adherence to quality indicators, particularly those in nephrology specialty guidelines (⁴³). Quality of care performance in nephrology clinics specifically was unable to be assessed in our analysis. The prevalence rates of uncontrolled hypertension in patients with and without CKD are known to be high in the United States, and treatment of hypertension has been shown to not be concordant with guidelines (⁴⁴). Prevalence of chronic pain is high for patients with CKD, which may in part account for NSAID use in this population (⁴⁵,⁴⁶).

However, evidence-based interventions to improve the quality of CKD care have been shown to be effective. A systematic review of randomized trials of quality improvement interventions showed a decrease in progression to ESKD and increased use of ACEi/ARBs (⁴⁷). The majority of CKD is treated in primary care settings, and therefore, efforts toward improved CKD management must involve primary care physicians as a central component of multispecialty care teams (⁴⁸,⁴⁹). New initiatives, such as the National Kidney Foundation’s CKD intercept may enable improved CKD management in primary care. Significant barriers to improving quality of care in CKD remain. Primary care physicians may have skepticism that monitoring CKD improves care, which may drive worse quality (⁴¹). Additional barriers include limited time, competing demands, and difficulties obtaining and using data (³⁹). Effective management of hypertension and diabetes requires medication adherence and patient behavior change, which is difficult to achieve. In addition, population health–based interventions to improve medication prescribing or chronic disease management in CKD are currently not well supported by payment models or care delivery systems (⁵⁰).

There are several limitations in our results. The NAMCS dataset does not contain creatinine or urine albumin-to-creatinine ratio measurements to confirm CKD, which is a limitation of our study. Appropriateness of ACEi/ARB prescribing for proteinuria was unable to be fully assessed due to lack of laboratory data. Data interoperability standards in electronic medical records may allow for laboratory values to be incorporated in future datasets. BP measurement is not standardized in routine clinical practice, which may have been a source of measurement error. The NAMCS sample only represents patients routinely seen in office-based ambulatory medical care, and therefore, it may exclude patients who visit a hospital-based outpatient facility. It does not represent veterans or those in institutionalized settings. Although the NAMCS does capture over-the-counter medication usage, the observed increase in NSAID use may be reflective of improved medication documentation, particularly of over-the-counter medications. For example, recorded NSAID use was higher in primary care compared with medical specialty care, which may reflect increased documentation (Table 4). When all medications were included in the analysis, there was an increase in statin use, but this is likely an artifact of the data due to an increase in survey collection from eight to 30 medications (Supplemental Table 3). Finally, our analysis was limited to visit-based data rather than longitudinal patient data, which precluded the ability to assess relevant quality measures measured in the same person over time. Despite these limitations, this study provides a nationally representative picture of ambulatory care practice as it relates to CKD management in the United States.

In summary, in a nationally representative dataset, we found that quality of care for patients with CKD was not concordant with guidelines and did not improve over time. Our findings indicate that, even when physicians are aware of a patient’s CKD diagnosis, there are substantial gaps in quality of CKD care. The lack of improvement over a decade in quality of care in CKD highlights a more urgent need for CKD-specific quality measures and implementation of quality improvement interventions.

Disclosures

Dr. Keyhani, Dr. Powe, and Dr. Tummalapalli have nothing to disclose.

Funding

Dr. Tummalapalli was supported by research funding from the National Institute of Diabetes and Digestive and Kidney Diseases 2T32DK007219-41 Training Grant to the University of California, San Francisco.