Original Investigation Improving Care Coordination Between Nephrology and Primary Care: A Quality Improvement Initiative Using the Renal Physicians Association Toolkit William E. Haley, MD,1 Amy L. Beckrich, BA,2 Judith Sayre, PhD,3 Rebecca McNeil, PhD,4 Peter Fumo, MD,5 Vijaykumar M. Rao, MD,6 and Edgar V. Lerma, MD6 Background: Individuals at risk for chronic kidney disease (CKD), including those with diabetes mellitus and hypertension, are prevalent in primary care physician (PCP) practices. A major systemic barrier to mitigating risk of progression to kidney failure and to optimal care is failure of communication and coordination among PCPs and nephrologists. Study Design: Quality improvement. Longitudinal practice-level study of tool-based intervention in nephrology practices and their referring PCP practices. Setting & Participants: 9 PCP and 5 nephrology practices in Philadelphia and Chicago. Quality Improvement Plan: Tools from Renal Physicians Association toolkit were modified and provided for use by PCPs and nephrologists to improve identification of CKD, communication, and comanagement. Outcomes: CKD identification, referral to nephrologists, communication among PCPs and nephrologists, comanagement processes. Measurements: Pre- and postimplementation interviews, questionnaires, site visits, and monthly teleconferences were used to ascertain practice patterns, perceptions, and tool use. Interview transcripts were reviewed for themes using qualitative analysis based on grounded theory. Chart audits assessed CKD identification and referral (PCPs). Results: PCPs improved processes for CKD identification, referral to nephrologists, communication, and execution of comanagement plans. Documentation of glomerular filtration rate was increased significantly (P 5 0.01). Nephrologists improved referral and comanagement processes. PCP postintervention interviews documented increased awareness of risk factors, the need to track high-risk patients, and the importance of early referral. Final nephrologist interviews revealed heightened attention to communication and comanagement with PCPs and increased levels of satisfaction among all parties. Limitations: Nephrology practices volunteered to participate and recruit their referring PCP practices. Audit tools were developed for quality improvement assessment, but were not designed to provide statistically significant estimates. Conclusions: The use of specifically tailored tools led to enhanced awareness and identification of CKD among PCPs, increased communication between practices, and improvement in comanagement and cooperation between PCPs and nephrologists. Am J Kidney Dis. 65(1):67-79. ª 2014 by the National Kidney Foundation, Inc. INDEX WORDS: Quality improvement; chronic kidney disease (CKD); co-management; communication; primary care practice; primary care physician (PCP); nephrology practice; nephrology referral.

C

omanagement of patients with chronic kidney disease (CKD) is both desirable and necessary.1,2 A diagnosis of CKD is ascribed to an estimated 10 million US persons, with many more at risk.3,4 At-risk persons, including those with diabetes mellitus and hypertension, are highly prevalent in primary care practices. Early CKD identification allows for improved management and optimal care, which mitigates the risk of progression to advanced CKD and improves outcomes.5,6 CKD is common and associated with excess mortality.4,6-10 Despite best-practice guidelines becoming available more than a decade ago,11,12 outcomes remain poor.2,4 Previous work noting poor adherence to guidelines suggested that improving comanagement among primary care physicians (PCPs) and nephrologists offered promise.13-15 A recent study noted increased comanagement processes following the institution of estimated glomerular filtration rate Am J Kidney Dis. 2015;65(1):67-79

(GFR) reporting; however, care remained suboptimal.16 Similarly, use of electronic health record alerts in 30 PCP practices did not engender appropriate proteinuria assessment and nephrology referral in patients with moderate to advanced CKD,17 and a test

From the 1Division of Nephrology and Hypertension, Mayo Clinic, Jacksonville, FL; 2Renal Physicians Association, Rockville, MD; 3University of North Florida, Jacksonville, FL; 4Department of Pediatrics, Medical University of South Carolina, Charleston, SC; 5Delaware Valley Nephrology, Philadelphia, PA; and 6Associates in Nephrology, Chicago, IL. Received February 7, 2014. Accepted in revised form June 30, 2014. Originally published online August 30, 2014. Address correspondence to William E. Haley, MD, Division of Nephrology and Hypertension, Mayo Clinic, 4500 San Pablo Rd, Jacksonville, FL 32224. E-mail: [email protected]  2014 by the National Kidney Foundation, Inc. 0272-6386/$36.00 http://dx.doi.org/10.1053/j.ajkd.2014.06.031 67

Haley et al

of an enhanced versus standard laboratory CKD prompt was ineffective in 93 PCP practices.18 Likewise, a recent report comparing a nurse-coordinated model, which included a nephrologist in addition to the usual care provided by PCPs, revealed little impact on outcomes over 24 months.19 Conversely, educational intervention by pharmacists and nurses was shown to improve blood pressure control in patients with CKD in PCP practices,20 and 2 primary care practices demonstrated improvement in the recognition of CKD, anemia, and improper use of medications using a combination of practice enhancement assistants, academic detailing, and computer decision support.21 Lack of communication and coordination among PCPs and nephrologists has been cited as a major systemic barrier to slowing CKD progression and associated premature cardiovascular disease.22 Focus groups of patients with CKD, nephrologists, and non-nephrologists engaged in their care revealed the need for improved relationships between PCPs and nephrologists.13,22,23 Meanwhile, the large proportion of patients with CKD managed exclusively by PCPs fared poorly in terms of recommended best practices.1,14,24 Prompted by these needs, the Renal Physicians Association convened a Stakeholders Consensus Panel of PCPs, nephrologists, and other experts to explore common goals for these patients (Item S1, available as online supplementary material). The most salient concerns were general dissatisfaction with the care of patients with CKD and the referral process; proposed tasks were to improve CKD identification and appropriate referral to nephrologists, improve communication, and advance comanagement among PCPs and nephrologists. In consensus, a set of specific PCP tools was developed using Facilitated Process Improvement methodology25 to promote such improvement. Berwick26 has noted that improving health care delivery is an example of “complex, unstable, non-linear social change.” He argues for rejecting the traditional evidence paradigm as inadequate in quality improvement trials in favor of Pawson and Tilley’s27 construct: context 1 mechanisms 5 outcomes. The latter relates a program’s success to the introduction of “appropriate ideas and opportunities” and contrasts with the classic rules of inference, which favor the status quo over change.26 Berwick further asserts that changes in care processes are best measured with qualitative methods, which are better suited to providing essential information about such mechanisms and contexts.26 In keeping with the imprecise nature of the science of health care delivery, this work was constituted as a pilot learning project. At the project’s launch, we initiated a conversation between PCP and nephrology practices, aiming to engender a focus on CKD and improve communication and comanagement. We focused on the 68

functional specifications for improving CKD care— what processes are necessary and sufficient.25 The project plan provided a mechanism for resolving barriers and aimed to minimize time investment required on the part of PCPs. Our objective was to assess practice patterns in the identification of patients with CKD, communication between nephrologists and PCPs, and management of patients with CKD and to test the hypothesis that use of the specifically tailored PCP tools would result in better care as gauged by enhanced identification of patients with CKD and improvement in communication and satisfaction between nephrologists and PCPs.

METHODS The Mayo Clinic Institutional Review Board determined that this quality improvement study (ID 09-003536) was exempt, and it was approved by the clinical leadership of the practices, whose experience is reported. We investigated nephrology and primary care practices, testing the intervention of providing tools specifically created to improve CKD identification, referral, communication, and comanagement. The experiment included 9 PCP and 5 nephrology practices over 12 to 15 months and concluded on March 31, 2011. Two cities were selected from different regions of the United States (Northeast: Philadelphia; Midwest: Chicago), and nephrology practices were identified within those communities and selected based on willingness to participate in a quality improvement project and to recruit from among their primary care referring groups a number of those also willing to participate. Four referring PCP practices in Philadelphia and 5 in Chicago were recruited by respective nephrology practices. Nephrologist training and PCP education sessions were followed by selection of physician leaders and “site champions” at each participating practice. Site visits and pre- and postintervention questionnaires provided practice-level information. Additional information on practice patterns, perceptions, and tool use was obtained from taped interviews with physician leaders and site champions pre- and postimplementation and also from monthly teleconferences with study staff; audit instruments developed for quality improvement focused on CKD identification and referral for the PCP practices. Each PCP practice was responsible for identifying 20 patients for the pre- and post- audits; 7 of 9 completed both. Two of the 9 PCP practices, one each from Philadelphia and Chicago, completed only the preaudit. Both pre- and postintervention samples were selected using tools provided in the patient identification assessment and evaluation section of the Renal Physicians Association Advanced CKD Patient Management Toolkit.28 Practices screened consecutive charts of patients seen beginning on a convenient date, selecting those older than 50 years, with a diagnosis of diabetes mellitus and hypertension, and followed up by the practice for at least 12 months. Data were collected from 292 eligible patient records, 157 audited pre- and 135 audited postimplementation, and included the following information. Was a serum creatinine level ordered within the past 1 year? Is there documentation of GFR within the past 1 year? Is the GFR # 30 mL/min? Was the patient referred to a nephrologist? Also included were sex, age, race, and employment and insurance status. The stakeholders’ panel developed a consensus on the needs and objectives. Tasks derived according to methodology described by Matchar et al25 included developing tools to address specific barriers and root causes. The panel recommended modifications to tools from the Advanced CKD Patient Management Toolkit and creation of an additional tool (CKD Screening Protocol/When to Refer; Table 1) We identified barriers, which were recorded and Am J Kidney Dis. 2015;65(1):67-79

Coordination Between Nephrology and Primary Care then addressed, along with tool utilization, during the monthly teleconferences. For example, each PCP practice developed a process to route the Faxback Form within the practice and back to the nephrologist. Most PCP practices found this form to be beneficial and effective, designating it for continued use. Nephrologists’ use of the CKD Post-consult Letter tool was promoted by results of the audits of preimplementation consult letters, which demonstrated deficiencies in content and clarity of goals. Following the receipt of these letters, PCPs flagged CKD charts and implemented recommended actions in comanaged patients. Implementing the CKD Chart Flag was facilitated in one PCP practice with electronic health records by creating an electronic flag. The CKD Patient Diary, an education and comanagement tool, was used by all except one nephrologist, with barriers reportedly time constraints and perceived value. (For full tools description and instructions, access reference28.) All interviews were conducted by one study staff person (A.L.B.). Preimplementation interviews included open-ended questions about care processes for chronic disease management, guidelines and tools in use, perceptions, readiness for change, and satisfaction. Postimplementation interviews explored processes for CKD identification and referral, comanagement and goals of care, communication between nephrologist and PCP regarding referral and coordination of care, patient education, overall level of CKD care provided, tool use, barriers, overall burden of implementation, and satisfaction (Item S2). Using grounded theory,29,30 a qualitative expert (J.S.) reviewed interview transcripts for themes that emerged from the narrative. Themes were not deemed emergent unless they appeared in a majority of narratives for the respective group (PCPs, nephrologists, and site champions). Demographic characteristics of participating practices were summarized using means and percentages. PCP audit values were analyzed as percentage, 95% confidence interval (CI), and P value. CIs and P values were adjusted for practice-level clustering using generalized estimating equations. A total sample of 100 to 200 charts was estimated as the target sample size for each of the preand postimplementation times. After adjustment for intraprovider correlation,31 this minimum target sample size was calculated to provide an absolute margin of error of 6 0.1 to 0.4.32

RESULTS Demographic characteristics of participating practices were derived from site visits and preintervention questionnaires and are presented in Tables 2 and 3. Location was urban or suburban and practices were either group single specialty or solo; numbers of physicians and years in practice varied widely. Nephrologists’ first CKD visit averaged 40 minutes, with 20 minutes for follow-up, and PCPs estimated 35 minutes for first visit, with 20 minutes for follow-up. CKD management styles varied. Only one nephrology and one PCP site used electronic health records. PCP patients were younger, with, as expected, a lower percentage of CKD stages 3-5 compared with nephrology practice patients. Practice patterns in CKD identification, management, and communication were compiled from questionnaires, site visits, and monthly call worksheets. At the outset, care processes and mechanisms related to the project tasks varied among PCP practices. Most identified diabetic patients as at risk for CKD (7 of 9), but not other high-risk patients. The absence of a list of high-risk conditions was notable (7 of 9), as was the staff’s lack of Am J Kidney Dis. 2015;65(1):67-79

appreciation of that need (9 of 9). A point-of-care process for identifying patients with CKD also was lacking (9 of 9). Likewise, there were no available criteria for referral of patients to nephrologists (9 of 9). Referral processes varied; some provided laboratory test results, but most simply instructed patients to call the nephrologist for an appointment. Patient education processes were well developed in nephrology practices, but were rudimentary or nonexistent in primary care. In the 8 practices for which postimplementation PCP processes were analyzed (the physician leader and site champion for site 9 left soon after collection of the baseline data, and the practice decided not to participate further), remarkable improvement was observed for CKD identification, referral, and communication and execution of comanagement plans. A total of 166 of the 171 tasks and 124 of the 144 subtasks were in place postimplementation, compared with 78 and 51 preimplementation, respectively (Table 4). Nephrology practices likewise improved postimplementation, particularly in their referral processes and communication (Table 5). Postintervention questionnaires confirmed that patients with CKD were being referred earlier—none later than stage 4. Interviews were conducted with 12 physician leaders (7 PCPs and 5 nephrologists) and 13 respective site champions prior to implementation and 24 individuals (7 PCPs, 5 nephrologists, and 12 site champions) after implementation of the toolkit. Both sets of interviews averaged 32 minutes. Key themes are shown in Box 1, and relevant quotes are included in Box 2. Preimplementation, few PCPs reported familiarity with CKD clinical practice guidelines, and CKD screening was limited mostly to diabetic patients. Nephrologists and site champions expressed awareness of the importance of early CKD identification by PCPs. Postimplementation, all PCPs reported increased awareness of risk factors for kidney disease. Preimplementation, few PCPs used specific triggers for nephrology referrals, although several cited creatinine level . 2.0 mg/dL or when dialysis questions arose. Postimplementation transcripts revealed increased consistency of referral timing, with PCPs providing more vigilant monitoring of high-risk patients: managing CKD up to stage 3 and all reporting referral by stage 4. Included in the referral process theme, several nephrologists and their site champions noted the need for timely nephrology appointments. Postintervention nephrology interviews revealed heightened attention to communication and comanagement. PCPs reported that the project altered the content of nephrology postconsult letters, advanced comanagement goals, and improved teamwork among office staff. Increased communication between practices was associated with enhanced satisfaction scores. On a Likert scale (range, 1-5, with 5 the highest), satisfaction with comanagement 69

70 Table 1. Tools Used by PCP and Nephrology Practices

Function

Identification

Communication

Name of Tool

CKD Identification and Action Plan Poster

Intended User Relationship

Referring physicians/ clinicians/ nephrologists

Intent of Tool

Format of Tool

Notes on Use

Post in examination rooms, work Assists in identifying high-risk patients: Poster for physician’s office; includes stations; use to educate patients identifies patients with CKD and their definition of CKD, classification of CKD and staff about CKD stage, defines CKD, communicates stages and action associated with need for referral and comanagement; each stage, keys to identification of an information and reminder tool that patients at high risk for CKD, indicators for kidney damage, risk factors for summarizes essential elements of care progression, and potential complications

Referring clinicians CKD Screening Protocol/When to Refer

Assists in identifying patients with CKD and provides clarity on when CKD patients should be referred to a nephrologist for consultation and/or comanagement, including risk factors that necessitate early referral

Referring Clinician Faxback Form

Clarifies reason for referral and ensures 1-page faxback form for communication Originates from nephrology office nephrologist receives important clinical between nephrologist and referring once patient is referred; should be data; this is the communication tool clinician; can be individualized by completed by PCP, then faxed that identifies patients selected by adding practice fax header or inserting back to nephrologist office; site PCPs for comanagement and triggers clinic stamp champion needs to make sure this development of comanagement plan form is received for all referred patients

Nephrologist to referring clinician, faxed back to nephrologist

1-page document with CKD stages and risk factors

Post in work areas where high-risk patients are seen or include in charts

Outlines goals of care for anemia, bone 1-page document with goals of care disease, hypertension, nutrition, lipids, counseling, and timing

CKD Post-consult Letter

Nephrologist to referring clinician

CKD Chart Flags

Referring clinicians

Originates from nephrology office Clarifies goals, duties, responsibilities of Available in 2 formats: 1-page form once patient has been seen; PCP/ clarifying respective roles of comanagement plan based on the site champion should compare to nephrologist and referring clinician, reason for referral, if patient has been postconsult letter expectations available for both opinion-only and selected for comanagement; for document to verify it includes comanagement, and bulleted list to patients seen for opinion only, provides complete information; can be remind nephrologist while drafting his/ description of major issues and goals added to EHR her own letter to be addressed Paper charts: stickers to place on outside Develop/add EHR pop-up or template Plays unique role as identifier and of patient medical record; EHR: pop-up reminder mechanism for CKD patients with comanagement plan; serves as or CKD office template reminder to address specific duties for patients with advanced CKD (Continued)

Originates from nephrology office, may be included with postconsult letter; may also be posted for reference

Haley et al

Am J Kidney Dis. 2015;65(1):67-79

Concise Guidelines Nephrologist to referring clinician

Am J Kidney Dis. 2015;65(1):67-79

Abbreviations: CKD, chronic kidney disease; EHR, electronic health record; PCP, primary care physician; RPA, Renal Physicians Association.

Originates with nephrology office For educating patient about CKD goals, Card with patient and provider names (although also can be originated by and so patient knows who is treating and a series of diagrams shaded to PCP); should be brought in by reflect the patient’s degree of kidney what; diary involves patient in ongoing patients once they have seen care, is reminder of comanagement function and CKD stage; inside the nephrologist; someone in PCP’s responsibilities and serves as a card is a flow sheet similar to that in the office will need to assist patients in patient-initiated physician reminder patient chart; also has lay version of filling it out goals of care and recommendations from the RPA guideline and handy medication list CKD Patient Nephrologist to patient Education Patient Diary Patient Education

Format of Tool Intent of Tool Function

Name of Tool

Intended User Relationship

Table 1 (Cont’d). Tools Used by PCP and Nephrology Practices

Notes on Use

Coordination Between Nephrology and Primary Care

reported by nephrologists improved from 2.6 preimplementation to 4.3 postimplementation. Corresponding satisfaction levels of PCPs were relatively high preimplementation (4.3) and increased to 4.7. Of the 16 respondents with preimplementation levels less than “satisfied,” 15 noted improvement, with 6 improving from “somewhat unsatisfied” to “satisfied” or “very satisfied.” Audits were completed for 292 PCP charts, yielding 157 pre- and 135 postimplementation. Demographics of these randomly selected pre- and postimplementation samples were similar (age, race, sex, and employed and insured status). Audits were designed to provide feedback on opportunity for improvement, not a statistical sample. Analysis of PCP audits with complete pre- and postimplementation data revealed significant improvement in GFR documentation (preimplementation, 82% [95% CI, 53%-95%]; postimplementation, 99% [95% CI, 95%-99%]; P 5 0.01); most performed well with respect to ordering creatinine levels within 1 year for these high-risk patients preimplementation and all postimplementation (97% [95% CI, 89%-95%] and 99% [95% CI, 96%-99%], respectively; P 5 0.2). The overall rate of referral of patients with CKD was 24% (95% CI, 12%-41%) preimplementation and 39% (95% CI, 19%-63%) postimplementation, not significantly different (P 5 0.4). However, 4 of 7 showed an average 45% increase (95% CI, 20%95%) and a fifth, with no net increase in percentage of patients referred, nevertheless showed a 100% increase in referral of patients with CKD stage 4. The rate of patients with GFRs # 30 mL/min who were referred to nephrology was higher than the overall rate of referral of patients with CKD and did not increase significantly (50% [95% CI, 2%-79%] pre- versus 58% [95% CI, 24%-85%] postimplementation; P 5 0.7), whereas the percentage of those with GFRs # 30 mL/min was similar pre- and postimplementation (10% [95% CI, 3%-30%] vs 24% [95% CI, 8%-55%]; P 5 0.3). Although there were too few practices to support a formal comparison, those consistently using all tools performed better in terms of achieving project goals or improvements than those that did not. Site champions and physician leaders were contacted 3 years after completion of the study and asked specifically whether the improvements in awareness of CKD, communication, the referral process, and comanagement of patients with CKD had been sustained and whether tools were still in use. Of those who had remained active in their respective practices over that time frame, 5 responded (3 site champions and 2 physicians), with 14 of 15 answers being affirmative.

DISCUSSION In this study, the use of specifically tailored tools was associated with enhanced awareness and identification 71

Haley et al Table 2. Demographics and Characteristics of Participating Nephrology Sites Characteristic

Location Setting Academic affiliation Ownership CKD clinic Time spent with CKD patients First visit (min) Follow-up (min) Management of CKD patients (stages) Site champion

Site 1

Site 2

Site 3

Site 4

Site 5

Urban

Urban

Suburban

Urban

Suburban

Group single specialty

Solo

Group single specialty

Solo

Solo

No

No

Yes

No

No

Independent Yes

Independent Yes

Independent Yes

Independent Yes

Independent Yes

60 20-30

30-45 15

30 15-20

30 15

30 15

Co- (1-5)

Co- (1-5)

Co- (1-5)

Co- (1-3); Full (4-5)

Co- (1-5)

Nurse

Office manager

Physician leader

Physician leader

Nurse

Use of information systems Scheduling and billing Laboratory values Medication ordering Clinic notes/EHR

Yes Yes Yes Yes

Yes No No No

Yes No No No

Yes No No No

Yes No No No

Provider information No. of physicians No. of allied health Female sex Nonwhite ethnicity Mean no. of years in practice

6 0 0% 17% 17

1 1 0% 100% 30

7 2 57% 86% 13

1 2 0% 0% 5

1 1 0% 100% 7

Patient features Median age (y) Female sex Nonwhite ethnicity CKD stage 3 CKD stage 4 CKD stage 5

68 50% 82% 40% 20% 3%

60 60% 80% 85% 10% 5%

75 60% 10% 68% 28% 6%

50 50% 40% 60% 30% 10%

56 44% 85% 48% 20% 6%

Payor mix Managed care Indemnity Medicare Medicaid Fee for service Other

45% 12% 38% 2% 0% 3%

25% 0% 70% 5% 0% 0%

50% 0% 45% 4% 0% 1%

30% 0% 30% 10% 30% 0%

8% 0% 60% 30% 2% 0%

Abbreviations: Allied health, physician assistant, educator, nutritionist; CKD, chronic kidney disease; Co-, comanagement; EHR, electronic health record; Full, full management.

of CKD among PCPs and led to increased communication and improvement in comanagement and cooperation between PCPs and nephrologists. The PCP practices improved CKD identification, referral to nephrologists, communication, and execution of comanagement plans. Nephrologists improved referral and comanagement processes. Postintervention interviews of PCPs documented increased awareness of risk factors, the need to track high-risk patients, and the importance of early referral. Final nephrologist interviews revealed heightened attention to communication and comanagement with PCPs, along with increased levels of satisfaction among all parties. Previous work has linked poor guideline adherence to a lack of or poor comanagement between PCPs and nephrologists, even after institution of automatic 72

estimated GFR reporting. Preintervention, nephrologists noted that patients generally did not know the reason for referral and that test duplication was common. They asserted the need for better awareness of CKD and earlier referral. There was not a comanagement process in place in most of the PCP or nephrology practices. Investigation of preferences among PCPs and nephrologists in the care of patients with CKD revealed that while most favored collaboration, preferences differed; PCPs preferred to seek guidance on diagnosis, testing, medication use, and nutrition.33 Studies of PCP-specialist collaboration in diabetes care have noted that good relationships, insurance coverage, and continuity of care with PCPs were important, and that good communication was associated with improved outcomes.34,35 However, a Am J Kidney Dis. 2015;65(1):67-79

Characteristic

Site 1

Site 2

Site 3

Site 4

Site 5

Site 6

Site 7

Site 8

Site 9

Location

Urban

Urban

Urban

Suburban

Suburban

Suburban

Urban

Suburban

Urban

Setting

Group single specialty

Group single specialty

Group single specialty

Solo

Group single specialty

Group single specialty

Group single specialty

Solo

Group single specialty

Academic affiliation

Yes

No

Yes

Yes

Yes

No

Yes

Yes

Yes

Time spent with CKD patients First visit (min) Follow-up (min)

25-30 20

15 15

20 5

60 30

30 15

30-45 20

25-30 20

45 30

NA NA

Management of CKD patients (stages)

Full (1-3); Co- (4-5)

Co- (1-5)

Op (1-3); Co- (4-5)

Full (1-5)

Full (1-3); Co- (4-5)

Co- (1-5)

Co- (1-5)

Co- (1-5)

Co- (1-5)

Site champion

PA

Medical assistant

Office staff

Medical resident

Medical resident

Office staff

Medical resident

Medical assistant

Nurse

Use of information systems Scheduling and billing Laboratory values Medication ordering Clinic notes (EHR)

Yes Yes Yes No

Yes Yes Yes Yes

Yes Yes No No

Yes Yes No No

Yes No No No

Yes No No No

Yes Yes No No

No Yes Yes No

Yes No No No

Provider information No. of physicians No. of allied health Female sex Nonwhite ethnicity Mean no. of years in practice

3 1 67% 34% 28

4 0 50% 25% 18

3 0 0% 0% 20

1 0 0% 100% 20

2 0 50% 100% 20

3 0 0% 100% 28

28 0 57% 32% 5

1 0 100% 100% 27

3 0 0% 0% NA

Patient features Median age (y) Female sex Nonwhite ethnicity CKD stage 3 CKD stage 4 CKD stage 5

51 60% 85% 8% 3% 1%

45 55% 60% 5% 4% 1%

50 50% 90% 20% 5% 2%

60-70 40% 30% 10% 2% 1%

55 60% 30% 20% 20% 10%

401 50% 60% NA NA NA

45 60% 45% 10% 5% 1%

.30 60% 100% 15% 10% 5%

61 41% 82% NA NA NA

Payor mix Managed care Indemnity Medicare Medicaid Fee for service Other

25% 0% 30% 40% 0% 5%

55% 25% 10% 0% 0% 5%

20% 10% 30% 10% 30% 0%

5% 40% 50% 0% 0% 5%

20% 5% 65% 5% 0% 0%

20% 40% 40% 0% 0% 0%

50% 0% 5% 15% 15% 15%

28% 0% 70% 0% 0% 2%

NA NA NA NA NA NA

73

Abbreviations and definitions: Allied health, physician assistant, educator, nutritionist; CKD, chronic kidney disease; Co-, comanagement; EHR, electronic health record; Full, full management; NA, data not available; Op, nephrology opinion only; PA, physician’s assistant.

Coordination Between Nephrology and Primary Care

Am J Kidney Dis. 2015;65(1):67-79

Table 3. Demographics and Characteristics of Participating Primary Care Sites

Haley et al

Table 4. PCP Practice Patterns in CKD Identification, Management, and Communication Postimplementation Process in Placea

Baseline Process in Place

Site Site Site Site Site Site Site Site Site Site Site Site Site Site Site Site Site 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8

Project Tasks and Subtasks

Process for CKD Identification O X

O O

O X

O X

O O

O X

X X

—

X

O X

O O

O O

O X

O O

O O

O O

O O

O O

X O

X O

X X

X O

X O

X O

X O

X X

X O

O O

O O

X O

X O

O O

O O

O O

—

O O

O O

X X

O O

O O

O O

O O

X X

X X

O O

O O

X O

O O

O O

O O

O O

X O

2. Creatinine/GFR and urine (ACR) ordered O for each high-risk pt Staff understand need and what intervals X Someone is responsible for ordering these O lab tests

O

O

O

O

O

O

X

O

O

O

O

O

O

O

O

O

X O

X O

X O

O O

O O

O O

X O

X O

O O

O O

X O

X X

O O

O O

O O

O O

O

O

X

O

O

O

O

X

O

O

O

O

O

O

O

O

O

X

X

X

X

X

X

X

X

X

O

O

X

X

O

O

O

O

X

X

X

X

X

X

X

X

X

O

O

X

X

O

O

O

O

X

O

X

X

O

O

O

X

X

O

O

O

O

O

O

O

O

1. High-risk pts identified Accessible list of high-risk conditions available Staff understand need and means Mechanism exists to coordinate with existing care Staff has reminder mechanism Someone is responsible to flag these pts for screening

3. CKD pts identified and documented in their medical record Staff understand need to identify pts with CKD and how this is done (CKD definition available) Staff has means to identify pts with CKD at point of care and a place for this to be documented Someone is responsible for adding to problem list or flagging chart as reminder

O

Process for Referral to Nephrologist O O

O O

O X

O X

O X

O X

O X

O X

O X

O O

O O

O X

O X

O O

O O

O O

O O

X O

X O

X O

X O

X O

X O

X O

— O

X O

O O

O O

X O

X O

O O

O O

O O

O O

2. Nephrologist with CKD interest and expertise selected

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

3. Reason for referral and pertinent test results are communicated to nephrologist

X

O

X

O

O

O

O

X

O

O

O

O

O

O

O

O

O

4. Someone identified as responsible to initiate the referral and communicate to the pt

O

O

O

O

O

O

O

—

O

O

O

O

O

O

O

O

O

1. Pts requiring referral are identified Staff have clear understanding of who to refer A definition/list of whom to refer is available Staff understand means by which referral will be accomplished

Communication of Comanagement Plan 1. Nephrologist and PCP staff understand X need for comanagement and the means by which this will be done

O

X

O

O

X

O

X

X

O

O

O

O

O

O

O

O

2. CKD comanagement plan is developed by X nephrologist for all pts referred for this purpose

X

X

O

O

X

X

X

X

O

O

O

O

O

O

O

O

3. Comanagement plan communicated to PCP with selected goals and targets and suggested division of duties; PCP and nephrologist agree on plan

X

X

O

O

X

X

X

X

O

O

O

O

O

O

O

O

X

(Continued)

74

Am J Kidney Dis. 2015;65(1):67-79

Coordination Between Nephrology and Primary Care Table 4 (Cont’d). PCP Practice Patterns in CKD Identification, Management, and Communication Postimplementation Process in Placea

Baseline Process in Place

Project Tasks and Subtasks

Site Site Site Site Site Site Site Site Site Site Site Site Site Site Site Site Site 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8

4. Mechanism for coordinating care in place X that includes easy data sharing and communication of changes to medications and other instructions given to the pt

X

X

O

X

X

X

X

X

O

O

O

O

O

O

O

O

5. Staff has means to identify pts with comanagement plan in place at point of care and this is documented in medical record

X

X

X

X

X

X

X

X

X

O

O

X

O

O

O

O

—

6. Someone is responsible for adding to X problem list or flagging chart as reminder

X

X

X

X

X

X

X

O

O

O

X

O

O

O

O

—

7. Referred pt is aware of comanagement plan and who is responsible for what Mechanism exists for informing and reminding pts with comanagement plans; staff has means to identify these pts and communicate their plan

X

X

X

X

X

X

X

X

X

O

O

X

O

O

O

O

O

X

X

X

X

X

X

X

X

X

X

O

X

O

O

O

O

—

Execution of Comanagement Plan 1. Staff are aware that the pt has been referred, is being co-managed, and has a co-management plan; there exists a reminder mechanism

X

X

X

X

X

X

X

X

X

O

O

X

O

O

O

O

—

2. Relevant lab tests and parameters are ordered/tracked/shared Staff aware of what testing required, guidelines for optimal care, what should be done at each CKD stage; clarity on who will do what; mechanism exists for tracking; reminder mechanism in place

O

X

X

O

O

X

O

—

O

O

O

O

O

O

O

O

O

X

X

X

X

X

X

X

X

X

O

O

X

X

O

O

O

—

3. Treatment is adjusted based on agreed upon goals and targets

O

O

X

O

O

X

O

X

X

O

O

O

O

O

O

O

O

4. Pt is educated regarding disease stage, capabilities, and preferences Staff possess knowledge, skills and tools for teaching pts about CKD and comorbid conditions and importance of treatment and follow-up

X

O

X

O

O

O

O

O

X

O

O

O

O

O

O

O

O

X

X

X

O

X

X

X

X

X

O

O

X

X

O

O

O

—

X

O

X

O

X

X

X

—

X

O

O

X

O

O

O

O

O

5. Pt is informed re grading test results, changes in treatment, progress, and comanaging practitioner is kept current

Abbreviations and definitions: O, process in place; —, no data available; ACR, albumin-creatinine ratio; CKD, chronic kidney disease; GFR, glomerular filtration rate; PCP, primary care physician; pt, patient; lab, laboratory; X, process not in place. a There are no postimplementation data for site 9. The physician leader and site champion for site 9 left soon after collection of the baseline data, and the practice decided not to participate further.

recent report noted disconnects between PCPs and specialists on referral, consultation communication, and the need for improving information transfer.36 Late referral and suboptimal care resulting in higher mortality and hospitalization rates and the need for early recognition among high-risk populations have been the subjects of numerous reports,2,5,9,14,24,34,37-45 and the lack of knowledge about appropriate timing of referrals and poor communication have been cited as Am J Kidney Dis. 2015;65(1):67-79

key factors.46 Uncertainty in referral timing, variability in CKD diagnosis and treatment, and perceptions that the referral process is “disorganized” all contribute to poor communication and comanagement.13,43 The tools implemented during this project (Table 1) were designed to address these concerns. Comanagement hinges on effective communication, and the Post-consult Letter Tool served as a reminder of clinical issues to address in CKD. This 75

Haley et al Table 5. Nephrology Practice Patterns in CKD Management and Communication Postimplementation Process in Place

Baseline Process in Place

Site Site Site Site Site Site Site Site Site Site 1 2 3 4 5 1 2 3 4 5

Project Tasks and Subtasks

Referral Process 1. Reason for referral and pertinent test results are communicated to nephrologist

X

X

X

X

X

O

O

O

O

O

Communication of Comanagement Plan 1. Nephrology and PCP staff understand need for comanagement and the means by which this will be done

O

O

O

O

O

O

O

O

O

O

2. CKD comanagement plan is developed by nephrologist on all pts referred for this purpose

X

X

X

X

X

O

O

O

O

O

3. Comanagement plan is communicated to PCP with selected goals and targets and suggested division of duties; PCP and nephrologist agree on plan

X

X

O

O

O

O

O

O

O

O

4. Mechanism for coordinating care in place that includes easy data sharing and communication of changes to medications, other instructions given to the pt

X

X

X

X

X

O

O

O

O

O

5. Staff has means to identify pts with comanagement plan in place at point of care and this is documented in medical record

O

O

X

X

O

O

O

O

O

O

6. Someone is responsible for adding to problem list or flagging chart as reminder

O

O

O

O

O

O

O

O

O

O

7. The referred pt is aware of comanagement plan and who is responsible for what Mechanism exists for informing and reminding pts with comanagement plans; staff has means to identify these pts and communicate their plan

X

X

X

X

X

O

O

O

O

O

X

X

X

X

X

O

O

O

O

O

Execution of Comanagement Plan 1. Staff are aware that the pt has been referred, is being comanaged, and has a comanagement plan; there exists a reminder mechanism

O

O

X

X

O

O

O

O

O

O

2. Relevant laboratory tests and parameters are ordered, tracked, and shared Staff aware of what testing required, guidelines for optimal care, what should be done at each CKD stage; clarity on who will do what; mechanism exists for tracking; staff has reminder mechanism in place

O O

O O

O O

O O

O O

O O

O O

O O

O O

O O

3. Treatment is adjusted based on agreed upon goals and targets

X

X

X

O

O

O

O

O

O

O

4. Pt is educated regarding disease stage, capabilities, and preferences Staff possess knowledge, skills, and tools for teaching pts about CKD and comorbid conditions and importance of treatment and follow-up

O O

O O

O O

O O

O O

O O

O O

O O

O O

O O

5. Pt is informed regarding test results, changes in treatment, progress, and comanaging practitioner is kept current

O

O

O

O

O

O

O

O

O

O

Abbreviations and definitions: O, process in place; CKD, chronic kidney disease; PCP, primary care physician; pts, patients; X, process not in place.

tool facilitated explicit communication of a comanagement plan. The PCPs and their site champions reported improved satisfaction with communication, noting that letters were more descriptive and collaborative and were conducive to improved communication, enhanced comanagement, and better patient care. Postimplementation, processes improved in PCP practices regarding CKD identification, referral to nephrologists, communication, and execution of comanagement plans (Table 4). Aggregate comparisons between pre- and postimplementation PCP chart audit samples suggested improvement in PCP documentation of GFR and referral to nephrologists over the course of the study. Likewise, nephrology 76

practices improved processes, particularly those related to referral and communication of a comanagement plan (Table 5). The engagement and enthusiasm of site champions and physician leaders were critical to successful implementation of tools and changes in practice.2,29 Notably, the perceived burden of implementation and participation was found to be light; both nephrologists and PCPs gave favorable rankings (nephrologists, 4.5; PCPs, 4.7 [Likert satisfaction scale, with 5 being highest]). Criteria for qualitative research include credibility, transferability, reliability, and confirmability.30 In this study, triangulations through in-depth transcript analysis, persistent observation, and statistical methodology Am J Kidney Dis. 2015;65(1):67-79

Coordination Between Nephrology and Primary Care

Box 1. Key Themes Derived From Formal Analysis of Individual Interview Transcripts 1. Enhanced awareness of CKD 2. Increased and improved communication and comanagement including referral processes among PCP, nephrologist, and respective staff 3. Increased awareness of CKD guideline recommendations, with resultant changes in care and referral patterns 4. Individual variations in office practice, barriers, and use of communication tools 5. Improved satisfaction Abbreviations: CKD, chronic kidney disease; PCP, primary care physician.

supported these criteria. In describing key themes from the interviews, an attempt was made to identify how the participants understood, regarded, and responded to project tasks. Final interviews of PCPs and site champions documented increased awareness of risk factors, a need to track high-risk patients for early identification of CKD, and the importance of early referral of certain Box 2. Selected Nephrology and Primary Care Quotations Relevant to Project Goals PCP: “Generally more aware of identifying those patients and paying a little extra attention to them; staying on top of their renal function and maybe more aggressively screening people, particularly identifying those high risk patients.” PCP (site champion): “To me, the key is the communication between nephrologists and primary care and setting up a system that works for that primary care practice to liaise with the nephrologist’s office.faxback and post-consult letter working really well.” PCP: “Consult letters have been excellent, very clear and very useful as far as clarifying what we need to do and giving me a better understanding of the patient’s current status.” Nephrologist: “I think it clearly defined the areas where each person needed to concentrate.co-management was clear. [PCPs] also understood we’re not taking over the patient.” PCP (site champion): “It clarified everything and gave us a standardized way rather than every physician doing their own independent thing. I think it’s the fine details of the co-management of CKD. It’s having those tools and having the diary and the post-consult letters with a co-management plan articulated in the letter clearly. That helps me personally feel more confident about co-managing these patients.” Nephrologist: “It did two things. It forced us to interact more with the primaries. Some wanted formal teaching and they were very open to that. It made us aware that we have to be more explicit in our consult about the goals of care and targets. It made us more cognizant about making sure the responsibilities are laid down. I think we did a better job, as a result of this project, in patient education.” PCP: “It helped solidify our roles and the expectations of each party.it helped empower me to make more decisions about the care and what I would be able to manage on my own and what I should send back. I wish there was something like this for other specialties as well.” Abbreviations: CKD, chronic kidney disease; PCP, primary care physician. Am J Kidney Dis. 2015;65(1):67-79

patients. Postimplementation nephrologist interviews revealed a heightened sense of importance of communication and comanagement with PCPs, along with increased levels of satisfaction among all parties. We acknowledge limitations to our analysis. It is not surprising that this intervention involving multiple elements yielded improvements in identification of CKD and communication and comanagement, and it is possible that some amelioration could have occurred because of factors other than the intervention. Further, while our qualitative observations could represent true improvement in these areas, it also is possible that familiarity of the respondent and questioner over the course of the project might tend to result in response skewed to the positive side. In addition, as previously noted, there were too few practices to support a formal comparison regarding tool use. Our conclusions are based on data from 5 nephrology and 9 PCP practices in urban or suburban communities within 2 large cities that differ in practice environment, provider, and patient characteristics. A larger sample would have allowed additional comparisons, including practice settings, number of patients with CKD in the practice, use of electronic health records, and allied health care professionals. Another limitation is selection bias. Nephrology practices volunteered to participate and helped recruit their referring PCP practices. A volunteer bias is unavoidable in this type of study. Generalizability of these findings may be affected in so far as volunteering practices may be more motivated and more ready to make changes in processes and may be better staffed for such. This was devised as a learning project and we asked participants to share experiences and lessons learned during monthly contact between individual sites and study staff, and also at a midproject all-site teleconference. Given the nature and purpose of this study, we do not consider this an untoward contamination; rather, it appeared to strengthen the implementation effort. Bringing nephrologists and their primary care colleagues together and providing specifically tailored tools led to enhanced awareness and identification of CKD, increased and enhanced communication, and improved comanagement processes and cooperation between PCPs and nephrologists. Our methods and findings are largely qualitative and the audits were designed solely to provide feedback on opportunities for improvement. As noted, qualitative data may be better suited to the task of measuring changes in care processes than the traditional evidence paradigm.26 Further study designed to determine whether these changes in processes yield improved outcomes may be warranted. Current developments in the use of electronic health records, such as use of alerts built into electronic ordering and 77

Haley et al

best practice decision support alerts, may provide for some momentum in this direction. As outcome-based payment continues to gain traction, tools, including electronic versions, that affect clinical outcomes by virtue of comanaged care may be worthy of review and dissemination.

ACKNOWLEDGEMENTS This work was presented as an abstract at the American Society of Nephrology’s Annual Meeting in San Diego, CA, October 30 to November 4, 2012. The authors thank Rebecca J. Schmidt, DO, for critical review of the manuscript. Support: This study was funded by the Renal Physicians Association from an unrestricted education grant from Abbott Laboratories. Abbott Laboratories had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript. Financial Disclosure: The authors declare that they have no other relevant financial interests. Contributions: Research idea and study design: WEH; data acquisition: ALB, PF, VMR, EVL; data analysis/interpretation: WEH, JS, RM; statistical analysis: RM; supervision or mentorship: WEH. Each author contributed important intellectual content during manuscript drafting or revision and accepts accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of the work are appropriately investigated and resolved. WEH takes responsibility that this study has been reported honestly, accurately, and transparently; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained.

SUPPLEMENTARY MATERIAL Item S1: Stakeholders panel. Item S2: Interview questions, preimplementation and postimplementation. Note: The supplementary material accompanying this article (http://dx.doi.org/10.1053/j.ajkd.2014.06.031) is available at www. ajkd.org.

REFERENCES 1. Blantz RC. Handing out grades for care in chronic kidney disease: nephrologists versus non-nephrologists. Clin J Am Soc Nephrol. 2007;2(2):193-195. 2. Narva AS. Decision support and CKD: not there yet. Clin J Am Soc Nephrol. 2012;7(4):525-526. 3. Coresh J, Selvin E, Stevens LA, et al. Prevalence of chronic kidney disease in the United States. JAMA. 2007;298(17):2038-2047. 4. Collins AJ, Foley RN, Chavers B, et al. US Renal Data System 2013 annual data report. Am J Kidney Dis. 2014;63(1)(suppl 1): e1-e420. 5. Chan MR, Dall AT, Fletcher KE, et al. Outcomes in patients with chronic kidney disease referred late to nephrologists: a metaanalysis. Am J Med. 2007;120(12):1063-1070. 6. Komenda P, Levin A. Analysis of cardiovascular disease and kidney outcomes in multidisciplinary chronic kidney disease clinics: complex disease requires complex care models. Curr Opin Nephrol Hypertens. 2006;15(1):61-66. 7. Weiner DE, Tabatabai S, Tighiouart H, et al. Cardiovascular outcomes and all-cause mortality: exploring the interaction between CKD and cardiovascular disease. Am J Kidney Dis. 2006;48(3):392-401. 78

8. Go AS, Chertow GM, Fan D, et al. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med. 2004;351(13):1296-1305. 9. Nissenson AR, Collins AJ, Hurley J, et al. Opportunities for improving the care of patients with chronic renal insufficiency: current practice patterns. J Am Soc Nephrol. 2001;12(18):17131720. 10. Obrador GT, Ruthazer R, Arora P, et al. Prevalence of and factors associated with suboptimal care before initiation of dialysis in the United States. J Am Soc Nephrol. 1999;10(8):1793-1800. 11. National Kidney Foundation. K/DOQI Clinical Practice Guidelines for Chronic Kidney Disease: evaluation, classification, and stratification. Am J Kidney Dis. 2002;39(2)(suppl 1):S1-S246. 12. Bolton WK. Renal Physicians Association clinical practice guideline: appropriate patient preparation for renal replacement therapy: guideline number 3. J Am Soc Nephrol. 2003;14(5):14061410. 13. Patwardhan MB, Matchar DB, Haley WE, et al. Nonnephrologist and nephrologist perspectives on implementing a chronic kidney disease guideline. J Healthc Qual Web Exclusives. 2007;29(6):W6. 3-17. 14. Patwardhan MB, Samsa GP, Matchar DB, et al. Advanced chronic kidney disease practice patterns among nephrologists and non-nephrologists: a database analysis. Clin J Am Soc Nephrol. 2007;2(2):277-283. 15. Patwardhan MB, Matchar DB, Samsa GP, et al. Utility of the advanced chronic kidney disease patient management tools: case studies. Am J Med Qual. 2008;23(2):105-114. 16. Abdel-Kader K, Fischer GS, Johnston JR, et al. Characterizing pre-dialysis care in the era of eGFR reporting: a cohort study. BMC Nephrol. 2011;12:12. 17. Abdel-Kader K, Fischer GS, Li J, et al. Automated clinical reminders for primary care providers in the care of CKD: a small cluster-randomized controlled trial. Am J Kidney Dis. 2011;58(6): 894-902. 18. Manns B, Tonelli M, Culleton B, et al. A cluster randomized trial of an enhanced eGFR prompt in chronic kidney disease. Clin J Am Soc Nephrol. 2012;7(4):565-572. 19. Barrett BJ, Garg AX, Goeree R, et al. A nurse-coordinated model of care versus usual care for stage 3/4 chronic kidney disease in the community: a randomized controlled trial. Clin J Am Soc Nephrol. 2011;6(6):1241-1247. 20. Gallagher H, de Lusignan S, Harris K, et al. Qualityimprovement strategies for the management of hypertension in chronic kidney disease in primary care: a systematic review. Br J Gen Pract. 2010;60(575):e258-e265. 21. Fox CH, Swanson A, Kahn LS, et al. Improving chronic kidney disease care in primary care practices: an upstate New York Practice-based Research Network (UNYNET) study. J Am Board Fam Med. 2008;21(6):522-530. 22. Rettig RA, Norris K, Nissenson AR. Chronic kidney disease in the United States: a public policy imperative. Clin J Am Soc Nephrol. 2008;3(6):1902-1910. 23. Parker TF, III, Blantz R, Hostetter T, et al. The chronic kidney disease initiative. J Am Soc Nephrol. 2004;15(3):708-716. 24. Patwardhan MB, Matchar DB, Samsa GP, et al. Opportunities for improving management of advanced chronic kidney disease. Am J Med Qual. 2008;23(3):184-192. 25. Matchar DB, Patwardhan MB, Samsa GP, et al. Facilitated process improvement: an approach to the seamless linkage between evidence and practice in CKD. Am J Kidney Dis. 2006;47(3): 528-538. 26. Berwick DM. The science of improvement. JAMA. 2008;299(10):1182-1184. Am J Kidney Dis. 2015;65(1):67-79

Coordination Between Nephrology and Primary Care 27. Pawson R, Tilley N. Realistic Evaluation. Thousand Oaks, CA: Sage Publications Ltd; 1997. 28. Renal Physicians Association. Advanced CKD Patient Management Toolkit, table of contents. http://www.renalmd.org/ TOC---Advanced-CKD-Patient-Management-Toolkit/. Accessed June 20, 2014. 29. Glaser BG, Strauss AL. The Discovery of Grounded Theory: Strategies for Qualitative Research. Chicago, IL: Aldine; 1967:78. 30. Denzin NK, Lincoln YS. Introduction: the discipline and practice of qualitative research. In: Denzin NK, Lincoln YS, eds. Handbook of Qualitative Research. 2nd ed. Thousand Oaks, CA: Sage Publications; 2000:21. 31. Hanley JA, Negassa A, Edwardes MD, et al. Statistical analysis of correlated data using generalized estimating equations: an orientation. Am J Epidemiol. 2003;157(4):364-375. 32. Lenth RV. (2006-9). Java Applets for Power and Sample Size. 2006. http://www.stat.uiowa.edu/wrlenth/Power/. Accessed June 20, 2014. 33. Diamantidis CJ, Powe NR, Jaar BG, et al. Primary carespecialist collaboration in the care of patients with chronic kidney disease. Clin J Am Soc Nephrol. 2011;6(2):334-343. 34. Kinchen KS, Cooper LA, Levine D, et al. Referral of patients to specialists: factors affecting choice of specialist by primary care physicians. Ann Fam Med. 2004;2(3):245-252. 35. Foy R, Hempel S, Rubenstein L, et al. Meta-analysis: effect of interactive communication between collaborating primary care physicians and specialists. Ann Intern Med. 2010;152(4):247-258. 36. O’Malley AS, Reschovsky JD. Referral and consultation communication between primary care and specialist physicians: finding common ground. Arch Intern Med. 2011;171(1):56-65. 37. Allen AS, Forman JP, Orav EJ, et al. Primary care management of chronic kidney disease. J Gen Intern Med. 2011;26(4): 386-392.

Am J Kidney Dis. 2015;65(1):67-79

38. Charles RF, Powe NR, Jaar BG, et al. Clinical testing patterns and cost implications of variation in the evaluation of CKD among US physicians. Am J Kidney Dis. 2009;54(2): 227-237. 39. Lea JP, McClellan WM, Melcher C, et al. CKD risk factors reported by primary care physicians: do guidelines make a difference? Am J Kidney Dis. 2006;47(1):72-77. 40. Boulware LE, Troll MU, Jaar BG, et al. Identification and referral of patients with progressive CKD: a national study. Am J Kidney Dis. 2006;48(2):192-204. 41. Israni RK, Shea JA, Joffe MM, et al. Physician characteristics and knowledge of CKD management. Am J Kidney Dis. 2009;54(2):238-247. 42. Narva AS, Briggs M. The National Kidney Disease Education Program: improving understanding, detection, and management of CKD. Am J Kidney Dis. 2009;53(suppl 3): S115-S120. 43. Fox CH, Brooks A, Zayas LE, et al. Primary care physicians’ knowledge and practice patterns in the treatment of chronic kidney disease: an Upstate New York Practice-based Research Network (UNYNET) study. J Am Board Fam Med. 2006;19(1): 54-61. 44. Herget-Rosenthal S, Quellmann T, Linden C, et al. How does late nephrological co-management impact chronic kidney disease? An observational study. Int J Clin Pract. 2010;64(3): 1784-1792. 45. Winkelmayer WC, Owen WF, Jr, Levin R, et al. A propensity analysis of late versus early nephrologist referral and mortality on dialysis. J Am Soc Nephrol. 2003;14(2): 486-492. 46. Navaneethan SD, Aloudat S, Singh S. A systematic review of patient and health system characteristics associated with late referral in chronic kidney disease. BMC Nephrol. 2008; 9:3.

79