National data are obtained from the US Renal Data System (USRDS) from 2008 to 2016. P value comparing KPNC data between years is calculated using Cochrane-Armitage tests for trend.
aPD education programs developed and implemented for nephrologists, renal nurses, and patients and/or families.
bMost successful PD programs share best practices across nephrology leadership and medical center teams.
cUrgent-start PD programs launched; surgeons start attending Kaiser Permanente PD University.
dMedical group and health plan leadership visit each medical center to evaluate urgent-start PD program. Education is provided to renal case managers.
eExpanded PD training for surgeons; interventional radiologists attend Kaiser Permanente PD University.
fUpdated regional and medical center–level quality reporting metrics.
g2014-2017 National shortage of PD solution.
Rates are directly standardized to the 2008 population undergoing PD. P values comparing data over time are calculated using Cochrane-Armitage tests for trend.
Rates are directly standardized to the 2008 populations undergoing peritoneal dialysis (PD) and hemodialysis. P values comparing data over time are calculated using Cochrane-Armitage tests for trend.
eAppendix. Peritoneal Dialysis Playbook
Customize your JAMA Network experience by selecting one or more topics from the list below.
Pravoverov LV, Zheng S, Parikh R, et al. Trends Associated With Large-scale Expansion of Peritoneal Dialysis Within an Integrated Care Delivery Model. JAMA Intern Med. 2019;179(11):1537–1542. doi:10.1001/jamainternmed.2019.3155
What are the feasibility and outcomes of a system-level approach to expand the use of peritoneal dialysis as the preferred initial modality for advanced kidney disease?
This cohort study of 13 500 eligible patients describes a large-scale program implemented within an integrated health care delivery system that was followed by an increase in the proportion of new peritoneal dialysis initiation from 15.2% to 33.8% among patients starting dialysis during an 11-year period that was substantially higher than national rates. In addition, 331 of 394 patients (84.2%) who started peritoneal dialysis in 2017 were still receiving it at 1 year.
Large-scale expansion of peritoneal dialysis for advanced kidney disease appears to be feasible through a coordinated, integrated health care delivery framework that applies a multidisciplinary system-level approach.
Despite favorable national trends in the incidence of end-stage renal disease (ESRD) from 2008 to 2011, ESRD incidence has been increasing recently, and less than 10% of patients with ESRD start renal replacement therapy with peritoneal dialysis (PD) in the United States. Given known and potential advantages of PD over hemodialysis, the Kaiser Permanente Northern California integrated health care delivery system implemented a program to expand use of PD.
To describe the system-level approach to expansion of PD use and temporal trends in initiation and persistence of PD and its associated mortality.
Design, Setting, and Participants
This retrospective cohort study included adult members of a large integrated health care delivery system in Northern California who initiated chronic dialysis therapy from January 1, 2008, through December 31, 2018. Data were analyzed from March 1, 2018, through May 31, 2019.
From 2008 to 2018, Kaiser Permanente Northern California implemented a multidisciplinary, system-wide approach to increase use of PD that included patient and caregiver education, education and support tools for health care professionals, streamlined system-level processes, monitoring, and continuous quality improvement.
Main Outcomes and Measures
Temporal trends in the proportion of patients starting chronic dialysis with PD vs hemodialysis compared with national trends. Secondary outcomes included persistence of PD at 1 year in those initiating it and standardized 1-year mortality rates in those initiating PD or hemodialysis.
Among 13 500 eligible health plan members in the study population (7840 men [58.1%] and 5660 women [41.9%]; mean [SD] age, 64.3 [14.4] years), initiation of PD increased from 165 of 1089 all new dialysis patients (15.2%) in 2008 to 486 of 1438 (33.8%) in 2018, which was substantially higher than national trends (6.1% in 2008 and 9.7% in 2016). Among the 2974 patients who initiated PD from 2008 to 2017, 2387 (80.3%) continued PD at 1 year after initiation, with a significant increase in age-, sex-, and race-standardized rates from 2008 (69.1%) to 2017 (84.2%). Age-, sex-, and race-standardized 1-year mortality for patients receiving PD and hemodialysis did not change significantly across this 10-year period (17.3% to 15.5% for hemodialysis, P = 0.89 for trend; and 5.5% to 7.3% for PD, P = 0.12 for trend).
Conclusions and Relevance
This study suggests that large-scale expansion of PD is feasible using a multidisciplinary, integrated, coordinated care approach; we believe these findings represent a national opportunity to improve outcomes for patients with advanced kidney disease.
From 2007 through 2014, an estimated 30 million US adults had chronic kidney disease (CKD),1 and more than 124 000 patients with end-stage renal disease (ESRD) started renal replacement therapy in 2016.2 To date, no definitive randomized clinical trials have evaluated clinical outcomes between peritoneal dialysis (PD) and hemodialysis,3 but several observational studies suggest comparable or better outcomes and patient-centered advantages with PD.4-7 Despite this and recent Medicare value-based reimbursement changes favoring greater use of PD as the initial renal replacement therapy modality, less than 10% of patients with incident ESRD initiated PD in 2016 nationally, and only 7% with prevalent ESRD were receiving PD.2 The latter rate was substantially lower than that in other countries (eg, 70% in Hong Kong, 51% in Mexico [Jalisco region], 30% in New Zealand, 21% in Sweden, and 19% in Canada).8
Multiple factors have likely contributed to underuse of PD nationally, such as financial incentives for health care professionals to use hemodialysis, inadequate training of nephrologists and nurses in providing effective PD and related patient education, fragmented care in the transition from advanced CKD to ESRD in many practice settings, and recent shortages in the availability of peritoneal dialysate.9,10 Finally, a persistent misconception by some health care professionals is that PD leads to inferior clinical outcomes.11
With national efforts underway to reward value-based care through coordinated management strategies, effective health system–level solutions are needed to expand the use of PD nationally and to reduce unnecessary variation across providers. We herein describe what we believe is a successful approach used to substantially increase the incidence and prevalence of PD within a large, integrated health care delivery system and secondarily report on trends in persistence and short-term survival associated with this system-level strategy.
Kaiser Permanente Northern California (KPNC) is an integrated health care delivery system currently caring for more than 4.3 million members through 21 medical centers and more than 245 offices in the San Francisco and greater Bay Area. Its membership has broad age, sex, racial/ethnic, and comorbidity diversity and is highly representative of the local and statewide population.12 Kaiser Permanente Northern California consists of an exclusive partnership between the not-for-profit Kaiser Foundation Health Plan, Inc/Kaiser Foundation Hospitals and The Permanente Medical Group, with more than 80 board-certified nephrologists on faculty as of 2018. This study was approved by the institutional review board of KPNC, and a waiver of informed consent was obtained owing to the retrospective nature of the study. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
Each KPNC medical center is represented by a nephrology department chief who works as a part of the regional nephrology service line to identify system-level priorities that are then addressed through each medical center’s physician faculty, medical and support staff, and local infrastructure. Physicians, renal nurse case managers, and support staff within KPNC are responsible for assisting patients in their decision-making about whether to receive renal replacement therapy and the timing and type of dialysis therapy if they choose it, but long-term dialysis therapy is delivered primarily through contracted external service providers. From January 2008 through December 2018, a multiple-component strategy was implemented to increase use and persistence of PD by engaging patients, health care professionals, and regional health system and medical group stakeholders.
Using a comprehensive regional ESRD treatment registry,13 we identified all patients aged 18 years or older who initiated dialysis each year from 2008 through 2018 within KPNC. We excluded patients who received a kidney transplant at any time before initiating dialysis, and we focused on adults initiating PD vs in-center hemodialysis using manually confirmed dialysis modality and start date information found in electronic health records and dialysis therapy records. Furthermore, we considered patients initiating hemodialysis who then switched to PD within 90 days to presumably be receiving bridge hemodialysis therapy, and they were classified as undergoing PD. Persistence of PD was defined as the proportion of patients initiating PD who were alive and still receiving the modality at 1 year after starting or who died while receiving PD within 1 year after initiation without switching to hemodialysis. Deaths that occurred through December 31, 2018, were comprehensively identified from electronic health records, member proxy reporting, Social Security Administration files, and California death certificate information.14,15
In 2007 through 2008, based on available published evidence, KPNC nephrology leadership focused on expanding PD as the preferred initial modality among patients choosing to receive chronic dialysis therapy. Subsequent engagement with surgery and internal medicine stakeholders led to development of a systematic, coordinated approach to identify patients with advanced CKD using population management tools, dissemination of best practices for CKD management, processes for establishing necessary infrastructure and workflows, and tracking performance metrics for initiation and persistence of PD across all medical centers. Regular reporting and feedback to nephrology and medical center leadership on targeted performance metrics supported continuous quality improvement, with progressively higher target PD rates set during the study period. In 2012, medical group leadership also adopted expansion of PD as a program-wide priority.
Four main components constituted the system-level solution to expand PD within an integrated care infrastructure. The components are described briefly below, with more specific details provided in the Kaiser Permanente Peritoneal Dialysis Playbook (eAppendix in the Supplement).
More systematic education was provided to patients with CKD and their families about renal replacement therapy choices and to promote a PD-first approach if the patient chose to start dialysis. Components included enrolling all health plan members with CKD in educational programs, along with revision of educational resources to provide a consistent and accurate message about the potential benefits and risks of all renal replacement therapy options. Patients considering dialysis therapy were encouraged to meet with PD clinic staff to have their questions answered about this modality and to provide reassurance about an individualized long-term plan for renal replacement therapy and associated support that would be provided to the patient and their caregivers.
A multidisciplinary approach was implemented to educate health care professionals across different specialties about the importance of PD as an initial dialysis modality. In 2008, the regional Kaiser Permanente PD University program was launched to increase awareness of health care professionals about the availability and potential advantages of PD for patients, to eliminate possible misconceptions about the therapy, and to increase knowledge on how PD is delivered through a team approach within the health system (eAppendix in the Supplement). Given their critical role in the process, renal nurse case managers were provided additional in-depth training in patient education and how to tailor available support for patients based on the dialysis modality selected, along with regular educational and training updates.
Key operational improvements were adopted, including development of an urgent-start PD program in 2011, expanded use of embedded PD catheters, and increased internalization and reorganization of PD services tailored to each medical center’s infrastructure and population. In addition, health plan and medical group leadership routinely met with teams at each medical center to emphasize expansion of PD as a regional priority for patients choosing dialysis and to address any local logistical or other challenges. To further align priorities, KPNC nephrologists took on medical directorships at multiple PD units serving health plan members (without any additional personal or health system monetary compensation) to increase further direct oversight of quality of care and outcomes improvement.
Reports were provided to each medical center’s nephrology and leadership teams, which included monthly and 12-month rolling incidence and prevalence of PD use to track progress and to identify quality improvement opportunities. Regional and medical center–level performance was also reviewed at quarterly nephrology chiefs’ meetings to discuss and share best practices across medical centers and to address any remaining workflow, infrastructure, or other challenges.
Data were analyzed from March 1, 2018, to May 31, 2019. For each year from 2008 to 2018, we calculated the proportion of patients with ESRD initiating PD or hemodialysis in KPNC to compare with national benchmark rates reported by the US Renal Data System.16 Temporal trends in the incidence of PD were evaluated using Cochrane-Armitage tests for trend, with a 2-sided P < .05 indicating significance. One-year persistence of PD and all-cause mortality were calculated among those initiating dialysis from 2008 through 2017, to allow for at least 1 year of follow-up. The rates of persistence and all-cause deaths were directly standardized to the distributions of age (18-39, 40-64, 65-75, and ≥75 years), sex (male or female), and self-reported race/ethnicity (white, black, Asian or Pacific Islander, and other/unknown) of the KPNC population initiating PD in 2008. For reporting on temporal trends in 1-year all-cause mortality among patients receiving hemodialysis, the same approach was used through age, sex, and race/ethnicity standardization to patients in KPNC initiating hemodialysis in 2008. We evaluated for any temporal trends in the outcomes of interest adjusted for age, sex, and race/ethnicity using multivariable logistic regression with a calendar year term. All analyses were conducted using SAS, version 9.3 (SAS Institute, Inc).
From 2008 to 2018, we identified 13 500 eligible adult members initiating chronic dialysis, with 3460 (25.6%) starting with PD and 10 040 (74.4%) starting with hemodialysis overall. Mean (SD) age was 64.3 (14.4) years; 5660 (41.9%) were women and 7840 (58.1%) were men; 2944 (21.8%) were of Hispanic race/ethnicity, and the population was racially and ethnically diverse (4885 [36.2%] were white, 2176 [16.1%] were black, 3078 [22.8%] were Asian/Pacific Islander, and 3361 [24.9%] were of other or unknown race/ethnicity). Incidence of PD more than doubled during the study period, from 165 of 1089 patients (15.2%) in 2008 to 486 of 1438 (33.8%) in 2018 (P < .001), with a modest drop in incidence from 2014 (390 of 1239 [31.5%]) to 2017 (394 of 1425 [27.6%]) (Figure 1). These rates were notably higher than national rates, which increased but remained below 10% from 2008 (6.1%) to 2016 (9.7%) (Figure 1). Among 2974 KPNC patients initiating PD from 2008 to 2017, 2387 (80.3%) continued PD at 1 year after initiation, with age-, sex-, and race-standardized rates increasing from 69.1% in 2008 to 84.2% in 2017 (P < .001 for trend) (Figure 2). The proportion of patients alive and still receiving PD at 1 year also increased significantly from 108 of 165 (65.5%) in 2008 to 306 of 394 (77.7%) in 2018. (P < .001 for trend), whereas the proportion of patients dying or disenrolling while receiving PD before 1 year was not significantly different during the study period (Figure 2).
In contrast, no significant age-, sex-, and race-standardized 1-year all-cause mortality temporal trends by dialysis modality were noted for patients initiating PD or hemodialysis from 2008 to 2017 (17.3% to 15.5% for hemodialysis, P = 0.89 for trend; and 5.5% to 7.3% for PD, P = 0.12 for trend) (Figure 3). The mean absolute 1-year age-, sex-, and race-standardized risk of death was elevated (17.7%) for patients initiating hemodialysis throughout the study period (Figure 3).
Through a multidisciplinary, system-level approach, PD initiation increased from 15.2% to 31.5% within 6 years within a large, integrated health care delivery system in Northern California. The modest decline in new PD initiations from 2015 to 2017 was driven primarily by an unexpected national shortage of peritoneal dialysate solution by its manufacturers that precluded many patients from starting PD and had an extended effect on the care pathways that had been established. However, as of 2018, we were able to fully implement our program, which led to 33.8% of new dialysis starts being PD in that year. Despite expanding the use of PD to a broader patient population over time, 1-year persistence of PD increased over time as well, with 84.2% of patients starting PD in 2017 still receiving this modality a year later. In addition, no significant temporal trends were noted in age-, sex-, and race-standardized 1-year mortality in patients initiating PD or hemodialysis during the study period.
Nationally, various factors have driven greater use of in-center hemodialysis compared with PD. Fee-for-service and volume-based reimbursement has favored in-center hemodialysis for decades, but the transition to a value-based reimbursement system focused more on better clinical outcomes, patient satisfaction, and lower overall costs has made PD a more attractive modality, in addition to home hemodialysis. After implementation of the expanded bundle in 2011, the Centers for Medicare & Medicaid Services expanded financial incentives for home dialysis by offering identical payments to providers of dialysis services for PD and in-center hemodialysis, but this has contributed to only a modest increase in PD prevalence nationally.16,17 However, some concerns existed that clinical outcomes could be compromised with expanded PD use in settings with inadequate health system infrastructure and/or lack of trained health care professionals. We demonstrated the feasibility for large-scale expansion of PD as the preferred initial renal replacement therapy option within an integrated care delivery model, with a high level of therapy persistence and no significant temporal change in all-cause mortality associated with the expanded access to PD within our population.
Potential physician bias and inadequate training may also contribute to low PD rates nationally. Many physicians still have a misconception of hemodialysis being systematically more effective than PD,11 which, among other factors, may have contributed to self-reported perceptions of inadequate PD training among US nephrology fellows.18,19 Inadequate training and experience in providing high-quality PD could lead to less confidence caring for patients who receive PD or recommending it to patients with CKD. This possibility can be further complicated by late referral of patients with advanced CKD to nephrologists, which can negatively influence the likelihood of patients choosing PD.20
However, recognition of various benefits of PD compared with in-center hemodialysis (eg, greater patient engagement, improved cognition and quality of life,21 higher likelihood of retaining employment,22 and preservation of residual kidney function23) is increasing. Furthermore, high-quality PD is associated with lower medical costs24 and allows for broader access to renal replacement therapy, including within medical resource–limited environments. However, successfully achieving and maintaining this large-scale expansion required a multipronged approach to planning and implementation, comprehensive integration of care (including use of a single electronic health record system across all practice settings), alignment and commitment of multiple specialties (eg, nephrology, general surgery, interventional radiology, general internal medicine, family medicine, and cardiology), an adequate number of well-trained staff to facilitate PD initiation and enhance PD persistence, systematic performance tracking and sharing of best practices (eAppendix in the Supplement), and prioritization and strong support from medical group and health system leadership. These efforts followed previous successful population-based models for improving cardiovascular risk reduction within KPNC.25-28
Despite our large, contemporary, sociodemographically diverse community-based population, our ability to increase the proportion of PD among new dialysis initiations may not fully generalize to all practice settings in the United States, especially for those that do not have adequate infrastructure, trained physicians and staff, or coordinated education and support for patients and caregivers. Our health care delivery system uses a fully integrated care model with alignment among the physicians, staff, and insurer and with a single, comprehensive electronic health record system, which facilitates implementation of population health– and system-level quality improvement initiatives across inpatient, emergency department, and outpatient settings, but other forms of accountable care organizations are much less integrated and may require different solutions to increase PD use. The purpose of our temporal trend analysis of all-cause mortality over time in patients with PD was primarily to address possible concerns of worse outcomes with systematic expansion of PD to less selected patients, for which we found no evidence in our population. However, we acknowledge that our study does not directly address whether initial PD vs home hemodialysis vs in-center hemodialysis leads to differential net clinical outcomes in particular groups of patients with CKD; that would require conducting adequately powered randomized clinical trials, which is not feasible in the United States. We also lacked systematic information on PD access-related complications or on the individual-level use of acute PD.
Our findings suggest that PD remains underused nationally, with long-term efforts needed to expand the trained workforce and associated infrastructure to provide high-quality PD along with more systematic patient engagement and support. Our findings support the feasibility of coordinated expansion of the ability to provide PD within an integrated care delivery model to improve outcomes for patients with CKD.
Accepted for Publication: June 6, 2019.
Corresponding Author: Alan S. Go, MD, Division of Research, Kaiser Permanente Northern California, 2000 Broadway, Oakland, CA 94612 (email@example.com).
Published Online: September 9, 2019. doi:10.1001/jamainternmed.2019.3155
Author Contributions: Mr Parikh and Dr Go had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Pravoverov, Zheng, Tan, Bhalla, Reddy, Mroz, Jonelis, Go.
Acquisition, analysis, or interpretation of data: Pravoverov, Zheng, Parikh, Jonelis, Go.
Drafting of the manuscript: Pravoverov, Zheng, Parikh, Reddy, Jonelis, Go.
Critical revision of the manuscript for important intellectual content: Pravoverov, Zheng, Parikh, Tan, Bhalla, Mroz, Jonelis, Go.
Statistical analysis: Parikh, Go.
Obtained funding: Go.
Administrative, technical, or material support: Pravoverov, Zheng, Tan, Bhalla, Mroz, Jonelis, Go.
Supervision: Zheng, Tan, Jonelis, Go.
Conflict of Interest Disclosures: Dr Go reported receiving grants from the National Institutes of Diabetes, Digestive and Kidney Diseases and the Kaiser Permanente Northern California Community Benefit Fund during the conduct of the study. No other disclosures were reported.
Funding/Support: This study was supported in part by research grant U01 DK060902 from the National Institute of Diabetes, Digestive and Kidney Diseases of the National Institutes of Health and by the Kaiser Permanente Northern California Community Benefit Fund (Dr Go).
Role of the Funder/Sponsor: The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We thank all the physicians, nurses, and staff who provide care for patients with chronic kidney disease and end-stage renal disease. Farzien Khoshniat-Rad, BS, Kaiser Permanente Northern California, provided his technical assistance, for which he was compensated.