AJKDblog: What events led to the creation of the Peer Kidney Care Initiative (Peer)?
AC: Peer grew out of a collaborative, informal relationship among the chief medical officers (CMOs) of 13 dialysis providers across the United States. The CMOs initially met in March 2013, at a meeting organized by DaVita HealthCare Partners, Dialysis Clinic, Inc. (DCI), Fresenius Medical Care, and Renal Ventures Management. At that meeting, the CMOs discussed a number of clinical issues, including fluid control, infection, sudden cardiac death, and vascular access. The CMOs met again one year later in Baltimore. At that meeting, the group began to consider a more comprehensive effort centered on data sharing, relying on both Medicare and provider data to guide quality improvement efforts at the local and national levels.
The Chronic Disease Research Group (CDRG), which previously served as the contractor for the United States Renal Data System (USRDS) Coordinating Center, spoke at the Baltimore meeting and presented an array of issues related to morbidity and mortality that providers could consider for focused attention. In ensuing months, CMOs and CDRG jointly developed the Peer Kidney Care Initiative. Peer launched its internet home, www.peerkidney.org, in January 2015.
AJKDblog: Who is involved in Peer?
AC: There are currently 14 dialysis provider organizations that participate in Peer: American Renal Associates, Atlantic Dialysis Management Services, Centers for Dialysis Care, DaVita HealthCare Partners, Dialysis Clinic, Inc. (DCI), DSI Renal, Fresenius Medical Care, Health Systems Management (managing facilities in Emory Healthcare and Wake Forest Baptist Health), Independent Dialysis Foundation, Northwest Kidney Centers, Renal Ventures Management, Satellite Healthcare, The Rogosin Institute, and U.S. Renal Care. These organizations operate a 9-person Steering Committee, which works with CDRG to coordinate Peer analytics and efforts. I (Allan Collins, MD, former director of the USRDS Coordinating Center) am the Executive Director of Peer. A number of physicians, epidemiologists, and biostatisticians at CDRG participate in the day-to-day operations of Peer.
AJKDblog: What is the overall goal of Peer?
AC: The overarching goal of Peer is to improve the quality of dialysis patient care in the United States. Of course, quality improvement efforts can take many forms. In the case of Peer, quality improvement takes the form of using data and sophisticated epidemiologic analysis to discover trends and variations in clinical outcomes. Peer is really moving along two paths. One path involves taking a deeper dive into data than public reporting mechanisms typically have. For example, in our analyses of hospitalization among dialysis patients, we analyzed trends in 14 specific clinical events.
The second path involves helping provider organizations to better understand variations in clinical outcomes among their respective facilities. Dialysis Facility Compare (DFC) is a fine tool, but DFC is focused on rates of “all-cause” events: all-cause mortality, all-cause hospitalization, all-cause readmission, et cetera. On one hand, all-cause events are straightforward to quantify and insensitive to coding practices. On the other hand, such events may not offer a clear window to clinical problem areas. Is an elevated mortality rate due to an excess of sudden cardiac death; challenges with fluid overload and heart failure; too many infections? Peer aims to use retrospective data to its fullest potential by assessing specific clinical events that can be targeted for active intervention.
AJKDblog: What are the highlights of the supplement?
AC: The supplement is Peer’s first major publication. We used a few themes to present data that were truly unlike anything that the USRDS presents. The USRDS provides a wonderful service to the nephrology community, and there is no need to duplicate its efforts. Specifically, Peer wrapped its analysis in three themes:
- Secular trends: We included many displays of trends over 4-year, 8-year, and 16-year horizons. Trends, especially in local areas with limited year-to-year variation in the composition of the patient population, are useful tools for identifying clinical domains in which progress has been achieved and domains in which progress has not been realized.
- Geographic variation: The supplement includes many maps. This country is diverse; there is substantial state-to-state variation in patient characteristics, economic climates, and the quality of health care delivery. Although the mortality rate in dialysis patients has been declining, rates have not been declining in all parts of the country. States along the Ohio River are struggling to make as much progress in improving outcomes as other states, while Texas and its neighbors are achieving tremendous year-over-year progress.
- Seasonality: Almost everything in the dialysis patient population seems to vary across the seasons. New patient counts peak in the winter, mortality rates peak in the winter, and many – but not all – cause-specific hospitalization rates peak in the winter. (Admissions for access infections actually peak in the summer!) These patterns suggest the dialysis facilities can improve outcomes by orienting care toward reducing the risks of the season.
AJKDblog: What are the major obstacles in improving patient care that Peer is trying to overcome?
AC: Dialysis providers have EHR systems for recording what happens inside the dialysis facility. Understanding what happens outside of the facility is much more difficult, but clearly important. As many in the field can appreciate, simply obtaining discharge records from hospitals can be a challenging task. Peer helps dialysis providers better understand what happens in hospitals, emergency and observation rooms, and pharmacies. Peer is also interested in shifting the conversation about quality metrics toward patient-centered clinical outcomes. Among facilities operating in 2013, the average percentage of patients with Kt/V exceeding 1.2 was roughly 90%. Meanwhile, the average percentage of patient-months with hypercalcemia (> 10.2 mg/dL) was 4.1%. Improving performance on each of these metrics has limited potential impact. Peer aims to expand the conversation about metrics to common sources of mortality and morbidity, in order to maximize potential impact.
AJKDblog: Any thoughts on the new 5-star rating system of dialysis units covered in the supplement?
AC: The rating system ostensibly exists to help patients and their families evaluate dialysis facilities. If so, then the system should value what patients and their families value. In the supplement, we showed that even in the context of the current components of 5-Star, ratings vary when the formula that underlies the rating is perturbed. In other words, a single rating for a facility is necessarily built upon a specific conception of quality. Do CMS and patients see quality in the same way? The answer is unclear. At minimum, we need to think about whether a single rating is a good idea.
By its very nature, 5-Star names equal numbers of winning and losing facilities, because the methodology imposes a bell-shaped distribution of performance on each of the constituent metrics. This might seem intuitive, but it comes at a high cost: over time, it is impossible to discern whether facilities are improving outcomes on an absolute scale. Standardized event ratios, like the standardized mortality ratio and the standardized hospitalization ratio, likewise suffer from this problem.
An alternative approach that Peer seeks to advance is to examine trends in outcomes at the levels of facility, state, and region to determine if progress has been achieved within these units. This approach recognizes that steadily improving outcomes for all subsets and areas of the patient population is an important goal that should be formally recognized. From a methodological point of view, rating according to trends in local areas may obviate some of the (arguably intractable) challenges with risk adjustment that are inherent in administrative data; it is difficult to list and measure all of the factors that distinguish West Virginia from California, for example.
AJKDblog: What can we expect from Peer in the future?
AC: Since the supplement went to press, Peer has acquired more timely data, which will be featured in the next supplement and in manuscripts that will be submitted to journals across the nephrology landscape. Peer will continue to add content to www.peerkidney.org. For example, each month, Peer publishes an educational flier about a timely topic. Our May flier, for example, described the impressive progress that has been made toward achieving Healthy People 2020 objectives in the dialysis patient population. For June, we prepared a flier about access infections. These fliers are free to download and distribute to staff and patients alike.
Finally, Peer will be at ASN Kidney Week this year. In addition to having submitted several abstracts for presentation, Peer will participate with DOPPS and the SRTR in a 2-hour Clinical Practice Session on Friday afternoon about the use of databases in setting policy regarding ESKD.
To view the supplement (freely available), please visit AJKD.org.