July 21, 2008
MINUTES
Kidney Care Quality Alliance Steering Committee
July 21, 2008
Conference Call
Attendees
Gail Wick (Co-chair), ANNA
Barbara Fivush, ASPN
Barry Straube, CMS
LeAnne Zumwalt, DaVita
Ray Hakim (Co-chair), FMC
Dolph Chianchiano, NKF
Maureen Michael, NRAA
Allen Nissenson, RPA
Linda Keegan, Robyn Nishimi, Lisa McGonigal, Kathy Lester, KCP
Summary
Dr. McGonigal opened by taking a roll call. Drs. Hakim and Nishimi then advised the Steering Committee that the conference call would be used to review and discuss: (1) the recommended measure-testing plan for the five Kidney Care Quality Alliance (KCQA) performance measures that received time-limited endorsement from the National Quality Forum (NQF), and (2) the proposed protocol for identifiying the list of target facilities for testing. Dr. McGonigal led the ensuing discussion during which the Steering Committee deliberated and reached agreement on a testing plan and the proposed testing site distribution.
FACILITY-RELATED SAMPLING FACTORS
Dr. McGonigal informed the Steering Committee that five sampling parameters were identified as being important in the recommended testing protocol: facility type, proportion of facility types, number
of facilities, testing protocol type, and geographic distribution.
Recommended Facility Types
Dr. McGonigal reported that, per the protocol, testing will be performed on a nationally drawn sample reflective of the current dialysis facility industry profile. The sample will contain a mix of both for-profit and not-for-profit hospital-affiliated and freestanding facilities within large, small, and independent dialysis organizations in urban, suburban, and rural settings. Facilities both with and without EMRs will be included in the sample. However, Department of Veterans Affairs-affiliated and other public facilities will be excluded to streamline the recruitment process.
Recommended Facility Type Proportions
Dr. McGonigal next informed the Committee that data from the USRDS 2007 Annual Report was examined to determine the current relative proportions of each of the facility types identified above. She noted that limitations in the existing data, however, restricted the analysis to a comparison of large dialysis organizations versus non-LDOs and for-profit versus not-forprofit facilities. Moreover, as the collection of proprietor information was discontinued in 2002, she reported that it was necessary to coalesce ownership data from 2001 with profit status data from 2005 to generate a plausible conclusion as to the current industry composition. In addition, there have been a substantial number of mergers and acquisitions over the past few 2 years that are not reflected in existing data. Despite the lack of homogeneous data available for analysis, she noted that calculations suggest that the following proportions are a reasonably accurate representation of the current national profile: 75% LDOs / 25% non-LDOs and 80%
for-profit / 20% not-for-profit.
TABLE 1: FACILITY PROPORTIONS
| For-Profit | Not-For-Profit | Total | |
|---|---|---|---|
| LDO | 60% | 15% | 75% |
| Non-LDO | 20% | 5% | 25% |
| Total | 80% | 20% | 100 |
Based on this analysis, the recommended testing sample would consist of the following percentages of facility types:
- 60% for-profit LDOs;
- 15% not-for-profit LDOs;
- 20% for-profit non-LDOs; and
- 5% not-for-profit non-LDOs.
Recommended Number of Facilities
Dr. McGonigal next informed the Committee that a previous contractor’s recommendation that testing be performed at 50-100 randomly selected entities (facilities and physicians) was taken into consideration in determining the appropriate number of testing sites. Rather than limiting testing to a single ESRD network as was recommended by that contractor, however, Dr. McGonigal noted that a broader, national sampling representative of the current industry profile is recommended. Specifically, three facilities would be selected from each of the 18 ESRD Networks—a total of 54 facilities—plus six additional sites in Network 7 (Florida). She reported that Network 7 will be oversampled to take advantage of the location of the Centers for Medicare and Medicaid (CMS) contractor that developed the CROWNWeb 1.0 Data Dictionary so as to facilitate potential consultation with this organization during travel related to data auditing. Thus, a total of 60 facilities is proposed.
Recommended Testing Protocols
To minimize both burden to selected facilities and KCQA’s costs, Dr. McGonigal reported that the current recommendation is that all five measures not be tested at all participating sites. Rather, each of the following testing protocols will be implemented at 20 facilities (1 facility per Network plus two additional facilities in Network 7):
- Protocol 1: Test all five measures
- Protocol 2: Test three measures—the Influenza Immunization and two Vascular Access measures
- Protocol 3: Test four measures—the two Vascular Access and two Patient Education Awareness measures
Recommended Geographic Distribution
Dr. McGonigal again noted that facilities will be selected from each of the 18 ESRD Networks to generate a national sample.
PATIENT-RELATED SAMPLING FACTORS
Dr. McGonigal informed the Steering Committee that two patient-related sampling factors were identified: patient type and number of patients, as noted below and in the discussion.
Recommended Types of Patients
Dr. McGonigal reported that both peritoneal and hemodialysis patients will be included in the sample. Additionally, patients also will be categorized according to time on dialysis (i.e., 90 days, 90 days – 1 year, and > 1 year) to appropriately reflect distinctions made in the measures’ specifications.
Recommended Number of Patients
Dr. McGonigal noted that 25 patients per facility will be sampled.
PREPARATION FOR TESTING
Next, Dr. McGonigal advised the Steering Committee of the activities that are required in advance of the commencement of measure testing, summarized in the following sections.
Recruitment and Training of Facilities
Dr. McGonigal first noted that there will be a targeted invitation of facilities, with a goal of achieving a sample of 60 sites representative of the current national industry profile; invitations will be sent to 20% more than the target so as to achieve an N=60. Invitational letters will be sent from the KCQA and, where appropriate and feasible, from the KCP contact to which a facility belongs. Follow-up phone calls will be placed to non-respondents, and facilitation by Steering Committee and KCP members will be requested when necessary. The target for training is August 2008, and dry runs on small samples of patients will be performed in advance of the commencement of data collection, time and resources permitting.
Measure-Specific Near-Term Tasks
Dr. McGonigal next noted the KCQA’s near-term measure-specific tasks were:
- Review the KCQA Vascular Access data elements, as defined by CMS. As noted in previous communications, Dr. McGonigal reported that CMS has already defined and incorporated these data elements into the CROWNWeb 1.0 software and will begin data collection in February 2009. However, CMS has expressed its willingness to work with KCQA to revise these data elements prior to the release of CROWNWeb 2.0 in July 2009, should they not satisfactorily meet KCQA’s expectations or perform as anticipated during pilot testing. It is thus essential that KCQA begin its review of the CMS data elements and make plans for any changes to be incorporated into updates of CROWNWeb.
- Review the RPA/PCPI physician-level Influenza Immunization measure data elements and identify any changes or additions necessary to bring the measure into alignment with the analogous KCQA facility-level measure. Dr. McGonigal noted the specifications for the two measures are already known to differ. For the KCQA’s data elements to be incorporated into CROWNWeb, they must be provided to CMS in August 2008 for the announcement of the CROWNWeb 2.0 business requirements..
- Review the KCQA Patient Education Awareness measures’ draft specifications and data elements prepared by Drs. Nishimi and McGonigal and identify any changes or additions deemed necessary and appropriate. As with the Influenza Immunization measure, the data elements need to be provided to CMS in August 2008.
Testing Timeline
Dr. McGonigal next advised the Steering Committee of the proposed testing timeline:
- September 1, 2008 – Begin electronic or paper data collection for all measures
- November 30, 2008 – Q1 data due
- December 2008 – Verify/audit Q1 submissions and review data with KCQA Steering Committee and facilities
- February 28, 2009 – Q2 data due
- March 2009 – Verify/audit/review Q2 data
- May 31, 2009 – Q3 data due
- June 2009 – Verify/audit/review Q3 data
- August 31, 2009 – Q4 data due
- September 2009 – Verify/audit/review Q4 data
- October 2009 – Analyze data, prepare draft report for purposes of gaining permanent NQF endorsement, and review report with KCQA SC
- December 2009 – Finalize report
STEERING COMMITTEE DISCUSSION AND RECOMMENDATIONS
Dr. Hakim requested that the Steering Committee specifically consider the recommended testing
protocols. In particular, he opined that testing all five measures at all facilities would be preferable to the three separate protocols recommended above, each of which was to be implemented in 20 facilities. He noted that the current recommenation was intended to minimize the burden to participating facilities, but that testing all measures at all facilities would provide a more accurate sense of the time and effort required to collect the necessary data. Dr. Nissenson agreed that more robust information on the amount of work involved would be important when presenting results of testing to NQF, and ultimately, to CMS. The Steering Committee unanimously agreed with this opinion and it approved a plan to test all measures at all participating facilities.
Ms. Michaels next expressed concern that collection of the necessary data elements might be quite burdensome, particularly to small, independent dialysis facilities. Dr. Nishimi responded that measure testing is intended, among other things, to test the feasibility and burden of data collection, and further reminded the Steering Committee that testing will be limited to 25 patients per facility to minimize burden. Ms. Michaels agreed that this sample size should not prove problematic.
Ms. Michaels then remarked that it is unclear where and how clinician-level data for the Patient Education Awareness measures will be captured, given that data for CROWNWeb will be entered by the facilities. Multiple Steering Committee members agreed that neither the facilities nor the doctors will be accepting of or satisfied with this intended arrangement, and speculated that facilities will often not have this information available. Dr. Straube agreed that differentiating facility from clinician responsibility is difficult and that defining the accountable entity will be problematic where healthcare systems are not integrated.
Next, Dr. Straube noted that CMS is very supportive of the KCQA’s intent to test its measures and agreed that an assessment of the burden of data collection is very important. He noted, however, that as there is a push from Congress to proceed rapidly, testing timelines for measures are generally proving problematic. Specifically, he remarked that the timeline from inception to completion of testing for some recently endorsed (non-ESRD) measures (e.g., deep venous thrombosis measures endorsed by NQF) has taken more than seven years. He noted that if measure stewards are unable to find a way to expedite this process, Congress will find and impose a way. He suggested that collaboration might be a way to avoid such mandates.
Dr. Hakim acknowledged Dr. Straube’s concerns and assured him that if there is a feasible way to expedite testing without compromising the resulting data, it will be done. Dr. Chianchiano asked what effect testing all the measures at all facilities, in lieu of the recommended three testing protocols summarized above, would have on the timeline. Dr. McGonigal responded that this modification would have no impact on the timeline. Dr. Straube next queried whether testing on twice the number of patients in half of the time would be feasible and provide the same information. Dr. Nishimi responded that due to the technical construction of the measures (e.g., a 1-year data collection period or collection during a specific time of the year), this might prove problematic; she agreed that the option should be further investigated and given consideration.
Dr. Straube next suggested that involving the ESRD Networks might facilitate the recruitment process, as facilities might be more responsive. He noted that the Networks’ budgets are somewhat limited, however, and might not be able to support this additional activity. He remarked that he would explore this option further, noting that the Networks’ involvement would not include activities such as “data crunching”.
Ms. Lester next provide the Steering Committee with a brief update on the CMS Clinical Performance Measures (CPM) issues. She also provided a brief summary of CMS’ Quality Measure Development Overview—CMS’ written policy on measure development. Lastly, she noted that a letter from the KCP on the Conditions for Coverage (CFC) is currently being finalized.
Following a request from the Steering Committee for an update on the CPMs, Dr. Straube acknowledged that following introduction of the CPM requirements, CMS was afforded little time to create and finalize the list of CPMs relevant to the CFCs. Moreover, he noted there was, and continues to be, considerable confusion among providers between the CPMs and the Physician Quality Reporting Initiative (PQRI). Because of continued confusion on the inclusion of the anemia CPMs, Dr. Straube also clarified there is no statutory requirement that CMS utilize NQF-endorsed measures for the CFCs, and that decisions are ultimately at the discretion of the Secretary. He remarked that CMS of course prefers measures developed by experts in a given field, fortified with input from statisticians. He acknowledged that the anemia management measures were and continue to be a contentious issue. He reported that the fact that the NQF process did not reach consensus onan acceptable upper limit for hemoglobin was problematic for CMS, and so CMS arbitrarily selected the 10-12mg/dL range. He noted that this value is subject to change in the future.
Ms. Zumwalt remarked that it would be preferable if CMS were to select only one measure per category to minimize burden and suggested that this be considered in the future. Dr. Straube agreed that a goal for future efforts should be to define such concerns and refine the process accordingly in advance so as to avoid the necessity for regulatory changes or congressional statutes. He remarked that to achieve these goals, it is necessary for people to get involved far enough in advance to preempt problems. He suggested that KCP send CMS a letter outlining such concerns. Ms. Lester noted that the KCP in currently in the process of drafting such a letter.
NEXT STEPS
Dr. McGonigal advised the Steering Committee’s that facility identification and recruitment are the
highest priority, as will the defined measure-specific tasks. As per the Steering Committee’s decision, all five measures will be tested at all facilities in the pilot.
