2013 QRUR Report Analysis – LTC Medicine is Underwriting CMS’ risk for highest cost beneficiaries.December 3rd, 2014 by
Thanksgiving week I had the opportunity to meet with the CMS staff responsible for managing the Value Based Purchasing program. AMDA (the American Medical Directors Association) leadership asked me to participate in a dialogue with CMS. We were there to explore options for avoiding the ‘high cost’ classification LTC Medical Groups universally receive under the VBP cost analysis. Those classifications are found in the Quality Resource and Use Report (QRUR) for each group. The Quality component varies wildly because there was no standardized reporting strategy. Every Group received a Cost tiering (low, average, high). Based on copies of the reports for 7 different LTC Medical Groups, from 2012 and/or 2013, costs were uniformly high. Expressed as Standard Deviations from the mean (VBP terminology), the groups were always high ranging from +1.76 up to +2.67 deviations from the mean (VBP terminology). Any group with a score ≥1.0 for 2015 would be high cost and have a <2%> downward payment adjustment applied to 2017 Part B reimbursements.
CMS distributes these reports for the following stated purpose:
|As part of its value-based purchasing initiatives, for the past several years under the Physician Feedback/Value-Based Modifier Program, CMS has disseminated a limited number of confidential reports to physicians and groups of physicians that include measures of resource use and quality. CMS has pursued a phased approach to implementing physician feedback reporting as a way to expand understanding of policy issues related to measuring physician-driven costs of care and quality.|
The problem faced by LTC Medical Groups, as this blog has explored previously, is not the way we practice medicine but who we treat. PA/LTC patients fall into two distinct populations:
- Short-term post-acute care – traditional SNF ‘rehab’ patients. Nearly 100% of these individuals enter the Nursing Facility following a qualifying Part A Hospital stay. Each has significant institutional costs associated with Hospital and SNF care. If the LTC Medical Group provides the ‘plurality of primary care’ during the year, the patient is assigned to them under VBP.
- Traditional LTC patients – this is the population that resides in Assisted Living Facilities, Congregate Housing, or Nursing Facilities (previously called Intermediate Care Facilities). This population is in the last chapters of life and has a unique risk profile. CMS recognizes this different risk for Medicare Advantage (Part C) plans.
In preparation for the meeting with CMS, I took my deepest dive yet into QRUR supplementary reports. I wanted to see if the reports could “expand understanding of policy issues related to measuring physician-driven costs of care and quality.”
This is still relatively unfamiliar territory for me, and probably for most other practice managers and providers. Today’s blog is an introduction to our analysis of the 2013 QRUR reports that were issued by CMS for all group practice with 2013 Part B paid claims. The base QRUR report is accompanied by four Supplemental exhibits that list:
- Physician and Non Physician EPs billing under the medical group.
- Medicare FFS Beneficiaries attributed to TIN and Care Provided.
- Attributed Beneficiaries Hospital Admissions for Any Cause.
- Attributed Beneficiaries Attributed to the TIN for the Spending Per Hospital episode.
- Individual EP Performance on the 2013 PQRS Measures.
In addition to the reports themselves, CMS published a series of guides to the reports.
Each Part B Provider group gets access to its own QRUR. Based on discussions with other administrators, the registration and downloading process is obscure and frustrating. The biggest hurdles revolve around getting the correct registration in IACS, and then navigating to the reports themselves. They are housed in a different location than PQRS and ePrescribing reports.
Once you have gotten access to, and downloaded, the base QRUR documents, the fun begins. First, I will compliment CMS for the wide array of individual patient data they supplied – along with HIC numbers and index codes on each of the reports with patient data. This made it relatively easy (said by a guy with two IT departments and at least five colleagues to lend assistance along the way) to generate some interesting analyses. Basically, my colleagues created files from our own patient records with the patient’s HIC #, first and last DOS, and total 2013 encounters. That is not all the data we would ultimately want/need from our own records, but it let us do some initial analyses to better understand what information we are missing.
I say the analyses are interesting because they fall far short of being actionable, the stated objective from CMS. I will also state that the layout of the CMS excel tables was maddening. The tables included multiple merged cells, and numerous keyboard [space]s inserted for improved readability. Those elements make the tables, as presented, useless for Pivot Tables – the preeminent Excel tool kit for exploring large data files. I hope that we will have the opportunity to provide some detailed feedback to the CMS team working on these reports.
The 2012 report cycle gave us an initial opportunity to dig into the data. Based on that work a year ago, we knew we wanted to begin our analysis by asking this question;
– Is there a difference between Short Stay patients (SNF or POS 31), and Long Stay patients (NF/ALF or POS 32, 33, and 13)?
The follow-up question is how to display the results?
After hours of trial & error fiddling with the QRUR supplemental reports, we prepared the following information for our CMS meeting.
Of those patients that our group served in 2013 which were attributed to us – was there a difference between those served in Post-acute settings vs. LTC? Our assumption was that we would capture more of our LTC patient census, which was proved correct based on initial analysis:
Overall 53% of our census was ‘attributed’ in 2013; of that number, 61% were admitted prior to 2013.
Next, we examined the QRUR reports for those attributed patients and found distinct resource use patterns based on year of admission:
|Admission Year||Before 2013||2013|
|Hospitalization during 2013?||No||Yes||No||Yes|
|Standardized FFS Medicare Cost||$9,062||$56,013||$14,095||$67,068|
|Average HCC percentile*||80||84||63||66|
|% of Attributed Patient Census||44%||16%||9%||31%|
The HCC Percentile represents the relative acuity (risk) of the total Medicare Population. The group’s census had relative risks between the 63rd and 84th percentile. By definition, the mean is the 50th percentile.
There is much more analysis required before we can effectively advise our providers how to influence overall Medicare spending for Attributed patients, but what is apparent at this point:
– LTC Patients (those admitted prior to 2013, have lower costs than post-acute (2013 admissions), but higher HCC scores. That reflects the fact that ‘acuity or risk’ as reflected in the HCC percentile reflects the patient’s status in the preceding year. LTC Medical groups acquire these patients via attribution during the most acute phase of their illness (usually not reflected in the prior year’s acuity for community dwellers.)
– LTC Medical Groups are held responsible for the Hospitalization costs of Post-acute patients that often existed prior to ever seeing the patient. This almost guarantees our high cost classification.
I hope that this initial data will give interested readers food for thought over the holiday season. This is the last blog post for 2014; look for entries to resume in early January with more observations on the practice of LTC Medicine.
Regards and Best Wishes for a Happy and Prosperous 2015!