A hospital can achieve both clinical and financial performance improvements by decoding its data on Medicare spending per beneficiary and creating a multidisciplinary team to identify episode variations and opportunities.

This year, some hospitals will find it difficult to ignore the financial impact of their Medicare spending per beneficiary (MSPB) performance rates. Referred to as the efficiency measure by the Centers for Medicare & Medicaid Services (CMS), the MSPB measure was mandated by the Affordable Care Act and is a component of CMS’s Hospital Value Based Purchasing (VBP) program.  

For federal fiscal year 2015 (FFY15), 1,255 hospitals performed so poorly on MSPB measures that they received a 0 for their efficiency domain score in the VBP program, losing millions of dollars in total as a result. In addition, payments were reduced for 721 hospitals due to hospital-acquired conditions (HACs) under CMS's HAC Reduction Program, and for 2,610 hospitals due to avoidable readmissions under the agency’s Hospital Readmission Reduction Program. These are not the best of times.  

In May, hospitals should have received from CMS their 2014 MSPB measures, with detailed patient data. Hospitals should use this information to understand how MSPB affects their organization within the context of the CMS programs. The MSPB measure is the most heavily weighted individual VBP measure, accounting for 20 percent of hospitals' VBP adjustments in FFY15 and 25 percent in FFY16. Beginning in 2016, the MSPB measure also will be incorporated into the value-based payment modifier for physician payments, affecting 60 percent of physicians, according to CMS estimates.a To tackle these complexities, hospital finance executives and their staffs should understand the effects of the MSPB measure and lend their support to joint clinical-financial performance improvement teams.  

The Impact of MSPB

The MSPB measure is based on a hospital’s Medicare claims paid, and is compared with national averages that are price-standardized and adjusted for case mix.b Because the scoring process includes a two-year lag, MSPB performance in CY15 impacts the FFY17 MSPB measure. The urgency to delve into the MSPB data and understand how the organization can improve is real.  

Should addressing issues around MSPB be a priority? That depends on the relative impact that the measure has on an organization’s Medicare payments and the organization’s performance in other CMS programs. Consider these numbers: In FFY17, VBP could affect 2 percent of the Medicare severity DRG (MS-DRG) payment. Because MSPB will be 25 percent of the payment amount at stake, it potentially has a 0.5 percent impact on a hospital’s case-mix-adjusted base-operating payment. If a hospital is at 0, it still has an opportunity in 2015 to improve its 2017 score.  

Furthermore, physicians likewise will be challenged to optimize episode efficiency because in 2016, the Quality and Resource Use Reports (QRURs) for medical groups will include MSPB measures designated as “Spending per Hospital Patient with Medicare.”c 

On April 6, 2015, President Obama signed the Medicare Access and CHIP Reauthorization Act (MACRA), which modifies physician payments beginning in 2017. The secretary of the U.S. Deparment of Health and Human Services will implement a merit-based incentive payment system that will include MSPB-like measures.  

Mobilizing the Requisite Expertise

To make the best use of MSPB data, an organization should establish a performance improvement team that includes the following members:

  • A physician or nurse champion who has authority to guide quality improvement initiatives
  • A finance representative who understands the flow of Medicare fee-for-service claims data
  • An analyst to interface MSPB data with the hospital’s clinical and financial databases, and HAC data
  • A clinician with patient transition responsibilities.  

The team leader should have an in-depth understanding of analytic issues and clear reporting lines to both the physician leaders and senior finance managers.  

A performance improvement team is needed because MSPB results often demonstrate substantial variation and, thus, opportunities to improve. The National Quality Forum (NQF) carefully scrutinized CMS’s use of the MSPB measure before approving it through a vote of its member councils. NQF was concerned about attribution of cost and resource utilization to physicians and asserted that this efficiency measure should be paired with quality measures.d This is a clear call for joint financial-clinical team approaches.  

Decoding MSPB Data

The MSPB measure does not look solely at the hospital admission. Instead, it looks at "episodes," which include the index hospital admission, and compiles spending from the three days before admission and the 30 days after discharge. All of that spending, including claims paid at other facilities, is counted within a single MSPB episode and attributed only to the index hospital. A hospital’s standardized average spending per episode is then compared with national spending. More efficient patient care produces a lower MSPB measure, which is further compared with a hospital’s baseline performance, national averages, and national benchmarks.  

Three perspectives illustrate the paid claims that MSPB measures.

The beneficiary episode. Examination of an individual episode discloses the components that create MSPB measures, as illustrated in Exhibit 1.

The major diagnostic category (MDC) care pathway. CMS statistical models are developed at the MDC level and grouped to identify 30-day post-discharge care pathways, as shown in Exhibits 2 and 3. 

Physician profiles. Analyses at the physician level may reveal variations in care, as shown in Exhibit 4. 

Exhibit 1 illustrates a timeline for a Medicare beneficiary with a condition affecting the circulatory system, representing MDC 5. The patient was admitted to the hospital with syncope, discharged, and rehospitalized, and also received care at a skilled nursing facility (SNF). The callouts in this exhibit suggest further operational steps or questions for management and physicians.  

Exhibit 1: Medicare Spending Per Beneficiary: An Episode
Medicare Spending Per Beneficiary (MSPB): Analysis of an Episode of Care

The risk-adjustment panel (2) represents the point when hierarchical condition categories (HCCs)—risk-adjustment diagnoses recorded for 90 days prior to a patient’s episode—are considered. The MSPB risk-adjustment methodology adjusts the MSPB measure for age, severity of illness, and enrollment status indicators. Specifically, the methodology includes 12 age-categorical variables and 70 selected hierarchical condition category (HCC) variables derived from the beneficiary’s claims during the period 90 days prior to the start of the episode.  In this example, the presence of HCC 105, vascular diagnoses, would add $405 to the predicted payment, and chronic obstructive pulmonary disease (COPD) 108 would add $502. Full and appropriate documentation ensures that the hospital will maximize the multiple CMS risk adjustments.  

The episode (3) includes the patient’s index hospitalization, the three prior days, and the 30 days following discharge. The following factors affect the MSPB measure:

  • Efficiency of care prior to the admission
  • Correct DRG codes, given that in MSPB models, most DRGs without complications and comorbidities are negatively weighted (e.g., DRG 312, syncope, is –$2,872.), which means that insufficiently documented DRGs can lower the predicted payment amount
  • A discharge plan that minimizes the likelihood of readmission
  • The most efficient discharge venue for the patient, given the likelihood that most of the variances between total claims paid and expected payment (about 80 percent) occur within the 30-day period of post-acute care.e

The five highest Part B payments (4) are attributed to the claims' National Practitioner Identifier numbers. The financial scoreboard (5) combines the values for MDC, DRG, patient age, and HCCs to determine the predicted payment amount. This example includes MDC-5 for $11,334 and DRG 312 for –$2,871. The amount paid for the episode—$28,700—exceeds the predicted $11,700 by $17,000, with a resulting 2.45 MSPB rate (28,700/11,700). (Patients with multiple chronic diseases often have high MSPB measures. Patients with an acute encounter who are discharged home and not readmitted reflect efficient care and can offset the chronic disease patients.  

The Technical Aspects of MSPB Data

Each May, CMS provides hospitals with their Hospital Specific Report, which enumerates eligible episodes of Medicare fee-for-service patients discharged from Jan. 1 through Dec. 1 of the previous year. CMS calculates each hospital’s MSPB measure performance rate and compares it with the national expected value. CMS also models the average standardized spending and average expected spending per episode for each of the 24 MDCs. The MSPB national 50th percentile threshold reported in 2014 was 0.98.  

For MSBP measures, lower is better. These measures are later converted into efficiency domain scores for VBP from 0 to 100 percent, where higher is better. Because MSPB will count for 25 percent of the VBP score, a high efficiency domain score can mitigate poor performance in other VBP domains.  

To analyze hospital MSPB data, the CMS data first should be reconciled with the hospital’s internal records. Then the following three files, provided with the Hospital Specific Report, should be linked using the assigned episode number:

  • The Hospital Benefit Risk Scores file, which provides the DRG, MDC, and HCC regression weights for each patient and the resulting predicted payment amount
  • The Hospital Episode File, which documents the actual claims paid to each provider  

The resulting merged data file can be analyzed to detail episode timelines and potential drill-downs.  

First Drill-Down: Major Diagnostic Category

To find the lessons in the MSPB data, it is best to start with the MDCs that offer the greatest opportunities for efficiencies and then develop MSPB 30-day care pathways to identify these potential efficiencies.  

As illustrated in Exhibit 2, it is necessary to calculate approximate MSPB measures (column C) for each MDC by dividing the hospital average standardized spending per episode (column G) by national average expected spending per episode (column H). High-volume MDCs with an MSPB measure greater than 1.00 are of particular interest. Exploring individual patients with MSPB measures at the tails of the distribution will provide additional insights. In this example, the Respiratory System and Circulatory System MDCs would be good starting points.

Exhibit 2: Examples of Medicare Spending Per Beneficiary (MSPB) Measures by Major Diagnostic Category (MDC)
Examples of Medicare Spending per Beneficiary Measures by Major Diagnostic Category

Exhibit 3 shows an MSPB 30-day care pathway for a single MDC. This categorization algorithm can identify how often a post-discharge care pathway is used and the average MSPB measure for those episodes.

Exhibit 3: Medicare Spending Per Beneficiary (MSPB) 30-Day Care Pathways
Medicare Spending per Beneficiary 30-Day Care Pathways

Panel 1 represents all qualifying beneficiary episodes for the selected MDC. Panel 2 categorizes the episodes by whether CMS certification number identifies an additional inpatient provider.  

Panel 4  identifies patients with more than one inpatient provider. Such episodes increase the index hospital’s overall MSPB measure.  

Panel 5 episodes are patients first admitted to the index hospital and, within 30 days, readmitted to a hospital with a different CMS certification number, where the second admission has a lesser amount of claims paid than the index admission. Such circumstances raise the following questions:

  • Why was the readmission necessary?
  • Was care effective and sufficient?
  • Why did the patient "leak out" to a different hospital?  

Although the other hospital receives the DRG payment, the claims are reported under the index hospital. A high frequency of patients in this path indicates that it may be beneficial to follow up with attending and discharging physicians regarding the readmissions.  

Panel 6 episodes represent discharged patients admitted to an inpatient rehabilitation facility, where the claims paid to the rehabilitation provider exceed claims paid to the index hospital. In such instances, the aggregate MSPB measures for the patients often are 1.25 to 1.50.  

Panel 7 episodes occur when a patient is readmitted within 30 days to a hospital with a different CMS certification number, and the claims paid to the other hospital exceed claims paid for the index hospital admission. From a care management perspective, the same questions associated with Panel 5 apply. These patients' aggregate MSPB measures often exceed 2.00, negatively affecting the index hospital's MSPB measure. As in Panel 5, the second hospital will receive the DRG payment, while the claims paid for that readmission are reported under the index hospital’s MSPB measure.  

Second Drill-Down: Physician Profiles

Sufficient patient volumes make it possible to identify physicians’ pathway practice patterns using the physicians' national provider identifiers. Exhibit 4 compares the practice pattern of one orthopedic physician (Physician A) with the practices of other orthopedic physicians. For the same DRG, physician A has a lower use of rehab and SNF care, resulting in a MSPB measure of 1.05, while the other physicians have a MSPB measure of 1.40. (Because CMS models are developed at the MDC level, it is necessary to use caution when interpreting physician MSPB values. In this example using a single, homogenous DRG, there is little doubt that practice patterns differ, warranting a follow-up, including the use of inpatient, rehabilitation, and return of function quality measures.)

Exhibit 4: Physician Performance, Medicare Spending Per Beneficiary (MSPB) Orthopedics (Major Diagnostic Category [MDC] 8)
Physician Performance Medicare Spending per Beneficiary Orthopedics Major Diagnostic Category 8

The MSPB data and the 2012-14 trends  they reveal provide multiple opportunities to work with physicians, particularly to:

  • Study patients who "leak out" of the system
  • Compare the use of home health and SNF services for comparable patients
  • Understand the impact MSPB will have on the physicians’ 2016 value-based payment modifier
  • Understand the impact of HACs on MSPB results  

Physician involvement is highly dependent on existing institutional cultures, roles, responsibilities, and experiences with bundled or episode care.  

Time to Get Started  

Finally, maintaining an  episode- and value-oriented mindset is increasingly important to finance executives, senior managers, nursing, and physicians as CMS redirects VBP scores away from inpatient process measures to 30-day value oriented measures. Financial staff and clinical performance improvement staff can work together to change the impact of MSPB measures on VBP.  

Put your MSPB data to use now: Because MSPB performance in CY15 affects FFY17 VBP scores, there is no time to waste.  


Dale N. Schumacher, MD, MPH, is president of Rockburn Institute, Elkridge, Md.

Len Felgner is COO of Health Management Advisors, Inc., Ann Arbor, Mich.

Eric D. Dobkin, MD, is chief quality officer and vice president of quality and patient safety at Crozer-Keystone Health System, Springfield, Pa.

Fern E. Nerhood, is director of administrative services and senior analyst of Rockburn Institute, Elkridge, Md. 

Margaret W. Paroski, MD, is a professor of neurology at the University at Buffalo School of Medicine and Biomedical Services, Buffalo, N.Y.

The authors would like to thank Phillip Ryan, CFO at Crozer Keystone Health System, and Jon Swiatkowski, CFO at Kaleida Health, who provided valuable insights and perspectives; Maura Ciccarelli for editorial assistance; and C. Jean James of the Rockburn Institute for assistance and analytics. 

footnotes

a. "Medicare Program: Revisions to Payment Policies Under the Physician Fee Schedule, Clinical Laboratory Fee Schedule & Other Revisions to Part B for CY 2014 Final Rule," Federal Register, Dec. 10, 2013.

b. For details about CMS price standardization, see " CMS Standardization Methodology for Allowed Amount—v.3 for Services Provided During 2006-2013" (updated June 2, 2014)

c. See cms.gov, " 2013 QRUR (disseminated 9/14)" (page last modified March 31, 2015).

d. Opelka, F., and Travis, C.U., "Re: Cost and Resource Use Project," letter to National Quality Forum council members, Oct. 16, 2013.

e. "Cost and Resource Use Draft Report for Voting," memorandum from Acumen, LLC to National Quality Forum, Responses to Cost and Resource Use Steering Committee Concerns Medicare Spending per Beneficiary (MSPB) Measure (2158), June 27, 2013.  

Publication Date: Tuesday, May 26, 2015