Part IV: Cost and Improvement Activity Performance Categories

Editor’s note: This article is the first part of a five-part series on the Centers for Medicare and Medicaid (CMS) Quality Payment Program (QPP) in 2020.

  • Part I provides an update to the current requirements for MIPS, emphasizing key components that changed in 2020.
  • Part II will discuss the current requirements for the MIPS Quality performance categories.
  • Part III will focus on the current requirements for the MIPS Promoting Interoperability performance categories.
  • Part IV will detail the requirements for the cost and improvement activity performance categories of MIPS and MIPS audit considerations.
  • Part V will review the 2020 QPP Alternative Payment Models and MIPS APMs.

By Michael Stearns, MD, CPC, CFPC, CRC

The Centers for Medicare and Medicaid Services (CMS) made significant changes to the Quality Payment Program’s (QPP’s) Cost and Improvement Activity performance categories for the 2020 performance year. CMS added additional cost measures and modified the attribution rules for two of the most commonly scored cost measures. For groups reporting improvement activities, 50 percent or more of the clinicians in the group will be required to participate, a major change from the 2019 requirements.

This article will explore the updated requirements for these two performance categories.

MIPS audits started in June 2019 and will be ongoing. CMS requires that practices retain documentation of activities related to MIPS performance for at least six years, creating potential challenges for practices and compliance professionals. This article will also provide recommendations that may help to reduce the risk of negative audits and potential financial penalties.

Cost Performance Category

The cost performance category of MIPS is weighted at 15 percent of the total MIPS score in 2020, the same value it held in 2019. This category’s weighting has not been determined for 2021 but it will have a weighting of 30 percent in the 2022 performance year and in future years, as required by statute. Cost measure performance scores are based on patient demographic information and on data submitted Medicare Part A and Part B claims. There is no reporting requirement for practices.

CMS added 10 additional episode-based cost measures for the 2020 performance year (Table 1), bringing the total number approved for 2020 to 20 measures. CMS plans to continue adding additional cost measures in 2021 and subsequent years, including measures that cover additional specialties, such as behavioral health and other specific medical conditions.

Under the MIPS Value Pathways initiative (discussed later in this article) Quality, Improvement Activities, and Cost measures that address the same clinical condition will be grouped together and may result in additional cost measures being developed.

Table 1. 2020 MIPS Cost Measures
Initial Measures Case Minimum
Medicare Spending Per Beneficiary – Clinician (MSPB-C) (Revised) 20
Total Per Capita Cost (Revised) 35
Measures Approved Starting in 2019
Elective Outpatient Percutaneous Coronary Intervention (PCI) 10
Knee Arthroplasty 10
Revascularization for Lower Extremity Chronic Critical Limb Ischemia 10
Routine Cataract Removal with Intraocular Lens (IOL) Implantation 10
Screening/Surveillance Colonoscopy 10
Intracranial Hemorrhage or Cerebral Infarction 20
Simple Pneumonia with Hospitalization 20
ST-Elevation Myocardial Infarction (STEMI) with Percutaneous Coronary Intervention (PCI) 10
Measures Approved Starting in 2020
Acute Kidney Injury Requiring New Inpatient Dialysis 10
Elective Primary Hip Arthroplasty 10
Femoral or Inguinal Hernia Repair 10
Hemodialysis Access Creation 10
Inpatient Chronic Obstructive Pulmonary Disease (COPD) Exacerbation 20
Lower Gastrointestinal Hemorrhage (*only for groups) 20
Lumbar Spine Fusion for Degenerative Disease, 1–3 Levels 10
Lumpectomy Partial Mastectomy, Simple Mastectomy 10
Non-Emergent Coronary Artery Bypass Graft (CABG) 10
Renal or Ureteral Stone Surgical Treatment 10

Case minimums, as shown in Table 1, must be met for the measure to be attributed to a practice. Once a measure is attributed to a clinician or group, CMS will sum the total costs attributed to the practice as the “observed cost.” CMS will also determine an “expected cost” based on standardization, risk adjustment, and other parameters depending on the measure. The ratio of the observed costs to the expected costs is at the core of how CMS determines the performance score for the costs measure.

Performance in the Cost category has become a significant challenge for many practices for the following reasons:

  • CMS collects claims data during the performance year and then determines attribution and cost performance scores in the following year. CMS also calculates a national benchmark for each measure that is determined after the performance year. Practices have limited to no visibility as to how they are performing during the current year.
  • A proportion of costs assigned to each measure are typically generated by other clinicians and groups. To determine a reasonable estimate of cost performance during the performance year, practices would need access to the universe of real-time Part A and Part B claims and certain demographic data, regardless of which practice is submitting the claims. This information is restricted at this time.
  • Cost measures may be attributed to clinicians and groups that have relatively little control over healthcare expenditures for their attributed patients. However, CMS is making an effort to improve the attribution process in 2020, which is detailed later in this article.
  • Cost performance can be significantly influenced by risk-adjustment scores. Some MIPS practices may not have the staffing or level of familiarity with Hierarchical Condition Category (HCC) coding needed to achieve optimal risk-adjustment scores.
  • Cost score performance determination requires the number of attributed patients to reach a case minimum threshold only. Whether the practice has 35 or 3,500 attributed patients for a cost measure, the same calculation methodology is employed. This could impact scores based on outliers in smaller group of patients, although CMS makes an effort to limit the effect of outliers on performance.

CMS assigns attribution for cost measures at the clinician or group level based on models specific to each cost measure. If no cost measures are attributed to a clinician or group, the cost category weighting (15 percent in 2020) will be reallocated—in most cases, to the quality category. If one or more cost measures are attributed to a practice, CMS will use the average score on the attributed measures to determine the final cost score.

As compared to other areas of MIPS, CMS uses relatively complex methodologies to determine performance on cost measures. Practices can influence their cost scores through utilization containment programs (e.g., coordination of care efforts that reduce unnecessary admissions of emergency department visits) and through accurate risk-adjustment coding.

All cost measures undergo risk adjustment. Risk adjustment factor scores for cost measures are determined using a model similar to Medicare Part C (Medicare Advantage). Proper risk-adjustment coding can have a significant impact on performance in this category. Practices are encouraged to optimize their ICD-10-CM and corresponding HCC coding practices, in particular as the cost categories weighting will reach 30 percent by the 2020 performance year.

The two original cost measures in the MIPS program, the Total Per Capita Cost (TPCC) measure and the Medicare Spending Per Beneficiary (MSPB) measure underwent significant changes in their attribution model in 2020. CMS updated (and renamed) these two measures for the 2020 performance year.

Practices will need to gain familiarity with how cost scores are determined in order to optimize their performance. Additional information is available on the Quality Payment Program website.

The revised TPCC measure’s performance score is determined by the overall cost of care (based on Medicare Part A and Part B claims) delivered to Medicare beneficiaries for the entire calendar year, with an emphasis on primary care services. It is a payment-standardized, risk-adjusted, and specialty-adjusted measure.

Payment standardization takes into consideration variations in the geographic cost of providing care, funding for medical education, and other factors that allow cost to more comparable between clinician organizations. CMS uses the CMS-HCC risk adjustment model (v.22) to determine risk-adjustment scores for the TPCC model. Specialty adjustment is discussed below.

Patients are excluded from the TPCC measure if they are not enrolled in Part A and Part B for every month during the performance year (unless they were new enrollees or in the event of a patient’s death), if they are enrolled a private Medicare plan (e.g., Medicare Part C) and based on other criteria.

To increase the attribution rate of the TPCC measure to primary care providers, CMS created exclusions for non-primary care clinicians based on HCFA specialty designation codes. TPCC attribution now requires a combination of an E&M service and a general primary care service or a second E&M service from the same clinician or group within a specified time frame.

The TPPC measure was also modified so that the risk adjustment model accounts for changes in the patient’s health status throughout the performance year. This is done by compiling the risk-adjustment data over 12 months prior to a “candidate event.”

Table 2 provides a summary of steps involved with determining attribution for the TPCC measure in 2020.

Table 2. Attribution Steps for the TPCC Measure
Step Summary
Identify Candidate Events
  • A “candidate event” (e.g., an encounter) identifies the start of a primary care relationship between a patient and a primary care clinician
  • The candidate event is determined through evaluation and management encounter codes and primary care “services” (e.g. a screening chest X-ray) that are consistent with the services provided by primary care clinicians
  • Two encounters or an encounter and a primary care service must occur in close proximity
  • Multiple candidate events may be initiated during the performance year, including those initiated by the same primary care practice, a different primary care practice, or a specialty practice
Apply service category and specialty exclusions
  • Once the candidate events are identified, the clinician or group that billed for the E/M primary care services is attributed to the measure
  • However, as this may include non-primary care clinicians, CMS will exclude non-primary care providers based on service “category” (e.g., global surgery services) and HCFA specialty type
Construct Risk Windows
  • The candidate event (e.g., an annual wellness visit) opens a “risk window”
  • CMS divides the performance year into 13 four-week intervals, referred to as “beneficiary months.”
  • The allowed Part A and Part B payments for the patient are attributed to a clinician during beneficiary months where the risk window and performance year overlap
Attribute Beneficiary Months to Clinicians and Groups
  • CMS will determine the beneficiary months that fall within the risk windows of the candidate events that were initiated by the clinician/group
  • When more than one clinician or group is attributed to a patient’s care, CMS will attribute the measure to the clinician or group “responsible for the majority share, or plurality, of candidate events provided to the beneficiary”

Once the candidate events and attribution has been determined, CMS go through several additional steps (Table 3) to determine the score for the measure.

Table 3. Determining the Score for the TPCC Measure
Step Summary
Calculate Payment-Standardized Monthly Observed Costs
  • Monthly “observed” costs are the sum or all Medicare Part A/B claims billed for a patient during a beneficiary month
  • These costs go through standardization process to adjust for geographic differences in cost and other factors
Risk-Adjust Monthly Costs
  • The risk adjustment model for the TPCC measure uses the risk factor score from a year prior to a beneficiary month (i.e., it is updated more frequently than other models)
  • CMS uses different risk-adjustment models for new enrollees to Medicare, continuing enrollees, patients in long-term facilities (CMS-HCC V22) and ESRD patients (CMS-ESRD v.21)
  • CMS will then divide the observed costs for each beneficiary month by the average risk score
  • To remove outliers, risk-adjusted monthly costs are then “Winsorized” at the 99th percentile. (Monthly costs that are above the 99th percentile are reduced to match the cost of the 99th percentile.)
  • If the beneficiary is attributed to multiple clinician groups in a given month, the monthly costs are “normalized” to account for expected cost differences between practice settings
Specialty Adjust Monthly Costs
  • CMS uses a specialty adjustment approach that takes into consideration different costs associated with services provided by specialists and the “weighting” of the specialty clinician or group’s costs of care during the episode window
Calculate the TPCC Measure Score
  • The final score for the TPCC measure is determined by:
    1. Taking the primary care clinician’s or group’s average risk-adjusted monthly cost across all attributed beneficiary months
    2. Divide this value by the specialty-adjustment factor for the attributed clinician or group
    3. Multiple this resulting ratio by the national average per capita cost the average (non-risk-adjusted, winsorized) observed cost across the total population of attributed beneficiary months

The revised MSPB measure is now referred to as the “Medicare Spending Per Beneficiary – Clinician” (MSPB-C) measure. It compiles expenditures associated with acute care hospital admissions (episodes) throughout the calendar year, using all Medicare Part A and Part B allowed costs for the three days prior to admission to 30 days after hospital discharge.

The MSPB-C measure’s attribution methodology was revised in 2020 to assign a higher weighting to clinicians that were more likely to be in a position to control costs for an episode. Prior to 2020 each episode of care was attributed to the clinician billing the “plurality” Part B services during the inpatient stay.

For 2020, episodes were split into medical and surgical episodes. Medical episodes are attributed to any clinician or group billing at least 30 percent of the evaluation and management (E/M) services during the inpatient stay. Surgical episodes are attributed to the clinician or group that performed the main procedure for the episode.

CMS is also excluding unrelated services from the episode’s cost specific to groups of DRGs aggregated by MDC level, e.g., the costs associated with an orthopedic procedure during a hospitalization for a gastrointestinal disorder (based on DRG) is excluded.

Like all cost measures, the MSPB measure calculation methodology uses standardization, risk adjustment and winsorization (outlier removal) to determine more equitable levels of performance on this measure across all primary care providers. The final score determination for the MSPB measures is then determined via the following (simplified) steps for each attributed clinician or group:

  1. Determine the sum of the ratio of all observed episode costs/all expected episode costs.
  2. Multiply this sum by the national average payment-standardized observed episode cost.
  3. Divide this value by the total number of episodes attributed to a clinician or group.
Episode-Bases Measures

The remaining 18 cost measures approved for 2020 are divided into acute inpatient and surgical measures (Table 1). For acute inpatient episode-based measures (e.g., the Lower Gastrointestinal Bleeding cost measure) attribution will be assigned to the group billing at least 30 percent of the inpatient E/M services during the inpatient stay.

If this is not applicable, the clinician the clinician or group that billed at least one inpatient E/M service during the stay is attributed. For procedural episodes (e.g., Femoral or Inguinal Hernia Repair) the episode will be attributed to the clinician or group that bills the code that triggers the episode.

Cost Performance Final Score Determination

Performance in the cost category will be based on scores achieved for up to 20 attributed measures. CMS will determine the performance for each measure by comparing its calculated performance to national benchmarks. The benchmarks are determined by performance during the same calendar year as the cost measures. Performance scores for each measure range from 1-10 points once the benchmark is applied. The average performance on each measure (as compared to the benchmark) determines the overall performance in the cost category.

See Table 4 for cost performance score example.

Table 4. Cost Category Score Calculation Example
Cost Measure Points Earned by Clinician or Group

(Range 1-10 points)

Medicare Spending Per Beneficiary – Clinician (MSPB-C) (Revised) 5.5
Total Per Capita Cost (Revised) 6.2
Elective Outpatient Percutaneous Coronary Intervention (PCI) 7.8
Knee Arthroplasty 7.6
Revascularization for Lower Extremity Chronic Critical Limb Ischemia 5.9
Routine Cataract Removal with Intraocular Lens (IOL) Implantation 8.7
Screening/Surveillance Colonoscopy 5.4
Intracranial Hemorrhage or Cerebral Infarction 7.4
Simple Pneumonia with Hospitalization 9.1
ST-Elevation Myocardial Infarction (STEMI) with Percutaneous Coronary Intervention (PCI) 3.4
Acute Kidney Injury Requiring New Inpatient Dialysis 5.6
Elective Primary Hip Arthroplasty 7.8
Femoral or Inguinal Hernia Repair 9.3
Hemodialysis Access Creation 3.4
Inpatient Chronic Obstructive Pulmonary Disease (COPD) Exacerbation 5.8
Lower Gastrointestinal Hemorrhage (*only for groups) 6.2
Lumbar Spine Fusion for Degenerative Disease, 1–3 Levels 5.9
Lumpectomy Partial Mastectomy, Simple Mastectomy 8.5
Non-Emergent Coronary Artery Bypass Graft (CABG) 4.5
Renal or Ureteral Stone Surgical Treatment 8.2

In the example above the practice averaged 6.61 points out of possible 10 points per measure. Only measures that meet the case minimum are scored. The cost category performance rate for the practice is 66.1 percent. To determine how many MIPS points the practice will received from the cost category, the performance rate is multiplied by the weight of the cost category, which is 15 percent in 2020. This yields a total of 9.92 MIPS points for this practice from this category. However, the practice’s best possible total MIPS score will be 93.92.

Preventing a loss of total MIPS points secondary to suboptimal performance in the cost performance category will become increasingly difficult for practices. In 2022 performance year, when the cost category will be weighted at 30 percent, the same performance as detailed in the example above will yield a total MIPS score of 19.8 out of potential 30 points, or a loss of 10.2 potential total MIPS points. Assuming their performance (combined with the complex patient bonus) yielded the equivalent of a perfect score in the other three categories, the highest possible total MIPS score for the practice would be 89.2 points.

Performance in the cost category can be improved by:

  • Avoiding all unnecessary expenditures such as a reduction in the number of preventable admissions and emergency department visits, as well as selective use of medications, durable medical equipment, diagnostic procedures, and therapeutic procedures in a manner that does not adversely affect patient care.
  • Optimizing risk adjustment scores for their patient population through education, chart reviews, and tools designed to improve risk adjustment performance.

The audit risk for the cost category of MIPS is relatively low, although auditors could review claims and documentation that supports levels of risk adjustment, much as they have done for a number of the Medicare Part C programs.

MIPS Improvement Activity Performance Category

The weighting of the Improvement Activity performance category remained at 15 percent for the 2020 performance year.

For most measures the performance period is any 90-day continuous period during the performance year. Activities can be a denovo initiative that practices start during the performance year or an established program they have had in place for several years.

To earn the maximum number of points in this category, clinicians or groups need to earn 40 points. There are two types of improvement activity measures; high-weighted activities that earn 20 points and medium weighted improvement activities that earn 10 points. These point values are doubled for clinicians and small groups of 15 or fewer clinicians.

Prior to 2020 only one clinician in the entire group had to meet the requirements for an improvement activity and the whole group would earn credit. Starting in 2020, at least 50 percent of the clinicians in each group need to participate in each improvement activity for the group to receive credit. However, clinicians in the same group do not need to choose the same 90-day reporting period for each reported improvement activity.

CMS added two new improvement activity measures (Table 5) and removed 14 improvement activity measures (Table 6) for the 2020 performance year.

Table 5. New Improvement Activities in 2020
Improvement Activities New in 2020 ID
Drug Cost Transparency lA BE 25
Tracking of clinician’s relationship to and responsibility for a patient by reporting: MACRA patient relationship codes. ·       lA CC 18



Table 6. Removed Improvement Activities in 2020  
Improvement Activity Description (Removed in 2020) ID
Participation in Systematic Anticoagulation Program lA PM 1
Implementation of additional activity as a result of improving care coordination lA CC 3
Participation in Quality Improvement Initiatives lA PSPA 14
Annual Registration in the Prescription Drug Monitoring Program lA PSPA 5
Initiate CDC Training on Antibiotic Stewardship lA PSPA 24
Unhealthy alcohol use lA BMH 3
Participation in a QCDR, that promotes use of processes and tools that engage patients for adherence to treatment plan lA BE 11
Use of QCDR to support clinical decision making lA BE 2
Use of QCDR to support clinical decision making lA BE 2
Use of QCDR patient experience data to inform and advance improvements in beneficiary lA BE 9
Participation in a QCDR that promotes implementation of patient self-action plans lA BE 10
Use of QCDR to promote standard practices, tools and processes in practice for improvement in care coordination IA CC 6
Leveraging a QCDR for use of standard questionnaires lA AHE 4
Leveraging a QCDR to standardize processes for screening lA AHE 2
Use of QCDR data for quality improvement such as comparative analysis reports across patient populations lA PM 10
Participation in CMS Transforming Clinical Practice Initiative IA CC 4
MIPS Value Pathways

CMS has proposed adding MIPS Value Pathways (MVPs) starting in the 2021 performance year. CMS would align and connect measures from all four categories of MIPS (Quality, Cost, Improvement Activities and Promoting Interoperability) associated with a specific condition (e.g., type 2 diabetes). Clinicians would be required to report fewer measures and focus on MVPs that match the type of care they provide. CMS may allow a single MVP to address all MIPS reporting requirements for a clinician or group.

Table 7 provides an example of closely related measures that may be aligned to form an MVP. All Promoting Interoperability Measures would be required and used as the “foundation” for the MVP. CMS has not determined if they plan to make MVPs mandatory or optional in 2021, however, they may represent an attractive alternative for some practices.

Table 7. MIPS Value Pathway for Diabetes
Quality Measures Improvement Activities Cost Measures
Hemoglobin A1c poor control (MIPS ID #001) Glycemic Management Services (IA_PM_4) Total Per Capita Cost
Medical Attention to diabetic nephropathy (MIPS ID #119) Chronic care and preventive care management for empaneled patients (IA_PM_13) Medicare Spending Per Beneficiary
Controlling high blood pressure (MIPS ID #236)
Foundation: Promoting Interoperability Measures
QPP Audits

MIPS Data Validation Audits started in June 2019 and will audit MIPS data reported for the 2017 and 2018 performance years. CMS has advised practices to retain information related to MIPS reporting for at least 6 years after it is submitted. Multiple audit vulnerabilities have been identified in each of the four MIPS performance categories and within certain aspects of the Advanced APM programs.

Avoiding negative audits will require a detailed understanding of the requirements for each reported measure. Many of the requirements have ambiguities that may require clarification from CMS. Clinical documentation integrity (CDI) tailored tightly to MIPS measure requirements plays a key role in reducing the likelihood of negative audits.

MIPS Data Validation audit requirements are maintained for each performance category and provide guidance to auditors. The are available on the QPP website. It can be helpful to review these requirements as part of an audit risk-reduction program.

An audit file should be created for each MIPS performance year. The following represent examples of the information it may be helpful to include:

  1. Practice and provider information for that year, including business name, address, tax ID number (TIN), provider names and national provider identifier (NPIs).
  2. Copies of any email correspondence you may have had with the QPP Service Center
  3. A copy of the Security Risk Assessment document. This must be completed each year for practices that submit data on the Promoting Interoperability category of MIPS.
  4. EHR certification ID.
  5. Screenshots that include the practice’s MIPS participation status on the QPP website. If a practice or clinician is listed as not being eligible for the MIPS on this website but the auditors do not agree, a screenshot will afford some level of protection.
  6. Screenshots demonstrating that the documentation requirements for each of the following was met (for several patients):
    • Submitted quality measures
    • Submitted improvement activity measures
    • Submitted promoting interoperability measures
  7. Documentation supporting and justifying a decision to apply for a Promoting Interoperability hardship exception.
  8. Results from internal MIPS audits conducted by the practice. Internal MIPS audits should review documentation, actions and reported results for each MIPS related activity.

The Centers for Medicare and Medicaid Services (CMS). “2020 Physician Fee Schedule Final Rule.” Federal Register, November 15, 2019.

Quality Payment Program Resource Library.

Ordonez, O. and M. Stearns. “Quality Payment Program.” Presentation on Quality Payment Program potential compliance vulnerabilities at the Health Care Compliance Association Enforcement Conference, November 4, 2019, Washington, DC.


Michael Stearns (, is the founder and CEO of Apollo HIT, LLC.