The Effect of Medicare's Physician Quality Reporting System (PQRS) on Quality of Care

Select to access the PowerPoint® presentation (135 KB).

Slide 1

The Effect of Medicare's Physician Quality Reporting System (PQRS) on Quality of Care

Bryan Dowd
Tami Swenson
John Kralewski
Robert Kane
Jon Christianson
University of Minnesota

Robert Coulam
Simmons College

Slide 2

What is the Physician Quality Reporting System?

• Beginning in mid-2007 physicians were given the opportunity to report a new set of quality measures for Medicare beneficiaries under the Centers for Medicare & Medicaid Services (CMS) Physician Quality Reporting System or PQRS.
• The PQRS measures are developed and approved by organizations such as the National Quality Form (NQF) and National Committee for Quality Assurance (NCQA).
• Examples of PQRS Measures:
  • Diabetes Mellitus: Hemoglobin A1c Poor Control in Diabetes Mellitus. Developed by the NCQA. A patient aged 18 through 75 years with diabetes mellitus whose most recent hemoglobin A1c was greater than nine percent.
  • Coronary Artery Disease (CAD): Oral Antiplatelet Therapy Prescribed for Patients with CAD. Developed by the American Medical Association-sponsored Physician Consortium on Performance Improvement. A patient aged 18 years and older with a diagnosis of CAD who was prescribed oral antiplatelet therapy.

Slide 3

PQRS Measures

• PQRS includes both process and outcome measures.
• PQRS includes measures indicating both desirable and undesirable outcomes.
• Currently, the PQRS reporting system is voluntary—physicians earn an incentive payment simply for reporting PQRS measures. 1.5 percent (2007-2008) increased to 2.0 percent (2009-2010):
  • Later decreased by ACA, to become a penalty for non-participation in 2015.
• No rewards or penalties tied to the performance on PQRS process or outcome measures.

Slide 4

An Example of a Part B Claim With PQRS Reports

Image: A section of a sample Part B claim form is shown. The CPT Codes for Evaluation and Management Visits are highlighted by red arrows.

Source: 2010 Physician Quality Reporting Initiative Implementation Guide: CMS.

Slide 5

The Specific Aims of Our Project

Quantitative Analyses:

1. Is there a relationship between the PQRS measures and the outcome measures (avoidable utilization and cost) in cross sectional data?
2. Was the onset of PQRS reporting associated with a change in the value of the outcome measures?
3. Was the frequency of PQRS reporting associated with a change in the value of the outcome measures?
4. Do these results vary according to the race, ethnicity, or sex of the patient? Do they vary by type of chronic illness?

Slide 6

Results So Far—Quantitative Analysis

To date, we have results on the second and third aims of our quantitative analysis:

2. Was PQRS associated with a change in outcome measures?
3. Was frequency of PQRS reporting associated with a change in outcome measures?

Slide 7

Data

• We originally planned to analyze claims data from 2000 to 2010, allowing us to establish a clear time trend for each practice in the pre-PQRS time period.
• Budget limitations restricted our beneficiary sample size to 1,999,999 beneficiaries per year for four years of data.
  • Pre-PQRS years: 2005 and 2006.
  • Post-PQRS years: 2008 and 2009.
  • 2007 was a transition year.

Slide 8

Revised Sampling Strategy

• Drew random sample of TINs from the population of TINs that reported a PQRS measure in either 2008 or 2009.
  • = The "cases" or "treatment group" TINs.
• Drew a sample of non-reporting TINS matched by zip code and practice size.
  • = The "control" TINs.
• Drew equal sized samples of treatment and control TINs from the 2008 and 2009 combined data.
  • 2,744 TINs were present in all four years.

Slide 9

Revised Sampling Strategy

• Then we:
  • Drew a 100 percent sample of all beneficiaries from treatment and control TINs who met the denominator criteria for at least one PQRS measure for caseloads less than 1000 beneficiaries.
  • Randomly sampled 1000 beneficiaries in TINs with larger than 1000 beneficiary caseloads.
• Thus, our results apply only to beneficiaries who were eligible for PQRS reporting.

Slide 10

Difference-in-Differences Estimation

Y_jt = X_jtβ_j + Y2006β₂₀₀₆ + Y2008β₂₀₀₈ + Y2009β₂₀₀₉ + PQRSβ_PQRS + T_j,2008β_T,2008 + T_j,2009β_T,2009 + u_jt

Y_jt = Tin-level outcome variable of interest.

X_jt = a vector of beneficiary and practice characteristics.

Y200t = 1 if data are from year 200t and 0 otherwise (2005 reference category).

PQRS_j,t = 1 if the practice was a PQRS reporter in either 2008 or 2009.

T_j,t,200t = (a) 1 if the data are from a PQRS reporting practice in 2008 or 2009; or (b) various measures of the volume of PQRS reporting (=0 for pre-PQRS years).

Errors are clustered by TIN.

Slide 11

Conclusions thus far—Quantitative Specific Aims 2 and 3

AIM 2: Was the onset of PQRS reporting associated with a change in the value of the outcome measures? No.

• There was no statistically significant change in the outcome measures associated with the onset of PQRS reporting.

AIM 3: Was the frequency of PQRS reporting associated with a change in the value of the outcome measures? Yes.

• When the reporting effects are allowed to vary by year, the number of PQRS reports per physician is consistently and negatively associated with ACS admissions and inappropriate ED visits, and negatively associated with risk-adjusted cost in 2008.
• The number of PQRS reports per physician also is consistently and negatively associated with ACS admissions, potentially preventable readmissions, and inappropriate ED visits, but is unrelated to risk-adjusted cost.
• The proportion of TIN physicians filing PQRS reports was significantly and positively associated with several measures of inappropriate use in 2008.

Slide 12

The Specific Aims of Our Project

Qualitative Analyses:

1. What influences practices participate or not?
2. What problems and opportunities do practices see in PQRS reporting?
3. Is PQRS reporting connected to quality-assurance (QA) efforts of practices and physicians?
4. Do factors identified in these interviews help to explain our outcome measures?
5. Do our interview results suggest ways to improve the PQRS process?

Slide 13

Data and Methods for Qualitative Analysis

• PQRS implementation issues from the physician group practice's perspective:
  • Analyzes questions of PQRS design, implementation, and process that influence PQRS as a quality-improvement reporting program.
• Data from two sources:
  • Eight focus groups with a national sample of 76 medical group practices—arranged with the help of the Medical Group Management Association.
    • 47 practices participating in PQRS.
    • 29 practices not participating.
  • Interviews at CMS with PQRS managers (coming attraction: October 2012).

Slide 14

Conclusions thus far

• Participation decisions usually made by administrators—little evidence of physician buy-in and few linkages to other quality assurance (QA) activities.
• Viewed as a business decision rather than a QA issue, consequently little QA spill over.
• Wide agreement that the program is not difficult to implement.
• Wide concern that measures are inadequate—specialist associations need to be more visibly involved for measures to have legitimacy.

Slide 15

Conclusions thus far (continued)

• Practice strategies regarding adoption of innovations trumps most other factors: early adopters vs. wait and see.
• Only three non-participants said that they might phase out care for Medicare patients when PQRS becomes mandatory.
• Participating and non-participating sites would like to have a more collaborative relationship with CMS.