Statement of American Academy of Ophthalmology

Introduction

Recent studies by the Government Accountability Office (GAO) and MedPAC raised valid issues about how to reform Medicare as an alternative to the sustainable growth rate (SGR) used in fee for service Medicare.  Our statement focuses on two major recommendations they discussed at the May 10 hearing—that Medicare should move to profile physicians and to bundle or group services to beneficiaries.  Policy leaders say that to make these tools effective, they must be tied to physician payment under Medicare.

 Profiling

The GAO report released this month called for a link of Medicare physician pay to efficiency – defined as providing and ordering a level of services that meets the patient’s health care needs, but is not excessive, given the patient’s health status.  The document claims the Centers for Medicare and Medicaid Services (CMS) has the tools available today to profile physician practices for efficiency.

Profiling is the collection of data to compare doctors on their costs of providing services and to rate them on the basis of the ratio of their actual costs to the expected costs for delivering a specified service or the care of a patient’s condition over a defined period of time.  Private purchasers have had recent experience with profiling.

Key problems with profiling are:1) who defines the “expected” costs 2) how is the patient population risk adjusted and 3) what is the appropriate number of episodes of care required to evaluate efficiency

 GAO says that if CMS had additional authority, it could pay physicians similarly to private sector plans which use profiling.  A recent report conducted for the Massachusetts Medical Society on a recent private sector experience, gives us concern about the real value of linking Medicare payment to profiling.  The Massachusetts study found questions about the accuracy of the data particularly related to patient diagnosis which is critical to determining patient risk or severity of illness.  In addition, the report found that physician profiling at the individual level caused increased administrative burdens for insurers and unintended consequences for both physicians and their patients that affected quality of care.

• Profiles must differentiate between sub-specialists and patients severity of illness

While we acknowledge the increased demand by consumers and payers for more transparency in order to enable them to value the delivered services, the use of billing profiling by CMS is today unable to differentiate sub-specialists from generalists and among patients with differing co-morbidities.  Grouper software often used in profiling, which purports to be able to compare doctors on the basis of cost on similar patient populations, makes assumptions of risk adjustment on the basis of administrative claims data which have never been validated because they are proprietary.

In particular, there needs to be adjustments for age, case mix and levels of chronic or acute conditions within the practice’s patient population.  Many ophthalmologists treat a high percentage of elderly patients with diabetes and the eye conditions associated with the disease.  The number of years with the disease should be taken into account when formulating any profile.  Furthermore, within the specialty of ophthalmology, those who are further trained within a subspecialty will likely see more severe or chronic patients. 

In December 2006, CMS provided the first confidential feedback reports containing reporting and performance rates to the physicians who submitted reports on measures in early 2006.  CMS also intends to give physicians who participate in its new Physicians Quality Reporting Initiative (PQRI), a larger bonus reporting program, confidential feedback on their performance on quality measures.  This early attempt at profiling will be received by the individual physicians in mid 2008.  At that time, the Academy and other medical groups will work with CMS to analyze the usefulness of the data. 

• CMS data will need significant refinement and validation before linking payment to profile

Strategies to measure and encourage quality services and understand resource use must be crafted carefully to avoid serious unintended consequences. We applaud CMS’s goal of encouraging physicians to provide the right care at the right time and in the right setting.  Demonstrations that are underway through CMS will give us much of the analysis we need in order to proceed correctly.  Congress should keep in mind that CMS is in the very early stages of an effort to properly measure physician resource use.

• Even as a feedback mechanism, after data issues have been addressed, impact and value should be evaluated.

Data used as part of a quality improvement program for educational purposes or feedback on review of medical record documentation should be presented to physicians in a user-friendly manner^.  The methods for collecting and analyzing the profile data must be fully disclosed to both the physician and the consumer. The methodology for determining the profiles must be explained to both providers and consumers in easily understandable language, because complex statistical analysis is the methodology often used. 

Any established norms should be based on valid data collection and profiling methodologies, and must use a sample size that is of sufficient statistical power. Interpreting results that are based on insufficient sample size may lead to erroneous conclusions and inappropriate actions.  

Data sources used to develop profiles of physicians have many limitations. This is especially true of surveys, medical records, and claims data because of their limited ability to assess patients' health status and wellness.  These limitations must be clearly identified and acknowledged by Medicare or any other payer and other reviewers to itself, its patients, and its enrollees.  Additionally, standards, guidelines, or practice parameters used for any physician profiling must be derived from the evidence-based publications that are developed and approved by the specialty organization that is the primary specialty of that physician.

 Bundling to Reduce Overuse

 MedPAC proposes payment reform that puts physicians at greater financial risk for services – giving physicians incentives to furnish and order services more efficiently.  Medicare already bundles preoperative and follow-up physician visits into global payments for surgical services.  Specifically, MedPAC suggests a bundled rate that includes separately billable drugs and laboratory services under the current payment method.  In fact, MedPAC is in the process of examining bundling the hospital and physician payments for a selected set of diagnostic related groups (DRGs) to increase efficiency and coordination of care.  For example, they plan to examine the physician services furnished to patients before, during and after inpatient hospitalizations for medical DRGs to assess whether a global fee should be applied, similar to surgical DRGs.

 The Academy, as a surgical specialty, has a lot of experience with bundling payment for surgical services and the disincentives under this approach for over utilization of ancillary services and visits related to a surgery.  Bundling an episode of care for medical diagnoses can be done if the tools are there – Ophthalmologists have done that for diabetic retinopathy laser surgery with a global fee.

The Academy, however, has concerns about linking physician payments to hospital services because of adverse experience physicians have with the way hospitals allocate costs for the provision of services.  Furthermore, it is unclear about how such a payment would work and whether or not it would place physicians at financial risk when it comes to allocation of payments.

Conclusion

We do not believe Medicare should move at this time to tie payment to physician profiles and efficiency measures. Data issues and the lack of adequate severity of illness adjustment currently threaten the relevance and the accuracy of a physician profile under Medicare. Because of this, we suggest pilot testing before proceeding on linking payment to profiles and measures. Even as a feed back mechanism, the impact and unintended consequences need to be studied before devoting significant resources to this endeavor.

For more information go to the Academy’s Web site at www.aao.org