CMS data dump tells policymakers nothing

This April, the Centers for Medicare & Medicaid Services (CMS) released the details of $77 billion in payments to physicians, labs, and other medical providers in an effort to increase transparency. But after the dust of the screaming headlines fades, what can policymakers really glean from this Medicare data dump? Quite simply: nothing.

All providers should embrace complete transparency of healthcare costs – this is particularly important as patients bear a higher payment burden through increased deductibles and co-pays.  Unfortunately, the release of the Medicare data is both incomplete and without context. As such, not only is the goal of empowering patients not advanced, but the presented data unfairly harms independent physicians, particularly those in single- and multi-specialty group practices. Of even greater concern is that misinterpretations of the data may be used to advance public policy recommendations, which would be nothing short of dangerous.

So what is missing from the Medicare data dump that renders it essentially useless for policymakers? For starters, an individual physician’s payment data are packaged by their National Provider Identifier (NPI) number, which allows for tracking of services by a single provider, but does not address utilization by practice – this overstates billing and payments for some providers while understating the same for others. If a physician performs services as an independent contractor for more than one practice (for example, a pathologist or radiation oncologist), that doctor’s information is listed under their NPI, so, segregating data from each individual site is impossible under the current format. Similarly, some practices bill part-time employees and more junior-level physicians under the NPI of the supervising physician, which inflates the numbers for the supervising physician while the utilization of the other physicians may not be captured at all.

Furthermore, CMS utilization data consists of simple procedure counts using lists of Common Procedure Terminology (CPT) codes.  This confounds the ability to evaluate physicians in group practices by failing to account for the development of internal specialization. Creating disease or procedure specific expertise within groups is a critical component to improving quality and reducing costs. For example, in our large urology group, we have physicians with special training in female urology, robotic surgery and infertility, to name a few; if a provider refers a patient to one of his associates with a particular expertise, that specialized associate may appear to be performing more services than his peers when, in reality, he is performing those services for his peers. This is particularly troublesome as the CMS data do not include any information if a particular service is provided for fewer than 10 beneficiaries – those doctors that are referring to their peers may well not register as providing any services for a particular CPT code although their scope of practice may well include diagnostic expertise relative to that procedure. There is simply no way to determine the quality of services provided from CPT data alone.

Just as billing and payment data do not reflect utilization or quality, so too do they offer an incomplete picture of income. The data presented reflects top line revenue only – there is no information presented regarding the costs of these services.  All independent physicians must pay staff salary and benefits, office rent, utilities, professional liability insurance, medical equipment and supplies.  A failure to capture these overheads is especially misleading when considering pharmaceuticals and ancillary services. In the case of pharmaceuticals, agents are billed on a cost-plus basis derived from their average selling price (ASP), plus 6 percent.  After the Sequester, reimbursements for these agents were reduced by 2 percent of the total billed amount – not the total cost amount. This reimbursement reduction effectively slashed the margin for these drugs to 4.2 percent. This narrow margin barely accounts for the cash flow impact of acquiring and storing these agents as well as potential losses due to product expiration and breakage.  Ancillary services such as advanced imaging, pathology and radiation oncology are extremely capital intensive and require not only substantial initial investments in technology but ongoing overheads for maintenance. In addition, these procedures generally require highly trained, well-compensated technical staff. All of these expenses must come out of the reported reimbursement.

Perhaps the greatest shortcoming of the data is the failure to completely disclose the difference in costs between physicians’ offices and hospitals.  Because hospitals operate under a separate fee schedule, they are able to charge substantially more for precisely the same service when compared to independent physicians.  Although CMS released data on hospitals last week, it chose to only publish the top 30 Ambulatory Payment Classification (APC) codes (as opposed to all services by all doctors in the provider data dump).  As the CPT codes bundled in each APC aren’t provided, it’s virtually impossible for the layperson to compare rates for outpatient services.   Furthermore, the APC doesn’t include the professional charge for an employee physician – this is billed in addition to the APC.  Looking at the data, this means that Medicare paid nearly $1.5 billion dollars in facility fees for outpatient clinic visits alone.  And good luck if you want to see where cancer treatment is less expensive;  while the exact same CPT code for intensity modulated radiation therapy is reimbursed 30% higher in the hospital than in the physician office, this information isn’t part of the data dump.

Performing a service, even at a high volume, is completely unrelated to appropriateness or medical necessity, and it is not an indicator of the quality of the service performed. Topline revenue data provides no insight into the costs associated with providing those services.  The incomplete information does not help guide patients to choosing more affordable alternatives when making complex healthcare decisions.  If transparency is the goal, then CMS should release complete data for all sites of service, along with comprehensible explanations – until then, policymakers would do well to consider the Medicare physician payment data in context, rather than jumping to incomplete assumptions and false conclusions.

Kapoor is chairman of Health Policy for the Large Urology Group Prwctice Association (LUGPA).