Scrapping Consultations Blew Medicare Budget

November 30, 2012

Next year primary care physicians will receive a Medicare pay hike, assuming that Congress postpones a scheduled 26.5% cut mandated by the program's sustainable growth rate formula. Instead of further bloating the budget deficit, the Centers for Medicare & Medicaid Services (CMS) will fund the pay hike by trimming Medicare rates for specialists. The plan is to achieve "budget neutrality," an accounting mantra in a time of fiscal anxiety.

CMS officials hopeful about budget neutrality might want to study an article published online November 26 in the Archives of Internal Medicine about another government initiative that boosted primary care reimbursement. Lead author Zirui Song, PhD, and coauthors describe the "unintended consequences" of a CMS decision to eliminate reimbursement for consultation services — mostly conducted in physician offices and other outpatient settings — in 2010 and to use the savings to increase fees for office visits with new and established patients by roughly 6%. Even with the raise, however, reimbursement for office visits still fell far short of what consultations once garnered.

There also was a 0.3% raise in 2010 for initial hospital and nursing home visits. However, the study excluded the inpatient side of the policy change because it was so minor, said Dr. Song, a healthcare economist and third-year medical student at Harvard Medical School in Boston, Massachusetts, in an interview with Medscape Medical News.

Consultations — the province of opinions and advice sought by one clinician from another — were billed mostly by specialists, whereas office visits remain the bread and butter of primary care physicians. Under the new policy, specialists have to bill what were once consultations as lower-paying office visits, so ostensibly they are taking a pay cut. In contrast, primary care physicians enjoy a pay raise because office visit rates are somewhat higher.

In 2009, CMS said increasing or decreasing reimbursement for any particular wing of medicine was not the reason for ditching consultations. Instead, the agency had concluded that "there is no substantial difference in work" between consultations and office visits. Even so, the CMS policy had the effect of narrowing the pay gap between primary care physicians and their specialist brethren.

As with Medicare raises for primary care in 2012, CMS said its plan to stop paying for consultations and boost rates for office visits would be "budget neutral." Medicare outlays would neither decrease nor increase, according to CMS.

Dr. Song begs to differ. When he and his colleagues compared Medicare spending per beneficiary on consultations and office visits from 2007 to 2009 with spending on office visits in 2010, he found a 6.5% increase.

What blew the plan for budget neutrality? Dr. Song and colleagues write that higher spending did not reflect an increase in the overall volume of outpatient encounters from 2007-2009 to 2010, which remained flat. Instead, roughly two thirds of the spending increase stemmed from the higher office visit fees and one third was attributable to patient encounters coded at a higher level of complexity than before. In other words, physicians were more prone to select, say, billing code 99214 for an office visit with an established patient than a lower-paying 99213.

The overall uptick in coding levels was not dramatic, but it was big enough for specialists to recoup roughly one third of the Medicare reimbursement they were expected to lose because of the demise of consultations.

Although CMS failed to achieve budget neutrality, it did advance the cause of pay equity for primary care physicians, which the agency has championed. Dr. Song and colleagues write that primary care physicians accounted for 58% of the 6.5% spending increase in 2010 compared with a 42% share for specialists.

Risk for "Potentially Undesirable Coding Behavior"

Higher coding levels explain only about one third of the unexpected increase in Medicare spending, but they receive a disproportionate share of attention in the study by Dr. Song and colleagues. Then again, physician coding is in the news.

CMS worries about a decade-long trend of Medicare spending on evaluation and management (E/M) services — which includes office visits and, at one time, consultations — outpacing the growth of Medicare spending on all physician services. In addition, recent investigations by the Center for Public Integrity and the New York Times have detailed how electronic health records can inflate Medicare bills for E/M services.

When CMS announced in 2009 that it would stop reimbursing consultation services the next year, it arrived at budget neutrality on paper by assuming that these patient encounters would convert to other E/M services at the same level of complexity. In other words, a midlevel consultation coded 99243 would become either a midlevel office visit with a new patient (99203) or a midlevel office visit with an established patient (99213). The CMS math also assumed that half the consultations would become new-patient office visits and the other half would become established-patient office visits.

As it turned out, CMS was off the mark on the latter assumption. Most consultations converted to new-patient office visits, according to the study. CMS reimburses these visits at a higher rate than established-patient office visits, so the agency's miscalculation could help account for the unexpected increase in Medicare spending, Dr. Song told Medscape Medical News.

CMS cautioned physicians that they should not use these budget neutrality assumptions as a guideline for how they should actually code encounters previously reimbursed as consultations. Instead, they should follow the coding guidelines for office visits and other E/M categories.

As the study by Dr. Song and colleagues indicates, some specialists did not simply change their midlevel consultations into midlevel office visits but, instead, went a floor higher. The biggest increase in "coding intensity" occurred in the shift from consultations to new-patient office visits.

The study authors write that this bump-up need not suggest something shady, because CMS regulations give physicians more than one way to skin an E/M code. When more than half of a visit consists of face-to-face counseling with a patient, for example, a physician can base the code on the total time of the visit as opposed to meeting tedious requirements for the medical history, physical exam, and medical decision-making, which is the usual method. The time threshold for a level 3 outpatient consultation (99243) had been 40 minutes. A physician "may justifiably determine" that an appropriate substitute for such a visit — if it lasts 45 minutes — is a level 4 office visit with a new patient (99204), which has a time threshold of 45 minutes, the study authors write.

There is a downside to this coding flexibility, however. Any government policy that tries to economize only by adjusting the price component of healthcare spending is "susceptible to coding changes that (at least partially) offset its intended effects," the authors note. Whether the change represents a good faith selection of a higher code or illegitimate "upcoding" is hard to say, because the rules afford so much leeway.

"[T]he inherent flexibility and subjectivity of code definitions," Dr. Song and colleagues write in conclusion, "could lead to potentially undesirable coding behavior in response to fee-based policies."

Dr. Song told Medscape Medical News that his findings do not necessarily predict the outcome of the CMS decision to raise Medicare rates for primary care physicians at the expense of specialists in 2013, all calculated not to increase Medicare spending. The reimbursement changes next year go beyond the E/M services that he analyzed. In addition, the increased spending on outpatient encounters that he identified in 2010 after the demise of consultation codes may not be permanent.

Nevertheless, said Dr. Song, the study "is a cautionary tale about future price-based policies...that don't take into account what the behavioral response on the coding side might be."

Several of the study authors have disclosed a variety of relevant financial relationships. A full list may be found on the journal's Web site.

Arch Intern Med. Published online November 26, 2012. Full text