08/13/2020
Elsa PearsonElsa Pearson is a senior policy analyst with Boston University School of Public Health. Follow her on Twitter: @epearsonbusph. Research for this piece was supported by the Laura and John Arnold Foundation.
The complexity and opaqueness of health care billing seems to be an unfortunate given. Patients expect to be flummoxed by hospital bills and providers expect to fight to get reimbursed. Though this complexity, while frustrating, is not itself fraudulent, it provides the opportunity for more nefarious billing.
Specifically, providers and payers can sometimes hide behind that complexity to intentionally bill erroneously. Upcoding is one potentially fraudulent form of billing. Upcoding occurs when more intensive and expensive diagnoses or treatments are documented than what was actually provided to the patient or medically necessary.
Upcoding is happens and often. But before diving into the research, it is worth noting up front that not all billing investigations have found evidence of upcoding. Not every bill has been upcoded.
For example, one study of anesthesia services reviewed almost 50,000 surgical records from patients aged 50 to 79 years with non-deferrable conditions, expecting to see upcoding trends in patient physical status scores (a measure of patient complexity). Because private insurers that serve the under-65 population pay higher rates than Medicare does for more complex patients, the authors hypothesized that providers would upcode more in younger patients. However, after conducting several regression analyses, they found no evidence of a statistically significant shift in physical status score at age 65, suggesting no upcoding in younger patients as a way to increase payment from private insurers. (Of course, their findings still do not rule out more general upcoding across the age gradient, independent of physical status scores and/or payer incentives.)
Exceptions aside, research overwhelmingly suggests upcoding happens nationwide and often. Massachusetts recently found that a steady increase in coded patient acuity and patient risk scores contributed to higher state health care spending over the past few years. These increases could not be explained in totality by changes in actual patient demographics or disease prevalence, suggesting inappropriate upcoding rather than a true change in the patient population.
Surgical specialties seem to be ripe for upcoding, perhaps given the number of unique services delivered and coded for in a finite period of time compared to nonsurgical specialties. A study of over 1.3 million gastroenterology surgeries completed from 2001 to 2011 analyzed the number of codes assigned to each admission. The authors found a systematic increase in the number of codes assigned to each case over time, even when they isolated only low risk patients (no comorbidities, elective procedures only). While the former may suggest simply more accurate coding practices, the latter suggests unnecessary upcoding given no available explanation due to patient complexity or disease severity.
Another study of over one million gastroenterology procedures and their associated anesthesia services found that the number of patients with higher physical status scores increased by almost 8 percentage points from 2005 to 2013. The authors note this increase “cannot be explained by the severity of patients’ conditions… [nor by] changes in the physician population.”
A 2018 NBER working paper found that Medicare Advantage upcoding costs taxpayers $2 billion a year. The authors noted that “the ‘coding inflation’ correlated with how closely tied the doctors were to the insurance plan, with the [coded] risk scores for enrollees in physician-owned plans 16 percent higher than otherwise would be expected.”
ProPublica reported on an Inspector General report that found that Medicare overpaid $6.7 billion in upcoding in 2010 for evaluation and management clinic visits. What’s more, in a sub-analysis of Medicare claims, the Inspector General found that more than half of the claims were coded incorrectly. Nevertheless, the government said it wouldn’t investigate the physicians most responsible (those who upcode most often) given other active reviews.
ProPublica also found that there are physicians who notoriously, and unnecessarily, bill for more complicated visits and treatments more frequently than their peers. For example, one physician coded and billed for the most complex type of office visit 95 percent of the time when his peers only did so about 5 percent of the time.
Even in light of the evidence, it’s challenging to legally demonstrate that instances of upcoding are fraudulent. One JAMA commentary noted that proving coding fraud is difficult because of how the Centers for Medicare and Medicaid Services delineate fraud and abuse. One must prove intent to prove fraud; abuse is simply poor practice but not intentionally harmful.
There can also be good, or acceptable, forms of aggressive coding that might be viewed as upcoding. This occurs when hospitals, providers, or insurance plans simply try to code and bill more accurately. The administrative burdens of health care are substantial and coding can easily become an afterthought, leading to inaccuracies. Simply addressing this is not necessarily bad, particularly if it aligns payment with actual patient need and complexity or intensity of services delivered. But it can be confused with fraudulent or abusive upcoding.
For example, a lawsuit against the Baylor, Scott & White Health system that claimed the health system “inflated medical codes in order to maximize Medicare reimbursement” was recently defeated in court. The lawsuit argued upcoding between 2011 and 2017 led to $61.8 billion in false Medicare claims. But the court determined the health care system was simply implementing changes soon to be required by Medicare.
The next post in this series will discuss what is being done to limit upcoding now and what else can be done in the future.
