Print Friendly and PDFPrinter Friendly Version

Thursday, August 16, 2018

The False Choice Of Burden Reduction Versus Payment Precision In The Physician Fee Schedule

Bob Berenson, Visiting Scholar and Senior Advisor for Value Initiatives at Families USA, and Alan Lazaroff, American Geriatrics Society, explain in this Health Affairs blog that The Centers for Medicare and Medicaid Services’ proposed rule that makes changes to the Medicare physician fee schedule would worsen payment incentives for clinicians, compromising the quality of care and increasing costs for seniors.

This blog originally appeared in Health Affairs on August 15th. 

As co-author of a 2011 New England Journal of Medicine Perspective calling for the long-overdue elimination of the documentation billing guidelines for office visits paid under the Medicare Physician Fee Schedule, the lead author of this post might be expected to applaud the Department of Health and Human Services (HHS) proposal to greatly deemphasize the role of these guidelines in the proposed 2019 fee schedule rule. Regrettably, although HHS deserves credit for finally addressing this 20-year-old, festering problem, its recommended approach for diminishing the need for these intrusive guidelines would have serious negative consequences. If adopted, the proposal would move the fee schedule in the opposite direction from our broad objective of moving payment from “volume to value.”

Pick your metaphor for this proposal – “the tail wagging the dog,” “throwing the baby out with the bathwater.” For a fee schedule, the essential need is for rational, value-enhancing coding and payment. The documentation guidelines (described below) are a secondary effect of flawed coding. If the distinctions between levels in the code descriptions for office visits were clearer, HHS could eliminate the current documentation requirements altogether, while also improving payment.

Instead, the proposal would make payment incentives worse, thereby decreasing value. In short, HHS addresses the real problem of intrusive and burdensome documentation requirements by eliminating needed nuance on payment, to the detriment of beneficiaries and the health professionals paid by the fee schedule. Below, we review the problems caused by the documentation requirements and the flaws in HHS’s proposed remedy. Having tried to find better approaches to evaluation and management (E/M) coding for years, we conclude there are no perfect solutions; all approaches have weaknesses. Nevertheless, we conclude that the best approach would base office visits codes on the amount of time clinicians spend with patients.

Why The Documentation Guidelines Compromise Good Practice

In the Journal article, we argued that the office visit codes, a subset of what are called evaluation and management codes (E/M), were long outdated, not reflecting current practice. The code descriptions emphasize discrete elements of patient histories and physical examinations that often lack clinical importance, i.e., “busy work” that robs practitioners of already insufficient time to actually engage with patients on what matters clinically, and what matters to patients. Before adoption of the documentation requirements, physicians quickly figured out how to inflate the level of code by ignoring the details of the codes altogether and simply “upcoding” their visits to achieve higher payment from Medicare.

The Centers for Medicare and Medicaid Services (CMS) responded to upcoding by promulgating the documentation guidelines in 1995 and revising them in 1997, initially with the active participation of the American Medical Association (AMA). The guidelines essentially required clinicians to count the numbers of elements of histories and physicals they performed to qualify for each code level, with increasing associated payment. The more elements listed, the higher code they could bill.

The guidelines very quickly proved to be counterproductive, yet have been frozen in place for more than two decades. It has turned out that charting often irrelevant elements of the history and physical compromised the integrity of the clinical record; these data dumps often provided inaccurate information as physicians could easily cut-and-paste text from prior encounters in an electronic health record (EHR). There is broad consensus among clinicians that they have trouble finding clinically important information in the midst of a welter of text placed into the clinical record solely for billing purposes. This is sometimes referred to as “note bloat.”

The cut-and-paste modality and checkmarks in EHR data templates actually promote upcoding. The payment guidelines also distort how clinicians spend their time by conveying, especially to newly minted clinicians, that the codes and accompanying documentation specifications describe the appropriate content of the patient visit. They don’t.

Perhaps even worse in terms of moving from volume to value, EHR experts readily assert that the priority placed on documentation over two decades to justify payment – $23 billion dollars in 2017 for office visits alone – has diverted software designers’ focus from the important EHR decision-support functions that would actually improve the quality and efficiency of care (See pages 7-18). In short, these documentation requirements are a prominent reason why many physicians and other practitioners consider EHRs to be a burden rather than an important tool.

The Numerous Problems With HHS’s Proposed Solution

Now for the baby and bathwater part. HHS justifies reducing the role of the documentation guidelines by moving to a single payment, rather than the four different payment levels associated with the four codes currently recognized. (A fifth, low payment level which is used for short, nurse-provided encounters remains untouched). Under the proposal, payment is the same whether the clinician takes five minutes to care for a patient with a stuffy nose or 35 minutes to care for a patient with five serious, chronic conditions, taking many medications, and presenting to the physician with recent onset of unsteadiness and falling. In Medicare, the latter patient is not uncommon.

In other words, HHS solves the upcoding concern – and thus the need for the awful billing guidelines – by simply eliminating the differences in payment reflecting differences in clinicians’ effort and time. Medicare would pay a single fee – one for new patients and another for established patients – based on the volume-adjusted, weighted average of the four current codes’ fees. Practices would continue to pick among the four codes but the payment would be the same regardless of the code submitted on the claim form. Thus, for payment purposes there would be no need for documentation to differentiate among code levels.

In the proposed rule, HHS basically asserted that they were between a rock and a hard place. They could either reduce the burden of the documentation guidelines or maintain granular precision in coding by having multiple levels of payment corresponding to different levels of effort, or “work” in fee schedule terms. It argued that it could not both reduce burden and assure payment precision.

Yet, in the proposed rule HHS wrote, “Most significantly, we have understood from stakeholders that current E/M coding does not reflect important distinctions in services and differences in resources. At present, we believe that current payment for E/M visit levels… are increasingly outdated in the context of changing models of care and information technologies.” Despite this accurate summary of the current problem with visit coding, HHS chose to make the problem worse by ignoring any distinctions in services and differences in resources. In short, HHS is saying that because it doesn’t know how to improve coding, it is going to eliminate the role of coding altogether, an approach that violates the statute’s requirement that fees should be based on relative resources used by practices to provide each service approved for payment.

The perverse financial incentives on practitioners are stark. Why other than pure professionalism would practitioners spend 30 or 35 minutes with a complex patient when the payment rate is now calculated as if the visit takes about half that time? A professional who continually accepted a 50 percent discount for highly skilled work with complex patients could not financially sustain a practice. Such an approach might work out if different clinical specialties had a bell curve distribution of shorter and longer visits, but that is not the case. Some specialists, for example dermatologists and orthopedists, use their specialized expertise to handle their common clinical problems quickly. In contrast, geriatricians, neurologists, and some of the internal medical subspecialties such as endocrinologists and rheumatologists take much longer during an office visit to evaluate and treat their different patient populations.

In the proposed rule, HHS published a table finding little variation in revenues across specialties from this move to a single payment level (along with a few other minor – and somewhat ill-defined – coding changes that seem designed to minimize payment redistribution across specialties). However, the AMA’s analysis finds substantial redistribution of revenues across specialties. But both analyses address average payments across different specialties; there are large intra-specialty variations in the nature of the practice and of office visits. Some practices would do well with a single, average payment while others would face major revenue shortfalls, challenging their survival. The average hides this reality.

Of even greater concern would be the likely impact of the single payment approach on Medicare beneficiaries. The natural response of most practitioners to the reduction in payment levels from four to one would be to reduce the length of visits, asking beneficiaries and their caregivers to make more visits. Patients’ problems that might have been appropriately handled in one 30-minute visit might be instead addressed in two or even three shorter visits, generating more revenues for the practice and avoidable inconvenience and higher cost-sharing for beneficiaries.

That likely behavioral response would also compromise quality. Sometimes, a long visit is needed for practitioners to collect and analyze information to permit good medical decision making. Yet, the financial temptation to compromise on this need is powerful. Already many physicians decline to see new Medicare patients because of the relatively lower Medicare payment rates compared to private insurers’. For physicians who are not dependent on seeing Medicare patients, the move to a single payment level would likely reduce their willingness to care for them.

HHS’s published analysis of the financial impact of the single payment proposal is a static one, failing to account for the likely behavioral response to this radical, single-rate policy. The CMS Office of the Actuary has long found what it calls a "behavioral offset" that occurs when physicians respond to a reduction in visit fee levels by generating more office visits, or upcoding of the level of office visits, or both. In this proposal, upcoding is precluded – codes don’t matter – but because payment for five minutes would be the same as payment for a half-hour, surely physicians would schedule more short visits. Given the likelihood of this perverse response by practices, taxpayers should be concerned about this proposed policy as well. In budgetary terms, this proposal surely would be a “coster,” perhaps one with a high price tag. Of course, policy makers could develop an automatic mechanism that reduces the single payment level to make up for the volume-induced spending increase. It might be called something like “the sustainable growth rate.”

Time Spent Is A Better Approach

HHS argues that “our proposed documentation changes for E/M visits are intrinsically related to our proposal to alter PFS payment for E/M visits,” that is, that changing the documentation requirements necessitates the single payment approach. However, an alternative solution is available and straight-forward. Indeed, HHS almost got there in its own modification of documentation requirements. Even though under the HHS proposal there is no important reason for practitioners to distinguish code levels, HHS would permit practitioners to select among the following options as the basis to determine that any office visit was necessary: continued use of the 1995 or 1997 version of the documentation guidelines, the complexity of medical decision making, or time spent in the visit.

As suggested earlier, the relevant activities clinicians engage in during an office visit vary widely. Sometimes, a comprehensive history and physical are the core of the visit, such as with a neurological evaluation for a patient with new onset of seizures or tremors. Sometimes, medical decision-making dominates a visit, as with patients with multiple conditions on multiple medications. And sometimes, the visit is straight-forward, requiring a particular specialist’s unique expertise, as with a patient with a rash non-responsive to over-the-counter medication. The common denominator in these, and indeed all hypothetical patient vignettes, is that practitioner time spent varies with the required activities.

Currently, typical times spent for each code level are provided as advisory to facilitate correct coding and used more definitively only when the time spent in patient counseling or care coordination comprises the dominant practitioner activity during a visit. These (or revised) time intervals could readily be moved from advisory to prescriptive in assigning the appropriate code level, while permitting total elimination of the current documentation guidelines.

In recent years, HHS has introduced useful, new E/M codes, different than the standard office visit codes, that are explicitly based on the time spent in the visit. The new proposed rule also advances such codes. Most of the new codes for care coordination, telehealth and “communication technology-based” (e.g., “a virtual check-in” to determine whether a patient needs to come in for an office visit) have time thresholds to determine coding. For example, the virtual check-in can be coded and paid if the medical discussion takes “5-10 minutes.”

Similarly, HHS proposes new codes for “prolonged preventive services beyond the typical service time for the primary procedure” for the first 30 minutes and for each additional 30 minutes. And HHS proposes to modify a code used for “prolonged evaluation and management or psychotherapy beyond the typical service time of the primary office visit code,” again based on a threshold of time spent.

In short, time, with requisite documentation instructions has become essential for many codes. The solution of using time spent to document the different levels of office codes is hiding in plain sight. Adopting this approach would solve the documentation problem, avoiding the need to move to the flawed, single payment level approach. We believe that using time in this way would be far superior to either the current system or the solution proposed by CMS.

What Should Happen Now?

HHS proposed this radical approach to revision of how Medicare spends $23 billion dollars without prior consultation with physician and other provider groups, beneficiary advocacy organizations, commercial and Medicare Advantage insurers, or any other stakeholders. Physician organizations mostly have expressed opposition to the single payment approach even as they enthusiastically support reducing the importance of the documentation guidelines. Thankfully, there is a rule-making process that permits public input. However, a 60-day comment period is insufficient to adequately address the merits and demerits of the HHS approach and permit thoughtful consideration of alternatives.

The proposal seems well-intentioned, and HHS deserves credit for finally trying to solve the long-standing documentation guidelines problem. In addition, the department has proposed new, innovative codes and payment approaches in other parts of the proposed rule. Nevertheless, HHS should withdraw the single payment proposal. It is possible, as we have recommended, to use time to solve the documentation need so that revised code descriptions can be drafted based on clinical, rather than payment, considerations.